Customer Assist

Learn how to create, build, and manage conversational assistants.
  • Introduction to Customer Assist

    Optimize your organization's support experience with Customer Assist, our virtual conversational assistant (chatbot). You can add conversational assistants to your chat, SMS, WhatsApp, and Kustomer Voice experiences to promote customer self-service, boost agent productivity, and speed up response times.Design and apply custom rules, or flows, for conversations to accomplish the following with conversational assistants:Collect information from customers before they connect with agentsUse AI to respond to customer inquiries using information found in your knowledge baseAutomatically connect customers to the team they want to reachEquip agents with valuable information before they connect with customersRoute Kustomer Voice calls to the right teamReduce repetitive interactions for both customers and agentsProvide the customer with support articles related to their issueIf a customer wants to connect with an agent after interacting with the conversational assistant, the entire conversation history is sent to Kustomer for agents to view in the customer timeline.Once a customer completes their conversation with the conversational assistant, it will submit the conversation to Kustomer. Like any other interaction, your agents will see the entire conversation thread within the timeline. If the customer abandons their conversation with the conversational assistant before interacting with it (sending a message or clicking a button), nothing will be sent to Kustomer for agents to work on.Who can access this feature?User typesAdmins can access the Customer Assist page.Conversational assistants are available for Chat, SMS, WhatsApp, and Kustomer Voice. In this article, we'll explore how conversational assistants work in Kustomer and introduce you to the main concepts you need to know to understand and build conversational assistants in your organization. For a step-by-step conversational assistant tutorial, see Tutorial: Build a conversational assistant.For information on how to set up an IVR for Kustomer Voice, see this article.In this articleDialogsInteractionsFlows (or conversational flows)Volume controlCustomer Assist pageDialogsDialogs are the main building blocks for conversational assistants. Each conversational assistant has two dialogs by default: a Welcome dialog and a Complete Assistant dialog. These default dialogs cover the first and last conversational exchange your customer has with the assistant before the customer connects with an agent.Each dialog is a distinct unit of conversational exchange between the assistant and the customer. You can use the dialog editor to define each dialog's interactions and end interactions.You link dialogs together into a linear or branched flow for each conversational assistant.To learn more about dialogs and how to use the editor, see Conversational assistant dialogs.Interactions You can use interactions to gather information about your customers and guide them through the conversation path. There are two types of interactions you can use:Interactions without redirectsInteractions without redirects are the main building blocks for each dialog. An interaction is a message or question for the customer within a dialog. Interactions allow you to collect information from the customer and help you personalize the support experience.Some interactions allow the customer to reply with text or a response button. In your Kustomer organization, you can set customer button responses to be saved as text, a selected data attribute, or Klass.Interactions without redirects continue to the next interaction within the dialog.End Interactions (interactions with redirects)End Interactions are interactions with redirects. Based on their response, these end interactions direct the customer to another dialog in the flow or end the assistant for the customer (for example, to connect the customer with an agent).End interactions help you link dialogs together to create your conversational flow.For more information on each specific type of available interaction, see Conversational assistant interactions.FlowsFlows refer to the linear or branched conversational flows you can create for each conversational assistant.You link dialogs together with end interactions to create conversational flows.Flows reflect the order of the dialogs you want the customer to experience in their conversational exchange with the assistant.Volume controlVolume control settings allow you to automate the assistant's response to abandoned conversations and missed customer messages sent outside your normal messaging hours. You can offer customers options to submit support requests through an alternate channel.Volume control settings are available in the Assistant settings for your conversational assistant.For more information, please see Conversational assistant volume control settings.Customer Assist pageThe Customer Assist page allows you to create, view, and manage the conversational assistants available for your Kustomer organization.You can access this page by going to Settings> Kustomer IQ > Customer Assist. See Add a conversational assistant to learn how to add a conversational assistant.You can do the following from this page:Add a new assistant by selecting Add Assistant.Duplicate an assistant by going to More Options> Duplicate.Delete an assistant by going to More Options> Delete.Open an assistant's dialog editor and assistant settings by selecting the conversational assistant or the edit icon .Search for existing assistants.Copy the conversational assistant id.Filter through assistants based on status (Draft or Active).Conversational assistant IDThe Customer Assist page displays the name and unique id of each conversational assistant available for your organization.You can use the assistant id to customize Chat SDK settings for each brand. Select the copy iconnext to the assistant id to copy and paste the id.
  • About Lite and Premium assistants

    Customer Assist, our conversational assistant feature, helps you optimize your organization's support experience and promote customer self-service, boost agent productivity, and speed up response times. Using our code-free, visual editor, you can build an assistant based on your organization's specific needs.Who can access this feature?User typesAssistant Lite is available to all organizations for free. Assistant Premium is available to all organizations as a paid usage-based add-on.In this article:What is the difference between Lite and Premium assistants?Available interactions by assistant typeAvailable volume control settingsConvert your assistant typeWhat is the difference between Lite and Premium assistants?With Kustomer, you can create assistants based on your individual business needs.Lite assistants are our free, unlimited usage offering that is available to everyone. This option allows you to set up assistants, using available interactions, that gather the most important customer information upfront, such as their name, contact reason, or email address, before routing their conversation to an agent. You can also use Lite assistants to set up an IVR for Kustomer Voice.Premium assistants are paid and require a KIQ Conversations plan. It includes all the features and interactions available in Lite assistants while offering AI-powered interactions and more advanced automation capabilities.Note: Kustomer Voice interactions are not available with Premium assistants.Available interactions by assistant typeInteractions gather information about your customers and guide them through the assistant's conversation path. This table shows the interactions available for each assistant type. Please note that the channel type mainly determines available interactions. For example, if you create an SMS assistant that does not support the Button Question interaction, it will not be available regardless of the assistant type.InteractionLitePremiumSend a MessageXXCustomer text responseXXButton questionXXData attribute questionXXConditional branchXXBusiness scheduleXXPlay promptXObject SelectorXRun WorkflowXUpdate DataXXRequest feedbackXAI responseXCollect customer inputXCollect language preferenceXNavigation buttonsXXArticle deflectionXRedirect to dialogXXTransfer to an agentXXSend to voicemailXEnd conversationXAvailable volume control settingsYou can use volume control settings to automate how the assistant responds to customer-abandoned conversations, follows up with customers if an agent hasn’t been assigned to the conversation, or responds in the chat, after a set amount of time. Inactive customer reminder is available for both Lite and Premium assistants.Customer-abandoned conversation performs a different action, depending on which assistant type you have:Transfer conversations to an agent with Lite and Premium assistants.Mark conversations as done with Premium assistants.Agent missed conversation is available for both Lite and Premium assistants with the following exception:With Premium assistants, you can also mark a conversation as done if the customer doesn't select a follow-up channel within the specified time.Convert your assistant typeThere may be times when you create an assistant and decide afterward to create a similar, yet different type, of assistant. For example:Convert a Lite assistant to Premium when you are ready to help your customers with instant answers to your most asked questions using AI-generated responses and add automation.Convert a Premium assistant to Lite when you have an existing Premium assistant whose main purpose is to gather information and transfer it to an agent, and you are not ready to use AI or build automation on top of it just yet.You can convert your assistant type by going to Settings> Kustomer IQ > Customer Assist. In this example, we want to create a copy of a Lite assistant. Select More Optionsfor the assistant you want to copy and select Duplicate as Premium. Your assistant will be duplicated as the other type. Note: When duplicating a Premium assistant to Lite, the following features and interactions will no longer be available in the newly created Lite version:AI Response interactionIntentsObject Selector interactionRun Workflow interactionArticle Deflection interactionEnd Conversation interactionMark as done in the Volume Control Abandoned settingsWe recommend you review the assistant carefully after duplicating it to Lite and making any necessary changes to ensure the assistant follows the intended conversation path.
  • Understand KIQ Conversations

    KIQ Conversations is a Kustomer IQ add-on for Customer Assist, our intelligent conversational assistant (chatbot) you can build using a visual editor. With assistants, you can improve customer experiences and eliminate time-consuming tasks by pairing human support with automation capabilities on your Chat, SMS, and WhatsApp channels.Here is a quick summary of the KIQ Conversations add-on:Available Plan(s)Paid & FreeIncluded (Free) PlanRouting chatbots with Lite assistants (these are included as part of your base platform subscription fees and may be subject to limits).Premium (Paid) PlanBuild advanced chatbots with generative AI and automation capabilities with Premium assistantsUnitKIQ ConversationConversation that was not transferred to an agent, and the customer didn’t create another conversation in any channel in 24 hours. See What is a KIQ Conversation for more details.PricingSee Pricing details belowNote: These changes do not apply to customers with active subscriptions to one of our legacy assistant offerings: KIQ for Self-ServiceKIQ Conversations with Assistant ProSee this article for information on what interactions are available on Assistant Pro. Please reach out to support@kustomer.com if you have additional questions.In this articleWhat are the benefits?What is a KIQ Conversation?Full definitionPricing detailsSet up usage notificationsExamples of KIQ ConversationsWhat are the benefits?Our principle when designing conversational assistants is to Start Small, Start Today. You can start by designing a routing assistant with Lite assistants, and then expand to our AI-powered Premium assistants from there, one use case at a time with usage-based pricing.✅ Free unlimited routing botsLite assistants allow you to eliminate manual triage and expedite resolution by collecting information upfront and routing conversations based on user input and data. Lite assistants are free and available to any business customer as part of a paid platform plan, making getting started with Customer Assist easier than ever. Business customers on our Enterprise and Ultimate plans receive unlimited Lite assistants. Other platform plans may be subject to limits.You can convert your assistants from Lite to Premium anytime and start building automation chatbots without a commitment.See About Lite and Premium assistants for more information.✅ Level up with paid AI-powered, automation chatbotsUnlock advanced automation and AI capabilities with Premium assistants, our paid offering that has all of the features found in Lite assistants PLUS the ability to:Use AI to respond to customer inquiries using the information found in your public knowledge baseNotes: The AI response interaction is only available for the chat and WhatsApp channels.Automate end-to-end support workflows for top use cases such as order status, address updates, and delivery schedulingProvide the end customer with support articles related to their issueDeliver quick resolution by automatically detecting and routing conversations based on customer intentFree up agent inboxes by automatically ending conversationsNo commitment is required to start using Premium assistants.✅ Your assistant keeps running even during your busiest season.We’ll charge you for using your Premium assistants at our standard per KIQ Conversation rate. Our progressive pricing model adapts to your needs, even when managing a high volume of conversations. This ensures you pay based on your actual usage, and your price will progressively decrease the more you use, ensuring you receive the most value for your investment. What is a KIQ Conversation?An automated conversation that was not transferred to an agent and the customer didn’t create another conversation in any channel in 24 hours.Full definitionA KIQ Conversation is a conversation that has at least one back-and-forth message sent between your customer and the Assistant and is not transferred to an agent using the Human Takeover interaction within 24 hours after the customer’s first message.A conversation where your customer performs any of these actions within 24 hours after their first message is excluded from being counted as a KIQ Conversation:Creates another conversation in any channel.Provides negative feedback to an AI response interaction.Provides negative feedback to a Request feedback interaction.Provides negative feedback during the Article Deflection interaction, or abandons the conversation without selecting any of the articles shown.Abandons the conversation while the assistant gathers Email, Phone, or Name attributes.Abandons the conversation when asked about their issue during an Article Deflection or while selecting an object in an Object Selector interaction.Note: The AI Response, Request feedback, and Object Selector interactions are only available on the chat and WhatsApp channels for Premium assistants. Article Deflection is also only available on the chat channel for Premium assistants.Pricing detailsBusiness customers who wish to automate conversations using Premium assistants can pay as they go at our progressive per KIQ Conversation rate. As defined above, only KIQ Conversations using Premium assistants will count towards your monthly usage and be billed at our progressive per KIQ Conversation rate.Our progressive pricing model starts at $0.75 per conversation and is designed to adapt to your usage. As your business uses more KIQ Conversations, the price will progressively decrease as you enter into higher tiers based on your total usage. Your total monthly charge will be calculated based on the proportion of your KIQ Conversations within each band based on your total KIQ Conversations in any given month. Conversations per monthDiscountPrice per conversationFirst 0 - 5000.0%$0.75Next 501 - 2,5006.67%$0.70Next 2,501 - 10,00013.33%$0.65Next 10,001 - 25,00020.00%$0.60Next 25,001 - 50,00026.67%$0.55Next 50,001 - 100,00033.33%$0.50Next 100,001 - 250,00040.0%$0.45Next 250,001+46.67%$0.40For example, let's say you used 4000 KIQ Conversations in December: For the first 500 conversations: 500 x $0.75 = $375For the next 2,000 conversations: 2000 x $0.70 = $1,400For the remaining 1,500 conversations: 1,500 x $0.65 = $975Add up all costs for the monthly total: $2,750See this page for more information.Set up usage notificationsYou can set up a monthly threshold for KIQ Conversations to help manage your monthly cost. You will receive email notifications alerting you once you reach 50%, 80%, and 100% of the threshold set. When setting the threshold, you'll see an estimate of your monthly price.To avoid incurring additional charges that month, you can turn your Premium assistant(s) off from the Customer Assist page by going to and selecting Unpublish all Premium assistants.To change your threshold, go to Settings  > Billing > Manage Subscription and select Edit usage notification in the KIC Conversations monthly add-on section.Examples of KIQ ConversationsBelow are examples of common chat scenarios to help explain what counts as a KIQ conversation.Scenario #1: Customer asks about the status of their order and receives an automated answer. The conversation is marked as done automatically using Assistant Abandoned Settings.Customer didn’t create another conversationCustomer didn’t abandon the conversation after being asked for their email, name, or phone numberCustomer didn’t provide negative feedback in Request feedback or Article deflection interactionsCustomer didn’t abandon the conversation without selecting an article during the Article deflection interactionAssistant didn't transfer the conversation to an agentConsidered a KIQ Conversation?Scenario #2: Customer asks a question and the assistant detects the intent of the question and provides an answer to the customer. The conversation is automatically marked as done using Assistant Abandoned Settings.Customer didn’t create another conversationCustomer didn’t abandon the conversation after being asked for their email, name, or phone numberCustomer didn’t provide negative feedback in Request feedback or Article deflection interactionsCustomer didn’t abandon the conversation without selecting an article during the Article deflection interactionAssistant didn't transfer the conversation to an agentConsidered a KIQ Conversation?Scenario #3: Customer asks about the status of their order and receives an automated answer. The customer needs additional help, and the assistant transfers the conversation to an agent.Customer didn’t create another conversationCustomer didn’t abandon the conversation after being asked for their email, name, or phone numberCustomer didn’t provide negative feedback in Request feedback or Article deflection interactionsCustomer didn’t abandon the conversation without selecting an article during the Article deflection interactionAssistant didn’t transfer the conversation to an agentConsidered a KIQ Conversation?Scenario #4: Customer abandons the conversation when asked to input their email address.Assistant didn’t transfer the conversation to an agentCustomer didn’t create another conversationCustomer didn’t abandon the conversation after being asked for their email, name, or phone numberCustomer didn’t provide negative feedback in Request feedback or Article deflection interactionsCustomer didn’t abandon the conversation without selecting an article during the Article deflection interactionConsidered a KIQ Conversation?Scenario #5: Customer abandons the conversation during an article search in chat.Assistant didn’t transfer the conversation to an agentCustomer didn’t create another conversationCustomer didn’t abandon the conversation after being asked for their email, name, or phone numberCustomer didn’t provide negative feedback in Request feedback or Article deflection interactionsCustomer didn’t abandon the conversation without selecting an article during the Article deflection interactionConsidered a KIQ Conversation?Scenario #6: Customer abandons the conversation after doing an article search but doesn't select an article in chat. If they had selected an article to read, it would be considered a KIQ Conversation.Assistant didn’t transfer the conversation to an agentCustomer didn’t create another conversationCustomer didn’t abandon the conversation after being asked for their email, name, or phone numberCustomer didn’t provide negative feedback in Request feedback or Article deflection interactionsCustomer didn’t abandon the conversation without selecting an article during the Article deflection interactionConsidered a KIQ Conversation?Scenario #7: Customer abandons in the middle of the conversation and then sends an email 3 hours after having abandoned the chat conversation.Assistant didn’t transfer the conversation to an agentCustomer didn’t create another conversationCustomer didn’t abandon the conversation after being asked for their email, name, or phone numberCustomer didn’t provide negative feedback in Request feedback or Article deflection interactionsCustomer didn’t abandon the conversation without selecting an article during the Article deflection interactionConsidered a KIQ Conversation?
  • Conversational assistant interactions

    Customer Assist, our conversational assistant feature, provides multiple interactions that you can use to gather information about your customers and guide them through the conversation path. Depending on what you select, you can either have a customer continue to the next interaction within the dialog after answering a question or selecting an option, or you can choose to redirect them to another dialog entirely. The conversational assistant is fully customizable to your business needs.Who can access this feature?User typesAdmins can access the Customer Assist page.To learn the main concepts of an assistant, see Introduction to Customer Assist. For a step-by-step conversational assistant tutorial, see Tutorial: Build a conversational assistant.In this article:Available interactions by channelAvailable interactions by assistant typeAdd an interactionSend messageText questionButton questionAttribute questionConditional branchBusiness scheduleObject selectorRun workflowUpdate attributeAdd an end interactionRequest feedbackAI responseNavigation buttonsArticle deflectionList selectorRedirect to dialogTransfer to an agentEnd conversationAvailable interactions by channelThis table shows the interactions that are available for each channel. InteractionChatWhatsAppSMSKustomer VoiceTwilio WhatsAppMessageBird WhatsAppSend MessageXXXXXText questionXXXXXButton questionXAttribute questionXConditional branchXXXXXBusiness scheduleXXXXXXPlay promptXObject selectorXXRun workflowXXXXXUpdate attributeXXXXXNavigation buttonsXXAI responseXXArticle deflectionXRequest feedbackXXList SelectorXCollect customer inputXCollect Language preferenceXRedirect to dialogXXXXXXTransfer to an agentXXXXXXSend to voicemailXEnd conversationXXXXXAvailable interactions by assistant typeDepending on the type of assistant you have, you may have different interactions available. For more information, see this article. Please note that the channel type mainly determines available interactions. For example, if you create an SMS assistant that does not support the Button Question interaction, it will not be available regardless of the assistant type.Note: See this article for information on the interactions available for Kustomer Voice and how to set up an IVR.Add an interactionInteractions (interactions without redirects) are messages or questions the assistant sends the customer.Interactions without redirects continue to the next interaction within the dialog. The assistant runs through interactions from top to bottom for a dialog. You can reorder interactions. Use interactions without redirects to:Collect information from the customer to help you personalize the support experienceSend multiple messages to provide instructions or contextAsk multiple questions that require a customer text, button, or multi-level list responseYou can add an interaction by selecting Add Interaction in the editor and selecting one of the available options.Send MessageSend Message sends a message to your customer and automatically moves to the next interaction.Use Send Message to send messages that do not require a customer response. Attachments can only be sent in assistants created for WhatsApp. The supported file formats for attachments are JPEG, PNG, and MP4.Text questionText question sends a message to your customer and wait for a text response.You can select to save the customer response as a text attribute (that is, a message) on a standard Conversation or Customer object. This interaction can be used to detect intent with the conversational assistant.Button questionButton question creates custom response buttons for the customer to select.Similar to the Customer Text Response option, you can select to save the button text of the customer response as a text attribute (that is, a message) on a standard Conversation or Customer object.Attribute questionAttribute question creates custom response buttons based on the existing values for data attributes on your Customer and Conversation Klass.You can select the values for the response buttons based on the Klass (that is, Customer or Conversation), the data attribute on the selected Klass, and the data value (for example, Conversation [Klass] > Assigned Teams [data attribute, or field] > CX, Sales, Shipping [separate data values]).The assistant automatically saves the customer button response data to the selected Conversation or Customer Klass in Kustomer. You can select to skip the Attribute question interaction if the data for the selected attribute already exists for the Customer or Conversation object in your Kustomer organization.Conditional branchConditional branch creates personalized interactions based on if-else conditions using conversation and customer data. With this option, you can set up various conditions at any point in your conversation flow that connect and direct customers to any of our interactions or redirect them to another dialog.Using conditional branches, you can solve various use cases based on your customer's input. For example, you can create separate responses based on the email address they are using or if they are considered a higher VIP tier in your organization.Notes:If you redirect all of your conditional branches, you can't add additional interactions after this block or redirect the customer to another dialog.When building your conditions, the In all other cases then statement is optional.You need to verify new email addresses in Kustomer before you can use them in a condition.For more information on operators and to learn more on how to build your conditions, see Understanding filter operators.Business scheduleUse the Business schedule interaction to allow the conversational assistant to perform different actions based on your organization's business hours. For example, you can use this option to ask a customer for their name and email address if they contact you during your business hours or send a message notifying them that your business is currently closed if they contact you outside of those hours. Learn more about defining your business schedules in this help article.You can use this interaction to perform the following actions during or outside of your business hours:Send MessageText questionButton questionAttribute questionRun workflowUpdate attributeRedirect to dialogObject selectorObject selector creates a card in the assistant that looks up and presents authenticated customers with the most 3 recent objects created for a Klass on their timeline that match their input.  With this option, you can use the object selector to add custom object data, such as order information, into your conversation experience.Note: Customers must be authenticated via a secure JWT token.To set up this interaction, first select the Klass you want to use and then customize the card by mapping the attributes to each of the three available fields.Note: The Card Image field is optional. If you don't map anything to this field, the card will only display a title and description.If results that match the authenticated user are found, they can select the object and get more information right in the conversation window. If the customer isn't authenticated or results aren't found, you can redirect them to another dialog to continue the conversation.Run workflowRun workflow calls the workflow you select in the drop-down menu and uses data captured from a customer's interaction with the conversational assistant to return relevant data to your customer.Note: You can create the workflow from this interaction by selecting Create Workflow from the drop-down menu. You'll be taken to the Workflow page to configure the callable workflow.Once you select the workflow, you can specify the data the workflow needs to run and optionally determine where the information you return to the customer is saved. For more information, see Callable workflows and the conversational assistant.Update attributeUpdate attribute changes a value of the Conversation and Customer object attributes from directly within the conversational assistant. For example, you can use this interaction to update a conversation's priority and route it to the correct team, or change an attribute that's required for closing the conversation.Along with new custom attributes, you can use this interaction to update the following attributes:ConversationNamePriorityBrandTagsDefault LanguageCustomerNameDefault LanguageCompanyAdd an End Interaction End Interactions (interactions with redirects) direct the customer to another dialog in the flow or end the assistant for the customer based on their response (for example, to connect the customer with an agent). Use End Interactions to link dialogs together to create your conversational flow. You can create both linear and branched conversational flows.Note: You can only have one end interaction for each dialog. This means you must delete an existing end interaction before creating a new one.You can add an interaction by selecting Add End Interaction in the editor and selecting one of the available options.Request feedbackRequest feedback sends a message to your customers asking for feedback on their most recent support experience.You will need to select redirects based on the type of feedback they give. For example, if they choose the Failure button, you may want to redirect them to an agent to assist them.You can customize the interaction by selecting from 3 different button designs.AI responseAI response uses information from your organization's knowledge base to answer customer inquiries. See this article to Use AI to respond to customer inquiries.This interaction can also be used to detect intent with the conversational assistant. You can customize the AI's response by creating a new profile with custom prompts. Navigation buttonsNavigation buttons create buttons to redirect customers to other dialogs in the flow. Use Navigation buttons to create branched paths for the customer in the flow. For example, you can redirect sales and support inquiries to separate dialogs in the flow.Note: We recommend not using emojis for buttons if using this interaction in a WhatsApp assistant.You can select to save the button text of the customer response as a text attribute (that is, a message) on a standard Conversation or Customer object.Article deflectionArticle deflection searches your knowledge base and returns relevant articles based on the customer response or request. This interaction can also be used to detect intent with the conversational assistant.Use Article deflection to help customers resolve their support requests without an agent.You can select the knowledge base brand the assistant should use for article deflection. You will need to select redirects for the following article deflection scenarios:If there are relevant knowledge base articles:The assistant finds and shares relevant articles from the knowledge base:A dialog redirect for when the article helps the customerA dialog redirect for when the article is unhelpful or the customer needs additional helpIf there are no relevant knowledge base articles:A dialog redirect for when the assistant doesn’t find any relevant articles in the knowledge base.List selectorList selector creates a list of up to 10 buttons that redirect your customer to another dialog when they select a button in the message. You can select to save the customer response as a text attribute (that is, a message) on a standard Conversation or Customer object. This interaction can be used to detect intent with the conversational assistant.Redirect to dialogRedirect to dialog redirects the customer to any dialog in your assistant.Transfer to an agentTransfer to an agent ends the assistant and transfers the conversation to an available agent based on the customer response and agent routing.Use Transfer to an agent to end the assistant and to connect customers with an agent. The agent will have access to the customer information collected by the assistant.End conversationEnd conversation ends the conversation with the customer. The customer cannot reopen or continue the conversation. The customer will need to open a new conversation to request additional support.
  • Conversational assistant dialogs

    Dialogs are the main building blocks for your conversational assistants built through Customer Assist. Each dialog is a distinct unit of conversational exchange between the assistant and the customer. You can use the dialog editor to define the interactions and end interactions for each dialog.You can also create intents to help navigate your customers through the conversation path. With intents, you can train Kustomer to detect specific keywords entered in the conversation and redirect customers to a specific dialog based on what they enter.Who can access this feature?User typesAdmins can access the Customer Assist page.To learn the main concepts of an assistant, see Introduction to Customer Assist. For a step-by-step conversational assistant tutorial, see Tutorial: Build a conversational assistant.In this articleUnderstand the editorAdd a new dialogUse intents with a conversational assistantUnderstand the editorThe dialog editor is the main area of the conversational assistant editor. A new conversational assistant will always have two default dialogs: the Welcome dialog and the Complete Assistant dialog, as seen in the sidebar. You can edit both of these default dialogs in the main area of the editor. The Welcome dialog is always the first dialog in your conversational flow.In the dialog editor, you can:Edit the dialog name (if applicable)View and go to the previous dialog in the conversational flow (if applicable)Add and edit multiple interactions without redirectsAdd and edit the end interaction that links and redirects the dialog in the flowThe sidebar displays previews of the dialogs available for your assistant. Select a dialog in the sidebar to edit the dialog. You can also view the link status of the end interaction in each dialog:Green anchors indicate that the dialog is linked in the conversational flow. Hover over the green anchor to view the dialog linked as the redirect.Red anchors with a strikethrough indicate that the dialog is not linked in the conversational flow. Select an end interaction for these dialogs.Add a new dialogThe first step in creating a conversation path for your assistant is to create a new dialog:From the Dialogs sidebar, select Add Dialog.Enter a descriptive name for the dialog to easily identify the dialog in the Dialogs sidebar and other parts of the dialog editor.Select Add Interaction to add interactions without redirects. Optionally, determine if the dialog starts once customer intent is detected. For more information, see Create an intent.Select Add End Interaction to add an interaction with a redirect that links the dialog to another dialog in the flow or to an action that ends the assistant (for example, transfers the customer to an agent). For more information on each specific type of available interaction, see Conversational assistant interactions.Note: While you can add multiple interactions to a dialog, each dialog can only have one end interaction.Use intents with a conversational assistantWith intents, a conversational assistant can understand when customers enter specific words or groups of words in the conversation and redirect them to a specific dialog so that they can get the correct answer to their original query. You can create up to 30 intents. For more information, see Detect intent with a conversational assistant.
  • Conversational assistant rules and proactive chat

    Assistant rules provide a way for your customer experience team to engage website visitors with targeted messages. With rules, you can specify which dialog in your assistant launches when a customer starts a conversation based on a multitude of factors such as the current page they are viewing or their VIP status. You can also proactively initiate a conversation with a customer using this criteria instead of waiting for them to reach out to you. This helps you engage with your visitors and build a stronger customer relationship with them.Who can access this feature?User typesAdmins can access the Customer Assist page.In this articleCreate a ruleDefine your audienceStart a proactive chatSet up a rule for a specific SMS or WhatsApp phone numberOrder of assistant rulesBefore you create a rule, please note the following:An assistant can only belong to one channel at a time and cannot be used across multiple different channels.An assistant can have more than 1 rule.Each rule can only be used in 1 assistant.Create a ruleYou can create up to 200 rules for your chat, WhatsApp, and SMS assistants. For this example, we created an assistant that welcomes users with a different message depending on whether they are new or existing customers. The rule we create will start a conversation if the user is a new visitor. The following image shows where this dialog is along the conversation path.To create an assistant rule:Go to Settings> Kustomer IQ > Customer Assist.Select the Rules tab and then select Add Assistant Rule.Enter a name for the rule.Select the channel that has the assistant you want to define an audience for from the Channel drop-down menu. If creating a rule for the SMS or a WhatsApp channel, you can set up a rule that launches when it detects a specific phone number.Select the assistant you are creating the rule for and then select the dialog you want the assistant to start with if the rule matches. By default, the assistant will start with the Welcome dialog. The Preview pane on the right shows the dialog you selected.Note: You can only add a rule to a published assistant.Next, determine what criteria the rule must to start on that dialog or assistant. We are going to launch this dialog if the customer's email address isn't set, indicating they are a first-time visitor, and they've been on the current page for longer than 15 seconds.Optionally, select the Use proactive chat check box if you want to automatically start the conversation. For more information, see Start a proactive chat.Note: Proactive chat is only available for web.Select Save Changes. Please note that it may take up to 10 minutes to see your rules reflected in your assistant.As you can see, since the criteria are met, the assistant is skipping the other dialogs in the path and welcoming a new visitor.Define your audienceWith rules, you can use filters to narrow down the exact audience for your assistants. Depending on the channel you're creating the rule for, you can filter down by Company, Conversation, Customer, or Message attributes. Note: If you create a chat rule that filters your audience by customer, users must be authenticated so that we can identify them and match them to the rule. If you are creating a rule for a chat assistant, you can also filter down by Session type. The following fields are available:Current Page is the page the user is currently on. This field is only supported on web.Language is the language the browser is set to.Previous Page is the page that the user was on before landing on the current page. This field is only supported on web.Time on Page (seconds) is how long the website visitor has been on the current page (in seconds). This field is only supported on web.User Agent is information that identifies the user agent your customer is on, such as the browser, its version number, and the host operating system.For more information on creating filters and the available operators, see Understand filter operators.Start a proactive chatYou can proactively invite your website visitors to start a chat conversation instead of waiting for them to initiate one using our proactive chat feature. This provides a way for you to engage with your visitors and offer assistance in the proper context. You can use proactive chat in a variety of scenarios. For example, you can use this feature to:Prevent cart abandonment by offering help when you notice a visitor is on the order page for a certain period.Greet your returning customers and offer a personal experience that develops a customer relationship with them.Promote an offer that is occurring during a short period, such as a holiday sale.Announce a new feature while they are viewing your product page.Let's create a proactive chat for our cart abandonment example. Here, we want to initiate a conversation if the customer has been on the order page for longer than a minute and ask them if they have any questions before submitting their order.First, select the Chat channel and then the assistant you want to use.Note: Proactive messaging is only available for the Chat channel.Next, select the dialog that will proactively pop up. We want to start the conversation while the visitor is viewing their shopping cart, so we are using one named Shopping cart.Note: Proactive chat only supports certain dialogs. Please review how the dialog appears in the Preview section to verify its support.Enter the criteria that will trigger the chat to start. Select Session Time On Page (seconds) Is Equal To 60. Finally, select the Use proactive chat to start the dialog check box.Select Save Changes.Now, if a visitor is on your order page for a minute, the chat will automatically open asking them if they have questions and directing them to an agent.They can select a button or respond to free text from directly within the pop-up, or dismiss the message by hovering over it and closing it out.Set up a rule for a specific SMS or WhatsApp phone numberIf you have assistants for SMS or your WhatsApp integration, you may want to set up a rule that launches when it detects a specific phone number. To do so, select SMS or WhatsApp from the Channel drop-down menu and then select the phone number from the available list. You can also manually enter the number. If you do not select a specific phone number, the rule will launch whenever your audience contacts you via SMS or WhatsApp channels.Twilio for WhatsAppIf using Twilio for WhatsApp, enter the number in the Phone box in the +19191234567 format.MessageBirdIf using MessageBird, you need to enter the WhatsApp ID in the Phone box.To find your WhatsApp ID, log in to your MessageBird account, go to WhatsApp, and then select Install channel. Copy the ID for the WhatsApp channel you are using with Kustomer.Order of assistant rulesThe order of your rules will have an impact on which assistant launches when the criteria are met. The first rule that matches the set criteria is the assistant that will launch. There may be instances where you want an assistant to run before another so you'll want to reflect that in your sort order.If none of your rules match an incoming conversation, the default assistant will run. If you don't want the default assistant to run if none of the rules are matched, select None from the Assistant drop-down menu.Note: Existing iOS and Android chat implementations may use the default assistant. If you want to maintain backward compatibility, we recommend you do not set this to None.
  • Use AI to respond to customer inquiries

    Customer Assist, our conversational assistant feature, enables customers to self-serve and optimizes your organization's support experience. Powered by OpenAI, chat and WhatsApp assistants using the AI response interaction can respond to customer inquiries using precise answers found in your organization's knowledge base.Who can access this feature?User typesAdmins create conversational assistants.In this articlePrerequisitesHow does the AI response interaction workSet up an AI response conversational assistantFAQPrerequisitesTo use AI to respond to your customers' inquiries, you must have a public knowledge base with published articles.The AI response interaction is only available for Premium assistants, which requires a KIQ Conversations plan.How does the AI response interaction workThe AI response interaction uses OpenAI to answer questions in your chat or WhatsApp assistant using information found on your public knowledge base. You can set up an assistant to use this feature using the Respond with AI template, or you can add it as an End Interaction to an existing Premium assistant. You can customize the template to suit your needs best. You can also customize how the AI will respond by creating a new profile with custom prompts. By default, the template will direct users through the conversational flow based on whether or not the answer to their question was found and if feedback was received. Questions will only be answered using information found in published articles on your public knowledge base. In all cases, the assistant will ask the user if they have any other questions and attempt to respond again.After the second attempt, if the user selects negative feedback or the assistant cannot find the answer in your knowledge base, the customer will be transferred to an agent to provide additional support.Set up an AI response conversational assistantYou can start answering inquiries with AI by creating a new assistant using the Respond with AI template. Note: The AI response interaction is only available for Premium assistants on the chat and WhatsApp channels.Go to Settings   > Kustomer IQ > Customer Assist.Select Add Assistant and then select the Respond with AI template.Select the channel for which you want to create the assistant. This template is available for both the Chat and WhatsApp channels.Select the knowledge base you want to use to respond to customers from the drop-down menu. You can optionally select whether you want the template to be in English, Portuguese, or Spanish. By default, it is created in English.Select Use Template.Modify the template as needed. The new assistant comes configured with interactions and default messaging so you can use it immediately. Note: If modifying the text in the template, do not remove the wording that discloses that the assistant is not a person. For more information, see our Best practices for the conversational assistant.Once ready, select Save Draft (if you made any changes), followed by Publish.To start using your new AI-powered assistant, return to the Customer Assist page, select the Rules tab, and select Respond with AI from the Assistant drop-down menu.Once set, your assistant will use your public articles to answer customer questions.FAQQ: Which articles are used to respond to a customer's inquiries?The conversational assistant uses only articles published on your public knowledge base. Content found in internal articles, work-in-progress drafts, or unpublished content will not be used as a response. Q: How accurate are the responses?While we constantly work to improve our model, it's important to remember that AI models are not perfect and may sometimes generate an inaccurate response or "hallucinate." To minimize this as much as possible, the AI response feature conducts an initial search and only responds to questions related to content found in your knowledge base. If a customer asks a question unrelated to your help articles, the assistant will say it cannot find an answer. Q: Are the changes immediately available if I create a new article or update an existing one?In most cases, any updates to your public articles can immediately be used as a response.Q: What assistant type can I use the AI response interaction with? Is it available for all channels?The AI response interaction can only be used with Premium assistants for chat or WhatsApp.Q: Is there a limit to how many questions can be asked per day or month?There is currently no limit. We may reevaluate this based on usage.Q: Are my articles being used to train an OpenAI model?No, your data is not being used to train or deploy any OpenAI model. For more information, see Open AI's API data usage policies.Q. Can the assistant respond in a different language?The assistant can only respond to an inquiry in another language if:The customer's browser language is set to that language and You have public knowledge base articles written in the same language. For example, if a customer's browser is set to Spanish (es-es) and you also have articles published in Spanish (es-es), the AI can find an answer and respond in Spanish. However, if a customer's browser is set to French and you do not have any articles published in French, the assistant will not be able to respond.
  • Detect intent with a conversational assistant

    With intents, Customer Assist can understand when customers enter specific words or phrases in the conversation and redirect them to a specific dialog so that they can get the correct answer to their original query. You can choose from a list of predefined intents, or create up to 80 intents of your own. Who can access this feature?User typesOnly users with access to Intents can create a new intent to use with an assistant. For more information, see Define user access.In this articleUse predefined intentsCreate your custom intentsUnderstand the intentRedirect the customer if the intent is detectedManage your intentsDefine fallback behavior when an intent isn't detectedUse predefined intentsKustomer provides an automatic way for your conversational assistants to detect your customer's needs using a list of predefined intents. The language model can understand sentence variations, differences in capitalization, alternative spellings for words (color vs colour), and typos in words. This table lists all available intents and an example customer query for each one.Intent nameDescriptionCancel orderUse when a customer has inquiries about canceling an order. For example:What do I need to do to cancel order XXEE444?Change orderUse when a customer wants to make a change to an order. For example:Can I add something to order XXEE444?Change shipping addressUse when a customer wants to make a change to a shipping address. For example:I need help updating my shipping address.Check cancellation feeUse when a customer asks to check cancellation fees. For example:I don't know how to see the cancellation fee.Check invoiceUse when a customer is asking for an order invoice. For example:I want to check my latest invoice.Check payment methodsUse when a customer wants to know which payment methods are supported. For example:Help me check payment options.Check refund policyUse when a customer wants to know the current refund policy. For example:Help me check in what scenarios I can request a refund.ComplaintUse when a customer is unsatisfied with their service. For example:My experience with your service was bad, fI want to file a claim.Contact customer serviceUse when a customer wants to know how to contact your customer support service. For example:I need to see what hours I can reach customer support.Contact human agentUse when a customer wants to talk to a real person, not a chatbot. For example:May I speak to an agent?Create accountUse when a customer has problems creating an account. For example:I cannot create a free account.Delete accountUse when a customer wants to delete the account created in your service/web. For example:I do not need the gold account and want assistance canceling it.Delivery optionsUse when a customer wants to know the available delivery options provided by your company. For example:Is it possible to place an order from Seattle?Edit account informationUse when a customer wants to change the information they have in their account. For example:Edit details on my pro account.Get invoiceUse when a customer wants an invoice for their order. For example:I don't know how to download my bill.Get refundUse when a customer asks for an order refund. For example:I need assistance getting my money back.Newsletter subscriptionUse when a customer wants information about your newsletter or email communications. For example:How can I cancel my subscription to the company newsletter?Payment issueUse when a customer has a problem related to payments. For example:I'm having trouble with my online payment and need help.Place orderUse when a customer needs assistance to place an order. For example:Can you help me order some of your items?Recover passwordUse when a customer wants to recover their account password. For example:Can you help me reset the key for my user profile?Registration problemsUse when a customer has problems signing up for your service or website. For example:What do I need to do to notify you of issues with signup?Product review/feedbackUse when customers want to provide feedback related to a product or service. For example:I need help sending some feedback for your products.Set up shipping addressUse when a customer wants to add a new delivery address to your system. For example:The new delivery address isn't valid, can I get some help?Switch account and account upgradesUse when a customer with multiple accounts wants to log in with a different account type. For example:Change to Freemium account.Where is my orderUse when a customer wants to track the status of an order. For example:Want assistance checking my order status.Track my refundUse when a customer wants to know in which step of the process the refund is. For example:I am waiting for a refund of 299 dollarsCreate your custom intentsIf you have additional use cases that cannot be solved using our list of predefined intents, you can create your own using unique keywords and phrases. For more information, see Create custom intents.Understand the intentFor the conversational assistant to have the ability to detect a customer's needs, you have to add an interaction in the conversation path that will "listen" for the intent.  We recommend you add this step early on in your path.Note: The ability to understand intent is only available for the following interactions:AI response (See this article for more information on using AI to respond to customer inquiries.)Customer text responseArticle DeflectionLet's set up an intent that redirects customers looking to return an item to an agent.Open the assistant you want to use the intent with or create a new one.Since we are creating a conversation path to detect if a customer wants to return an item, let's add an interaction that prompts the customer to enter a few words describing why they are contacting you.In the Welcome dialog, select Add Interaction > Customer Text Response.Enter a message to prompt the customer to respond. In this example, we ask "What can we help you with today?"Turn on Understand Intents. The AI icon will appear, indicating that this interaction will help you understand your customer's input.Redirect the customer if the intent is detectedThe final step is to create the dialog that the assistant uses to continue the conversation based on the intent you want to detect.Select Add Dialog and then choose the desired intent from the drop-down menu.Select the dialog the assistant redirects the customer to if the intent is detected. In this example, we will transfer them to the next available agent.Note: You do not need to have this dialog linked to another one since it will continue the conversation only if the intent is detected.Now, the assistant will automatically detect if a customer requests a refund and redirect them to an agent. You can easily view which dialogs are listening for an intent by looking for the the robot icon , which indicates that an intent redirects the conversation to it. You can hover over the icon to see which intent is used.Define fallback behavior when an intent isn't detectedNote: This option is only available for WhatsApp.You can define what occurs if a customer's reply does not match an existing intent by going to an assistant, selecting Assistant Settings, and opening the Behavior tab. You can pick from one of two options:Redirect the customer to another dialog within the same conversational assistant.Continue to the next interaction in the dialog.Manage your intentsYou can manage your intents by going to Settings> Kustomer IQ > Customer Assist and opening the Custom Intents tab. Here, you can see all the intents you have created, the assistant(s) currently using it, and the number of actively detected languages.You can select the edit icon to modify the intent, check its training status, or delete it completely. When modifying the intent, remember that the changes will affect all of the assistants using it.
  • Callable workflows and conversational assistants

    Workflows make it easy to automate processes within the Kustomer platform. Unlike typical workflows triggered by an event, such as when a conversation is created, callable workflows are triggered by a conversational assistant. You can also use callable workflows to automate actions that depend on time-specific conditions. For more information, see Callable scheduled workflows.With callable workflows, you can use data captured from a customer's interaction with conversational assistants to return relevant data to your customer. In this article, we will create a workflow that accepts an order number from a conversational assistant and outputs the return status for that order.Who can access this feature?User typesAdmins can create and manage callable workflows.In this articleCreate the workflowSet up the conversational assistantCreate the workflowThe first step is to create a callable workflow that captures data from the conversational assistant, performs an action based on that data, and then returns output data to the conversational assistant.Go to Settings > Platform > Workflows.Select Add Workflow.Enter a name for the workflow and then turn on the Callable Workflow toggle. We're going to name it Return Status.Select Create.Next, enter the data field(s) that the workflow expects to receive from the conversational assistant. In this example, we want to receive input from the customer.Note: You can make any of these parameters required by selecting the Required Field check box.Select the Output Parameters tab and enter the name and data type of the parameter(s) that will be returned to the conversation assistant.Select Update.Add an Action step that takes the data entered into the conversational assistant and uses it to request information from a specific source. In this example, we're making an API call to get the status of the return.Add a final Action step that sends a response back to the conversational assistant by selecting Kustomer as the Action App and Workflow: Send Response as the Action Event.After you select the Send Response event, specify what parameters will be output to the assistant. In this example, we are going to tell the customer the status of their return.Select Save.Set up a conversational assistantNow that we have the callable workflow configured, we'll set up a conversational assistant that is going to call that workflow. Note: You may need to create custom attributes to save your information. Conversational assistants only support the Conversation and Customer objects. We created customer input1 and return status for this example.Add a new assistant and enter any desired dialogs and interactions, such as a welcome message or any questions you may want to ask your customers. For more information on creating an assistant, see Introduction to Customer Assist.Select Add Dialog.Select Add Interaction > Customer Text Response.Enter the message that will prompt your customers to input the data you need to run the workflow and then select where their response is going to be saved. In this example, we want to save customer input 1, which is going to be the customer's order number and is the input parameter that we defined in step 5 when we created the workflow, in the Conversation object.Select Add Interaction > Run Workflow.Select the Return Status workflow that we created earlier from the drop-down menu and then specify the data you need the customer to enter to run the workflow. You can manually enter it or select Snippets & Dynamic Text to select the dynamic text value from the available list.Tips:You can select Create Workflow from the drop-down menu to create the new callable workflow.To view an existing workflow, select the open icon next to the drop-down menu.Then, specify where to save the response that the workflow returns to the customer. In this example, we will save the customer's return status in Customer return_status.Add any additional dialogs to the assistant, as needed. For this example, we're going to select Add Dialog and then Add Interaction > Send a Message to configure the message the customer is going to receive with their return status.Add any additional dialogs to the assistant, as needed, and select Save and Publish.Now, whenever a customer contacts you via chat inquiring about the status of their return, the conversational assistant will call the workflow and use the order number entered to find the requested information and send it back to your customer.
  • Conversational assistant volume control settings

    Volume control settings allow you to automate how a conversational assistant responds to abandoned conversations and missed customer messages sent outside your regular chat hours. In your volume control settings, you can configure how the assistant responds when the customer abandons the conversation and when the agent misses a customer message over chat, SMS, or WhatsApp.In this articleInactive customer reminderCustomer abandoned conversationAgent missed conversationTo access Volume Control settings, create a new assistant or select an existing one, go to the bottom of the Dialogs sidebar and select Assistant Settings > Volume Control.Inactive customer reminderTurn this setting on to prompt the customer to respond to the assistant's last message if they haven't responded in the specified time. After the reminder is sent, the assistant will resend the last message sent before the customer went inactive.You can configure the following:Remind customer to respond after: The amount of time (in minutes) for a customer to be inactive before the assistant sends the reminder message. The default amount of time is 5 minutes.Reminder message: The message to send to the customer to remind them to respond to the last message that was sent. For example, "Are you there?".If using Kustomer Voice, this setting will automatically route the call to an agent if the customer fails to respond after the specified number of prompts, which repeat every 10 seconds.Customer abandoned conversation (available for chat, WhatsApp, and SMS)Automatically end the conversation if a customer hasn’t responded after a set time. Tip: Turn on this setting to prevent agent routing for abandoned conversations.You can configure the following:Abandon conversation after: The amount of time (in minutes) to wait for a customer response before the assistant ends the conversation. If the customer does not respond to the conversation within the set time, the assistant sends the customer the set Abandon Conversation Message and ends the conversation. The default amount of time is 10 minutes.Note: If you have Inactive customer reminder also turned on, the timer for this setting must be higher than the time set for the reminder message.Abandon conversation message: The message to send to the customer when the conversation has been abandoned. For example, "This conversation has ended because you've been inactive for 10 minutes".Agent missed conversation (only available for chat)Automatically offer the customer an alternate support reply channel, or contact method, if an agent hasn’t been assigned to the conversation or first responded in the chat after a set amount of time. Choose which follow-up channels to offer to customers for missed messages.Note: This setting does not apply to chats after a human user responds first. It is also not supported by an SMS or WhatsApp assistant. Tip: You can create a search for missed chats from customers who select an alternate follow-up channel. To triage or search for missed chats from the assistant chat automation, set a filter with the following search criteria: Conversation (Type) > Assistant Status (Property) > is equal to (Operator) > Missed.You can configure the following:Missed conversation after: The amount of time (in minutes) to wait for an agent's response or for an agent to be assigned to the conversation before the assistant ends the chat. If the agent is not assigned to the chat or does not respond within the set time, the assistant sends the customer a request to select an alternate follow-up method and ends the chat. The default amount of time is 10 minutes.Missed conversation message: The message to send to the customer when an agent hasn’t been assigned or responded in the set amount of time. For example, "Sorry, it looks like no one has become available in the expected time. Please select an alternate contact method so we can follow up with you."Configure missed conversations follow-up channels: The available channels the customer can select for follow-up messages from an agent for missed chats. Selected channels will appear as options for the customer after the Missed conversation message.The follow-up channels available on this list depend on the channels installed in your organization. You can select and configure from the following:Email channelPhone channelSMS/Text Message channelWhatsApp channelNote: Phone, SMS, and WhatsApp numbers must include all country and regional codes. For example +19171234567I'll wait: Allow the customer to stay on the chat and wait for an agent to reply.Mark conversation done if a channel isn't selected after: Automatically mark the conversation done in Kustomer if the customer does not select a follow-up channel in the set amount of time (in seconds). The default amount of time is 30 seconds.
  • Tutorial: Build a conversational assistant

    With Customer Assist, conversational assistants can help you collect information from customers before they connect with an agent to speed up response times and close conversations more quickly.In this article, we’ll learn how to build a conversational assistant that prompts the customer for basic information (for example, the customer’s name and the team they want to reach) and then, based on the collected information, connects the customer to an agent from the appropriate support team.Who can access this feature?User typesAdmins can access the Customer Assist page.In this articlePrerequisites Tutorial overview Add a new conversational assistant Create and edit dialogs Link dialogs into a flow Save, publish, and manage versionsPrerequisitesIf you are new to conversational assistants in Kustomer, see Introduction to Customer Assist before you begin this tutorial.Tutorial overviewYou can build conversational assistants to use across your brands with Customer Assist.Follow along with the tutorial below to learn how to set up a basic assistant with the following sections:Add a new conversational assistantCreate and edit dialogs and dialog interactionsLink dialogs into a flowSave, publish, and manage conversational assistant versionsThrough this tutorial, we will build a sample conversational assistant that asks the customer for their name and the team they'd like to reach before the assistant transfers the customer to an agent from the selected team. Add a new conversational assistantFirst, we’ll create a new conversational assistant from the Customer Assist page.To add a new conversational assistant:Go to Settings > Kustomer IQ > Customer Assist.On the Customer Assist page, select Add Assistant to create a new conversational assistant.In the Add Assistant box, enter an Internal Name, Description, and Public Name for the assistant:Internal Name: Required. Appears as the name listed for the assistant on the Customer Assist page in Kustomer. Use a short descriptive name that will help you manage your conversational assistants.Description: Optional. A brief description of the assistant that appears in your Kustomer app.Public Name: Recommended. The display name of your conversational assistant to appear for customers in the chat widget.Note: You can update the Internal Name, Description, and Public Name later under Assistant Settings in the assistant editor.Select Create to open the editor for the assistant.Create and edit dialogsNow that we have a new conversational assistant, we’ll use the dialog editor to create and add new dialogs for the assistant and also edit the default Welcome and Complete Assistant dialogs.Step 1: Edit the Welcome dialogThe editor automatically displays the Welcome dialog when you select or create an assistant on the Customer Assist page.By default, the Welcome dialog includes a Send a Message interaction with a default message and redirects the customer to the default Complete Assistant dialog.Let’s edit the welcome message to let the customer know that we’ll ask them some questions to help them connect with the right agent.To edit the Welcome dialog, select Welcome from the Dialogs sidebar to open the dialog editor. Edit the text in the Send a Message interaction.Select Save Draft in the dialog editor to save any updates to the conversational assistant.Step 2: Create and add new dialogsThe editor includes two default dialogs for each assistant, a Welcome dialog and a Complete Assistant dialog. The Welcome dialog is the first message the customer views when they start a chat. The Complete Assistant dialog is the last message the customer views before they connect with an agent.You can create and add dialogs for the customer to view after their first message and before their last message with the assistant.For each dialog, you can edit the name, interactions, and end interaction for the dialog.Let’s create and add two new dialogs to the conversational assistant. We’ll create one dialog to ask for the name of the customer and a second dialog to ask the customer for the support team they want to reach.Step 2a: Add new dialog: Ask Customer NameEach dialog contains a set of interactions. For this dialog, we’ll use a Customer Text Response interaction to ask the customer for their name.From the Dialogs sidebar, select Add Dialog to open the editor for the dialog. Enter a descriptive name for the dialog. For this example, we’ll use “Ask Customer Name” as the name for the dialog.From the dialog editor, select Add Interaction > Customer Text Response. Edit the message prompt in the interaction for the customer to enter their name.You can also save the customer’s reply as an attribute. In this case, to save the name as a customer attribute, select in the interaction the option to save the customer reply as an attribute in Customer (Klass) > Name (Field).Step 2b: Add new dialog: Ask for TeamEach dialog contains a set of interactions. For this dialog, we’ll use a Data Attribute Question interaction to ask the customer to select a team to connect with from the available options.From the Dialogs sidebar, select Add Dialog to open the editor for the dialog. Enter a descriptive name for the dialog. For this example, we’ll use “Ask for Team” as the name for the dialog.From the dialog editor, select Add Interaction > Data Attribute Question. Edit the message prompt in the interaction and select to display values from Conversation (Klass) > Assigned Teams (Field).For the Custom Values box, select the teams you want to provide as options for the customer. The customer will view these options as response buttons in the chat.Step 3: Edit the Complete Assistant dialogBy default, the Complete Assistant dialog doesn’t include any interactions and transfers the customer to an available agent.Let’s add a Send a Message interaction to the dialog to let the customer know that they will be connected with an agent from the selected team.To edit the Complete Assistant dialog, select Complete Assistant from the Dialogs sidebar to open the dialog editor. From the dialog editor, select Add Interaction > Customer Text Response. Edit the text in the Send a Message interaction to acknowledge the customer’s selection.Select Save Draft in the dialog editor to save any updates to the conversational assistant.Link dialogs into a flowNow that we have all of the dialogs we want for the conversational assistant, we’ll link the different dialogs into a flow to reflect the order you want the dialogs to appear for the customer.The available dialogs for the assistant appear in the Dialogs sidebar to the left of the dialog editor. The Dialogs sidebar displays the name, interactions, and end interaction for each dialog.Note: Red anchors with a strikethrough indicate unlinked dialogs. Green anchors indicate linked dialogs.Step 1: Review the dialog flowLet’s review the dialog order, or flow, we want for the dialogs in this conversational assistant example:Welcome → Ask Customer Name → Ask for Team → Complete AssistantWelcome dialog: The assistant sends the customer a welcome message.Ask Customer Name dialog: The assistant asks the customer for their name. The customer replies with their name.Ask for Team dialog: The assistant asks the customer for the name of the team they want to reach. The customer selects an option from the available response buttons.Complete Assistant dialog: The assistant acknowledges the customer’s selection and transfers the customer to an agent from the selected team.We’ll reference this flow to link the dialogs in order.Step 2: Link dialogs in orderTo link the dialogs in order, we’ll select and link the dialogs available in the Dialogs sidebar in order based on their position in the dialog flow:From Dialogs, select Welcome to open the dialog editor. Go to the Redirects to Dialog block in the editor and select the next dialog in the flow: Ask Customer Name.From Dialogs, select Ask Customer Name to open the dialog editor. Select Add End Interaction > Redirect to Dialog to redirect the customer to another dialog after their response. In the Redirects to Dialog block, select the next dialog in the flow: Ask for Team.From Dialogs select Ask for Team to open the dialog editor. Select Add End Interaction > Redirect to Dialog. We want the assistant to acknowledge the customer’s selection and transfer the customer to an agent. To reflect this, select the Complete Assistant dialog in the Redirects to Dialog block.Select Save Draft in the dialog editor to save any updates to the conversational assistant.All dialogs available in the Dialogs sidebar, except for the Complete Assistant dialog, should have green anchors next to the listed end interaction for the dialog.Save, publish, and manage assistant versionsLearn how to save, publish, and manage different versions of your conversational assistant.Save drafts of your conversational assistantSelect Save Draft in the dialog editor to save drafts of your progress at any time.Publish your conversational assistantPublish your conversational assistant when your edits are complete and you are ready to add the assistant to your chat widget.Select Save and Publish in the dialog editor to create a new version of the conversational assistant.Once you publish an assistant, you will be able to select and assign the assistant to your available brands.To select a default assistant for a brand, go to the Customer Assist page and select the Rules tab. Select the channel you want to set a default assistant for, and then select the desired brand and assistant from the drop-down menu.If a default conversational assistant is selected for a brand, the customer will chat with the assistant by default.For example, if we selected the sample conversational assistant we’ve built as the default assistant for a brand in Kustomer, the default customer chat would follow the assistant’s dialog flow:Manage conversational assistant versionsYou can save and manage multiple versions of your assistants to track changes and to switch between different versions easily.To create a new version of an existing conversational assistant, select Create New Draft from the dialog editor for the assistant. Select Save and Publish to publish the draft as the latest version.To manage the version of your assistant available for use in Kustomer, select More Options from the dialog editor. Select Delete Assistant to delete your conversational assistant. Select Latest Version to view and compare versions, edit a previous version, or select a new version to be available for use across your brands in Kustomer.
  • Assistant report

    The Assistant report helps you understand how Customer Assist, our conversational assistant feature, is performing by providing a quick way to view metrics, such as how many conversations were completed by the assistant and how many were transferred to an agent.Who can access this feature?User typesManagers have access to reporting.In this articleAssistant statusesAssistant report metricsAssistant Breakdown chartKIQ Conversations chartKIQ Conversations per assistant tableTo access the Assistant report, select Reporting, and then select Assistant.Assistant statusesThis report provides metrics and charts that give you insight into how your assistant is performing. To better understand this data, please review the following definitions for all of the available statuses the assistant can be in during its lifecycle:Running indicates conversations where an assistant is currently in progress. This is the assistant's default status and is final unless there is a change of status.Transferred indicates that the assistant tried to transfer the conversation to an agent.Once the assistant reaches this status, queue rules, business rules, and SLAs will resume. Agents can now be assigned to the conversation, reply to it, or mark it as done.If the agent doesn't reply to the conversation in the time specified in the assistant's volume control, the status will change to Missed.Ended indicates conversations where either the assistant or customer ended the conversation. The assistant ends when the customer reaches the End Conversation dialog. Customers can also end the conversation themselves if the Allow Chats to be ended option is turned on in their chat settings.Abandoned indicates conversations where the customer hasn't responded to an assistant question in the time specified in the Customer abandoned chat volume control setting. It is very common for customers to abandon conversations after they see an answer without acknowledging whether or not it is helpful.Missed indicates conversations where the agents didn't reply in the time specified in the Agent Missed Conversation volume control setting. This status can only be triggered after a conversation is transferred.Interrupted indicates conversations where an agent replied to the conversation in the Kustomer timeline while the assistant was running. This can occur when agents monitor assistant conversations in real-time.The following diagram shows the paths an assistant's status can take.You can use the Assistant Status conversation attribute to create a custom report that filters the exact assistant data you want to understand. With this attribute, your custom report can show conversations where this attribute is set to any of the available statuses (Assistant Status Is Set), or select from one of the available statuses when creating your report criteria.Our custom reporting feature also includes 9 conversational assistant chart templates to help you get started on building your report.Assistant report metricsThe following metrics are an overview of conversations in which an assistant participated.The following metrics are shown:Started is the number of conversations created where an assistant is involved. The total shows conversations segmented by Automated and Transferred to Agents.Automated is the number of conversations where the assistant didn't try to transfer the conversation to an agent. This metric shows conversations where the Assistant Status attribute is equal to Running, Ended, or Abandoned.Transferred to Agents is the number of conversations where the assistant tried to transfer the conversation to an agent. This metric shows conversations where the Assistant Status attribute is equal to Transferred, Missed, or Interrupted. The assistant will change to this status when the customer reaches the Human Handover dialog, or an agent interrupts the conversation. Conversations will also be automatically transferred to an agent if any error occurs with the conversational assistant.KIQ Conversations is the number of conversations that had at least one back-and-forth message sent between your customer and the assistant and is not transferred to an agent within 24 hours after the customer's first message. A conversation where your customer performs any of these actions within 24 hours after their first message is excluded from being counted as a KIQ Conversation:Creates another conversation in any channel.Provides negative feedback to a Request feedback interaction.Provides negative feedback during the Article Deflection interaction, or abandons the conversation without selecting any of the articles shown.Abandons the conversation while the Assistant gathers Email, Phone, or Name attributes.Abandons the conversation when asked about their issue during an Article Deflection or while selecting an object in an Object Selector interaction.Satisfaction Rating is the average Satisfaction rating, out of 5, during the selected period for conversations where the Assistant participated. Ratings are normalized across 2, 3, and 5-point survey scales.Assistant Breakdown chartThis chart shows the weekly count of conversations that have been transferred to an agent, abandoned by the customer, ended, or currently have an assistant in progress.KIQ Conversations chartThis chart shows the weekly count of conversations that were fully automated. A conversation is considered fully automated if it has at least one back-and-forth message sent between your customer and the Assistant and is not transferred to an agent within 24 hours after the customer’s first message.A conversation where your customer performs any of these actions within 24 hours after their first message is excluded from being counted as a KIQ Conversation:Creates another conversation in any channel.Provides negative feedback to a Request feedback interaction.Provides negative feedback during the Article Deflection interaction, or abandons the conversation without selecting any of the articles shown.Abandons the conversation while the Assistant gathers Email, Phone, or Name attributes.Abandons the conversation when asked about their issue during an Article Deflection or while selecting an object in an Object Selector interaction.Note: The Respond with AI, Article Deflection, Request feedback, and Object Selector interactions are only available for the Chat channel on Premium assistants.KIQ Conversations per assistant tableThis table shows the daily count of conversations that were fully automated by each of your assistants. 
  • Best practices for WhatsApp and SMS assistants

    With Customer Assist, you can create conversational assistants for your WhatsApp and SMS channels to help you collect information from customers before they connect with an agent to speed up response times and close conversations more quickly. This article provides tips to consider when creating menus for these assistants. See Introduction to Customer Assist to learn more about this feature.In this articleGetting startedFormat your interactions in WhatsAppCreate menus in WhatsApp and SMSExamples of menusAdditional tipsGetting startedHere are some basic tips to keep in mind before you create an assistant:Start each conversation with a menu to help guide users through the experience.Highlight the most common trigger for each option.Focus on one-touch and low-complexity resolutions.Use number emojis for menu options.Save the customer's response at a conversation level. We recommend creating an attribute called CA - Step Value - (Single line text type).Format your interactions in WhatsAppYou can format the text in your interactions to help highlight content for your users.FormattingActionExampleItalicPlace an underscore on either side of your text._text_BoldPlace an asterisk on either side of your text.*text*StrikethroughPlace a tilde on either side of your text.~text~MonospacePlace three backticks on either side of your text.```text```Create menus in WhatsApp and SMSMenus help guide your users through the conversational path. We recommended using a Conditional Branch interaction that evaluates the values for the CA - Step Value. Keep the following in mind when creating menus for these assistants:Use Match Any if you will evaluate more than one option in that interaction.Use the Is Equal To operator to evaluate a number or emoji option.Use the Contains operator to evaluate your keyword(s).Redirect customers to another dialog if they enter an unexpected value that doesn't match your criteria.We recommend not using emojis for buttons if using the Navigation buttons interaction in a WhatsApp assistant.Examples of menusThe following section shows examples of menus you can build for WhatsApp and SMS assistants. In all of these examples, we'll use the Customer Text Response interaction to format the menu the customer will see, followed by a Conditional Branch interaction that evaluates the customer's response and redirects them to the appropriate dialog. For more information on interactions, see this article.Starting menuStart the conversation by presenting a menu with the most likely reasons the customer may be contacting you.Example Customer Text Response interactionExample Conditional Branch interactionFeedback menuCreate a menu that asks the customer to provide feedback.Example Customer Text Response interactionExample Conditional Branch interactionAnother question menuCreate a menu that asks the customer if they have another question or want to end the conversation.Example Customer Text Response interactionExample Conditional Branch interactionAdditional tipsUse AI to respond to customers with information in your public knowledge base.Remember to use the business schedules interaction to change the message a customer sees if they contact you inside or outside business hours. You can also take different actions based on this conditional.You can use volume control with WhatsApp and SMS to automatically end conversations if a customer stops responding for a certain amount of time. You can also choose to mark these inactive conversations as Done automatically.Use Intent detection to understand user queries and jump from one dialog to another based on the end-user's keywords.Set up assistant rules to specify which dialog your assistant launches when a customer starts a conversation based on different factors.
  • Best practices for conversational assistants

    With Customer Assist, you can create conversational assistants that help optimize your organization's support experience, promote customer self-service, boost agent productivity, and speed up response times. This article provides tips for setting up these assistants. See Introduction to Customer Assist to learn more about this feature.Note: While we recommend these best practices, your legal team is the best resource to give you compliance advice for your specific situation.In this articleDisclose the assistant is not a personLet's talk about strategyTips when creating a menuRouting conversations and handover expectationsManage your dialogsTest, iterate, and deployReporting on the assistantConversation design and brandTips for SMS and WhatsApp assistantsDisclose the assistant is not a personYour assistant is an important part of the support you provide for your customers. When required by applicable law and to provide transparency to customers, it should be made clear that they aren't interacting with a human and that the person is interacting with a bot. Generally, we can call them “Brand Bot” to clarify that they are chatbots and express that they interact with an automated system. It should still feel like they are having a friendly conversation with a machine, but it must be made clear to customers that they are communicating with a bot:at the beginning of any conversation or message thread,after a significant lapse of time, orwhen a chat moves from human interaction to automated experience.Acceptable disclosures include but are not limited to: “Hello! Welcome to [Company name]. I am a bot and am not a real person, but I’ll do my best to answer your questions.”“Hello! I’m the [Company Name] bot,”“Hello! You are interacting with a bot,” “Hello! You are talking to a bot,” Note: Even where not legally required, we recommend informing users when they’re interacting with a bot as best practice, as this helps manage user expectations about their interaction with your messaging experience.Disclose assistant messages are not sent by a personBy default, every message the assistant sends in chat will display "Bot" to indicate to customers that they are not speaking to a person. Change this default setting in your Chat settings.Consent to record communicationsAs required by applicable law, you should obtain customers’ consent to record conversations and provide transparency to customers that you use Kustomer, Inc. to facilitate this chat function. Kustomer recommends you do this in your welcome message. An example message is: “By using this chat, you agree that your communications can be recorded. For more information, please refer to [company’s] privacy policy.”By default, your chat widget will show a short message in the footer notifying the customer that they are consenting to the conversation being recorded. You can customize this consent message in the Design tab of your Chat settings.Let's talk about strategyNow that you're ready to create a conversational assistant, you may wonder where to even begin. Below is a list of tips to remember when planning your assistant.Start Small. Start Today. There's no need to wait to build an advanced assistant or spend months planning the project.Think Big. Learn Fast. Constant iteration is the key to success. Iterating can make a massive difference in CSAT and automation rates.Solve your most common use cases. Analyze your contact reason data to help you determine where to start. You can view examples of common reasons to set up an assistant here.Start with one use case at a time. For example, you can start by creating an assistant that provides answers for one topic and hands over additional queries to your team.Reduce barriers to chat with your team. You don't need to try to answer 100% of chat questions. Instead, redirect to agents when needed. Avoid telling the customer to repeat their question.An assistant should reduce friction to resolution, not increase it.  Don’t just ask to ask - know how you are going to use the data.Keep it simple and effective. Do not loop your customers through too many questions or interactions.Tips when creating menusUsing menus in your assistant is a great way to help guide your customers to the best resolution. Start with buttons. Preferably, implement multi-level navigation with 1-3 levels of categories and topics.Create the menu from the customer's point of view. For example, labeling a button “My order is late” is better than labeling it “Shipping”.In your welcome menu, have a fallback option where you prompt the customer with another, more general question (for example, “Something Else”).Routing conversations and handover expectationsWhen it comes to handing a conversation over to an agent, make sure you clearly define timing expectations on whether the conversation is taking place during business hours or not, and whether you are going to get back to the customer via chat or email. Configure the Agent Missed Conversation volume control setting to provide alternative channels when your team can't respond in time. We recommend 5-15min for chat conversations. If you plan to have your chat respond outside business hours or you can’t respond promptly, you can hand over all your chat conversations to your email queue. Use the Update Data interaction and queue rules to route conversations to different teams. You can also use the Update Data interaction to save any helpful conversation attributes for your agents or help with routing.Additional tips:You can create notes for your agents when they receive the assistant conversation using callable workflows before handover.Anonymous chats will prompt agents to verify and merge profiles based on the provided email or phone number. Make sure you verify the information before merging.Manage your dialogsDialogs are a distinct unit of conversational exchange between the assistant and your customer. We recommend making small dialogs that belong to the same topic and that you keep them simple.Tip: A good way to think about when to create new dialogs: if you think a customer can ask a question about the topic at any point of the conversation, use a new dialog.You can use assistant rules to redirect customers to specific dialogs depending on the button they click or the current page they are on. This gives customers an incredible first impression, and one step closer to proactive support. You can also use proactive chat with assistant rules to engage with your customers at the right time.Support multi-language interactions by using snippets in all messages inside the conversational assistant.Ask for feedback after a final answer is given, usually with buttons such as “Helpful” and “I still need help”.It is generally good practice to end the conversation after a customer selects “Helpful” as a final answer.Using callable workflows with input parameters such as your customer, order, or conversation ID allows you to perform any action on behalf of the customer. You can also return information using output parameters to be displayed in a message using the Workflow: Send Response action event.Test, iterate, and deployIteration is key to your success. Over time, add new answers to your chatbot and measure its performance.Have multiple testers, and don’t have the person updating the assistant also be a tester. Having a pair of fresh eyes test the flows will help find inconsistencies.Use assistant rules to test specific start dialogs or assistants for specific customers. You can also use them to personally test an assistant using a specific URL or phone number.You can also upload the chat code snippet in codepen.com to test your assistant.Reporting on the assistantThe conversational assistant integrates seamlessly with our reporting features. Here are a few tips you can implement to better understand how your assistant is performing:Evaluate your assistant’s performance with our standard Assistant report or create custom reports that filter by the Assistant attribute. When creating custom reports on your contact reason or any other saved attribute, we recommend combining it with a specific Assistant or Assistant Status conversation attribute.Turn CSAT on for your assistant, and create custom reports by contact reason to understand where you need to improve your chatbot.You can also use the Update Data interaction to add tags or save information you can later report on. If you use this interaction with your contact reason attribute in different dialogs, you can keep your Assistant conversations categorized in the same way as your email or other channels.Conversation design and brandYour assistant is an important part of the support you provide for your customers. While they aren't interacting with a human, it should still feel like they are having a friendly conversation. Your assistant is another member of the team and should have a personality of its own. What is their name? How do they represent the brand? Develop a simple personality and be consistent with it in every message.Note: When developing a personality for your assistant, do not pretend the assistant is a human and use a human profile image as the avatar. When setting up your dialogs, use "we" instead of "I" and address the user by their first name, if the end user is authenticated. Set the expectations for the conversation flow early on. Clearly state what the assistant is designed to do to help people make the most out of the experience.Avoid conversation dead ends. When you reach the end of a conversation tree, try to still offer options. For example, you can move the conversation to another matter, return the customer to the menu, or just say bye if it was a final answer and end the conversation. Never block customers in a loop by repeating the same question. Likewise, clearly state when the conversation is handed over to an agent.Consider using emojis to express emotions in greetings, buttons, and menus. A first approach to the most simple emotions could be:Doubt: 😶😐😑🙄🤔🤐Negativity: 😱😣😕☹️😵🤧😭😔😣😓😤👎💩🤦Positivity: 😀😄👌😍😉😂🤗😎😋😇👍️😙😃👌✌️😁Don't ignore the assistant's style and personality when communicating errors! Messages such as, "Server error" or "Something went wrong, try again later" can spoil the chat experience and make the assistant look like an unreliable machine. Make sure you maintain the assistant's personality in all scenarios.Be concise, the less text the better. Check punctuation and try to use correct and natural expressions. Always double-check your spelling. Phrases are better than paragraphs and make sure you watch your word count.Tips for SMS, WhatsApp, and Kustomer Voice assistantsSee Best practices for WhatsApp and SMS assistants for more tips on creating menus for these channels. For tips on setting up an IVR, see this article.
  • Best practices for creating intents

    You can create intents that use common phrases your customers use when submitting an inquiry. These phrases are words, or groups of words, that a conversational assistant listens to during a chat conversation. Using natural language processing, the assistant uses these phrases to identify your customer's needs and redirect them to a specific dialog so that they can get the information that they need. This article provides tips to take into consideration when creating phrases for your intent detection. In this articleHave distinct examples across intentsUse a wide vocabularyMake sure the assistant has the context it needsHave distinct examples across intentsA common mistake when creating intents is to organize them into categories that are too granular. It's better to keep your intents more general to ensure variations are distinct. For example, let's say you've noticed that many customers contact you to inquire about topics such as, help with their password or email address. Instead of creating a different intent for each category and naming them Change password or Change address, you can instead create a more general My account category and save a variation of phrases for all of these inquiries.Example variations of phrases that you can save under a general My account category are:I need to change my passwordI want to change my e-mailWhat is my password?Can you  reset the password?How do I recover my password?How do I update my account?Upgrade my accountHow can I edit my personal information?I cannot login into my accountI forgot my passwordThese phrases provide a variety of ways to ask for information that can be grouped under account inquiries. When coming up with phrases for your intents, you should make sure that all of your intents are clearly defined and that a phrase can't be assigned to more than one intent category.Use a wide vocabularyA common mistake users make when creating intents is prioritizing the quantity of phrase examples, over the quality of the phrases themselves. It's much more important to make sure that the variations of your phrases cover a wide range of vocabulary and wording to ensure that the assistant can detect the various ways a customer might ask for the same thing. For example, let's say you want to create an intent that detects when a customer is inquiring about shipping costs. Example variations of phrases that you can save to capture the different ways a customer might ask for shipping cost information are: How much does shipping usually cost? How do you negotiate shipping costs? What is the cost of shipping and handling? How much does it cost to ship a package via USPS? USPS shipping cost per pound? What is the cheapest way to get the package delivered? How do I choose a shipping strategy for an online store? How much am I charged for shipping and handling? How do you calculate shipping costs? Can I reduce my shipping costs?Note: We recommend you review your real user data to make sure your variations reflect the way your customers talk.Make sure the assistant has the context it needsThe last tip to remember when creating an intent is that the assistant only understands what is said in the message and can't successfully understand a customer's intent if it needs to rely on external information to provide all of the context it needs. For example, let's say you are considering creating intents for Order delayed or Order on time because you've noticed customers regularly contact your organization for either of these reasons. If an agent is reviewing this inquiry, they are going to look at a variety of factors such as, whether the order is still being processed or tracking details to determine if the order is on time or delayed. Once they get this information, only then will they tag the conversation as Order delayed. The assistant, on the other hand, does not know the status from the original inquiry, even if the inquiry is "Is my order late?" In this scenario, creating an intent that directs the customers to a dialog apologizing for the late delivery could possibly be incorrect since more context is needed before making that conclusion. You should avoid creating intents that rely on additional information to correctly identify the customer's needs.
  • Migrate conversational assistants across instances

    If you have a conversational assistant set up in your sandbox or another Kustomer instance, you may find the need to copy the assistant across different organizations without having to recreate the entire conversational flow.You can use Postman to simplify this process.Who can access this feature?User typesAdmins and users with access can create conversational assistants and API keys. Users with access can interact with the Kustomer API using Postman.In this article:PrerequisitesCreate a conversational assistant in the new organizationCopy the assistant's JSON from the existing organizationUpdate the assistant ID on the JSONSend a PUT request to the assistant in the new organizationPrerequisitesBefore following this guide, you must sign up for a Postman account and set up your workspace: Setting up Postman.You will also need a user account with access to create and update conversational assistants in the organizations where you are replicating the assistant.Create a conversational assistant in the new organizationGo to Settings> Kustomer IQ > Customer Assist.Select Add Assistant and then choose Lite or Premium depending on the type of assistant you need to copy over.Select Continue.Enter a name for the assistant and select the Channel from the drop-down menu.Next, select Add Assistant. The new assistant will appear with the standard preconfigured flow.Go back to the Customer Assist page and copy the public ID of the assistant you just created.7. Open a new tab in the same browser where Kustomer is open and go to https://<your-new-org-name>.api.kustomerapp.com/v1/assistants?publicId=yourAssistantPublicID where:your-new-org-name is the company name of the Kustomer instance you created the new assistant in yourAssistantPublicID is the public ID you copied in the previous step The URI should return the assistant's data inside an array as seen in the example below. Copy the Assistant ID (which is different from the assistant publicId) and save it in a separate file for later use.Copy the old assistant's JSON from the existing organizationIn the organization that contains the existing conversational assistant you want to copy, go to Settings> Kustomer IQ > Customer Assist.Copythe public ID of the assistant that will be copied.3. On the same browser, open a new tab and go to https://<your-existing-org-name>.api.kustomerapp.com/v1/assistants?publicId=yourAssistantPublicID and replace yourAssistantPublicID with the Assistant Public ID you just copied. The org name is the company name of the Kustomer instance you have the assistant you'd like to duplicate, and the URL will return that assistant's data inside an array.4. Copy the assistant's data, starting from start, dialogs, and all the way to the end of settings.Update the assistant ID on the JSONIn the document where you pasted the assistant's data, find the assistantId attribute inside each node of the dialogs and replace that ID with the new Assistant ID (note: not the new Assistant public ID) which you saved previously when creating the new assistant.If the conversational assistant you are migrating has a deflection interaction, find the Knowledgebase ID in the new instance that you would like to use for the deflection. You can find that ID at this endpoint https://<your-new-org-name>.api.kustomerapp.com/v1/kb/knowledge-bases of your new instance.Send a PUT request to the assistant in the new organizationIn the org you are migrating the Assistant to, go to Settings > Security > API Keys and create an API key with org.admin.apps.write, org.permission.assistant.update, and org.permission.apps.update roles.Copy and paste the token that appears in a separate file for you to use in Postman.Open Postman and go to your Kustomer API workspace. Select the plus + icon in the toolbar to create a new request.In the Headers tab underneath the request URL bar, fill in the following key values:Content-Type: application/jsonAuthorization: Bearer API_KEY (the space is required after the word 'Bearer'). The API_KEY is the API key you just created.In the Body tab, create a PUT request to the endpoint https://<your-new-org-name>.api.kustomerapp.com/v1/assistants/yourNewAssistantIDwhere:your-new-org-name is the company name of the Kustomer instance you created the new assistant in yourNewAssistantID is the Assistant ID of the new assistant and the same ID you have been adding to the old assistant's dialogs.Copy and paste the assistant data you modified in your document, starting from start and all the way to the end of settings, to the Body section in Postman.Send the request and see that it is successfully sent.When you refresh your assistant in the new org, you will now see all the dialogs and interactions migrated over from your existing assistant.
  • About Assistant Pro (legacy assistant)

    Conversational assistants help you optimize your organization's support experience, promote customer self-service, boost agent productivity, and speed up response times. Assistant Pro gives you all of the functionality found in Assistant Lite, plus more advanced automation capabilities.Note: Assistant Pro is a legacy feature and is no longer available to new customers as of January 2024. Assistant Pro does not include any AI features. For information on current assistant types, see About Lite and Premium assistants.In this article:What is Assistant Pro?What interactions are available in Assistant Pro?Volume control settings available in Assistant ProConvert your assistant to LiteWhat is Assistant Pro?Assistant Pro is a legacy assistant that is paid. It includes all of the features and interactions available in Assistant Lite while offering more advanced automation capabilities. Assistant Pro does not include AI capabilities. For AI capabilities, you must upgrade to Premium assistants with our new KIQ Conversations progressive pricing.Note: If you have questions about your current KIQ Conversations with Assistant Pro plan, please reach out to support@kustomer.comWhat interactions are available in Assistant Pro?Interactions are used to gather information about your customer and guide them through the assistant's conversation path. This table shows the interactions that are available for Assistant Pro. Please note that the channel type mainly determines available interactions. For example, if you create an SMS assistant that does not support the Button Question interaction, it will not be available regardless of the assistant type.InteractionProSend a MessageXCustomer text responseXButton questionXData attribute questionXConditional branchXBusiness scheduleXObject SelectorXRun WorkflowXUpdate DataXRequest feedbackXAI responsen/aNavigation buttonsXArticle deflectionXRedirect to dialogXTransfer to an agentXEnd conversationXVolume control settings available in Assistant ProYou can use volume control settings to automate how the assistant responds to customer-abandoned conversations, follows up with customers if an agent hasn’t been assigned to the conversation, or responds in the chat, after a set amount of time. The following settings are available for Assistant Pro:Inactive customer reminderCustomer-abandoned conversation Transfer conversations to an agent Mark conversations as done Agent missed conversation With Assistant Pro, you can also choose to mark a conversation as done if the customer doesn't select a follow-up channel within the time specified.Convert your assistant to LiteThere may be times when you create an assistant and decide afterward that you want to create a similar, yet different type, of assistant. You can convert your assistant type by going to Settings> Kustomer IQ > Conversational Assistant. Your assistant will be duplicated as the other type. Note: When duplicating an Assistant Pro, the following features and interactions will no longer be available in the newly created Assistant Lite version:IntentsObject Selector interactionRun Workflow interactionArticle Deflection interactionEnd Conversation interactionMark as done in the Volume Control Abandoned settingsWe recommend you review the assistant carefully after duplicating it to Lite and making any necessary changes to ensure the assistant follows the intended conversation path.
  • Create custom intents

    Kustomer provides predefined intents so that you can quickly set up your conversational assistant to understand customers' needs and get them the information they need quicker. Your organization may have additional use cases that require you to create unique intents to better detect customer queries. Using both keywords and phrases, you can create up to 80 custom intents. Who can access this feature?User typesOnly users with access to Intents can create a new intent for use with an assistant. For more information, see Define user access.Plan availability✔  Professional✔  Business✔  Enterprise✔  UltimateIn this article, we will show you how a cookie store can set up an intent that redirects customers to an agent when they inquire about the available flavors.In this articleCreate keyword intentsDetect common phrasesUnderstand the intentRedirect the customer if the intent is detectedCreate keyword intentsYou can create intents that contain the words, or groups of words, you want the assistant to listen for during the conversation. We recommend using keywords when you know where the customer should be directed once their intent is detected. For example, a customer entering "help" should always match a help intent.Go to Settings> Kustomer IQ > Conversational Assistants.Go to the Intents tab and then select Add Intent.Enter an intent name and add the keywords you want to train the conversational assistant to detect. Since we are creating an intent that detects when a customer is inquiring about the cookie flavors available, suggested keywords for this can be: flavor, taste, sweet, salty, light, and ingredient. Please keep the following in mind when creating keywords:You can enter up to 15 keywords per language. You can activate additional languages in your Kustomer organization on the Localization page. For more information, see Customize localization settings.A keyword must have a minimum of 2 characters and a maximum of 35.An intent can only be used once in an assistant.A keyword can only be used in one intent.If entering a keyword with multiple words, the words must be used in that exact sequence for the assistant to detect it successfully. For example, "credit card" will not be detected if the customer enters "can I please get a credit on my card?"Select Save and Train Intent.Note: You can also create an intent directly within a dialog. If you do so, select Add Interaction to specify where the conversation directs the customer if one of the keywords is detected. For more information on what's available, see Conversational assistant interactions.Detect common phrasesYou can also create intents that detect customers' common phrases in their inquiries. Phrases complement keywords and help create stronger intents by training the assistant to understand how a customer might word an inquiry differently. When creating phrases, the more example phrases you have, the better the training will be. We require you to create at least 5 phrases and recommend creating at least 12 different variations of phrases to ensure you have enough variety to train the assistant properly. See Best practices for creating intents for tips on creating phrases for intent detection.Notes: English is currently the only language supported.From the Intents tab, select Phrases.Select Add Phrase and enter the phrases you want the conversational assistant to detect. For our example, we want to use phrases a customer would use when asking about cookie flavors. Suggested phrases could be:What flavors do you have?What do the cookies taste like?Do you have anything with almonds?Do you have light cookies?How many types of cookies are there?Select Save and Train Intent.Once you save your changes, training the assistant on your new phrases may take some time. You can check the training status by going to the Phrases tab for the intent.Understand the intentOnce you create the intent, you must add an interaction in the conversation path that will "listen" for the trained keywords. We recommend adding this step early on in your path.Note: The ability to understand intent is only available for the following interactions:AI response (See this article for more information on using AI to respond to customer inquiries.)Customer text responseArticle DeflectionTo add an interaction that detects intent:Open the assistant you want to use the intent with or create a new one.Since we are creating a conversation path to detect if a customer is inquiring about cookie flavors, let's add an interaction that prompts them to enter a few words describing why they are contacting you.In the Welcome dialog, select Add Interaction > Text question.Enter a message to prompt the customer to respond. In this example, we ask "What can we help you with today?"Turn on Understand Intents. The AI icon will appear, indicating that this interaction will help you understand your customer's input.Redirect the customer if the intent is detectedThe final step is to create the dialog the assistant uses to continue the conversation based on the keyword it understood from your saved intents.Select Add Dialog and then choose the desired intent from the drop-down menu. Select the dialog the assistant redirects the customer to if the intent is detected. In this example, we will transfer them to the next available agent.Note: You do not need to have this dialog linked to another one since it will continue the conversation only if the intent is detected.Now, if a customer uses one of the return keywords in the conversation, the assistant will automatically detect it and redirect it to an agent. You can easily view which dialogs are listening for intent by looking for the AI icon, which indicates that an intent redirects the conversation to it. You can hover over the icon to see exactly which intent is used.
  • Add a conversational assistant

    Conversational assistants can help you collect customer information before they connect with an agent, speeding up response times and closing conversations more quickly.Note: Only users with access to the Customer Assist page can create a new assistant. For more information, see Define user access.In this articleAdd a new assistantPreview an assistantTurn on a conversational assistantAdd a new assistantTo add a new assistant, go to Settings > KIQ > Customer Assist and select Add Assistant. Here, you can choose to create a Lite or Premium assistant. When you create a new assistant, you’ll need to provide the following:Internal Name is the name listed for the assistant on the Customer Assist page in your Kustomer organization.Description is a brief description of the assistant that appears in your Kustomer organization.Public Name is the display name of your conversational assistant that will appear for customers in the chat widget.Channel is the channel that the assistant will be on. Available options are Chat, SMS, and WhatsApp.In the assistant editor, you can update the Internal Name, Description, and Public Name under Assistant Settings.After you create an assistant, the editor page for the assistant opens. Here, you can add dialogs and interactions to build the different conversation paths the assistant will take. Once you are done, select Save and Publish and then turn on the assistant to start using it in Kustomer.For a step-by-step conversational assistant tutorial, see Tutorial: Build a conversational assistant.Note: You can create and manage multiple conversational assistants (chatbots) and assign them to multiple brands under our organization with Multi-brand Chat features.Preview an assistantYou can preview your conversational assistant and interact with it to test it before making it available. To do so, select Preview in the editor. You will be directed to a new page. Here, you can interact with your assistant as if you were a customer and test the flow you set up. You can also select Advanced settings and test other features, such as browser notifications, push notifications, browser language, and icon position and size.Turn on a conversational assistantSelect Save and Publish in your conversational assistant draft to make your assistant active in Kustomer. In Chat Management: Settings, you can assign a default conversational assistant for each brand from a list of your active assistants. Note: When you turn on an assistant for a brand, the assistant will start each newly created conversation with your customer unless you override the settings in the Chat SDK.
  • Common use cases for a conversational assistant

    A conversational assistant can handle many types of scenarios. Below is a list of the most common use cases for the assistant.Note: When setting up the assistant for your use case, remember that you can provide a better experience for customers and agents by authenticating your customers using JWT for chat when they are signed in to your website or app.Use case #1: Routing conversations to the right agentSet up your assistant to gather the following information before routing:Prompt users with a “How can we help?” question first, and then ask for their email or name (if needed) after.Create a welcome menu with some categorization to route to the right agents, and then ask the same questions as above.If you have very specific menu options, you can skip the “How can we help?” question and only gather the data that your agents usually request, such as the order number. Here's an example of an interaction prompting the user to enter their email address.Use case #2: Use a menu to help guide customers toward an answerInstead of asking customers open-ended questions that may result in a longer experience, use menus to present options for your customers to choose from and guide them to the best resolution.Expand your menu with many more contact reasons using a multi-level list attribute. You will need to use a conditional for the lowest menu options.Use the Update Data interaction in each menu option to save the contact reason you use for your other channels.For some menu options, start by providing short final answers to the high-volume, low-integration ones. Then, hand it over to your team.Short answers are better than providing a longer knowledge base article.By working off the menu options you can hand over the conversation to different teams based on the menu topic chosen.Use feedback buttons for all final answers such as “Helpful” and “I still need help”. We recommend ending the conversation with the “Helpful” button.Use case #3: Let AI help your customers using information found in your knowledge baseCreate a public knowledge base and use the AI response interaction to create an assistant that responds to customer inquiries using your information.Use case #4: Help your customers by using existing static content:Integrate with your knowledge base for free-text questions by adding an Article Deflection interaction in your "Something Else" menu fallback button.Create intents for your high-volume questions and provide step-by-step instructions on how to solve them. If you also have a knowledge base, turn on Intents in the Article Deflection interaction for maximum self-service. With this feature, the engine will match intents first, then suggest articles, and finally hand over the conversation to an agent as a fallback.Use case #5: Automate entire processes by performing actions and integrating with your CRM and systems:Use the Object Selector interaction to resolve common issues, such as where is my order, for authenticated users.Automate any process integrating with 3rd party APIs or your CRM information with callable workflows.
Powered by Kustomer