Internal Actions

This action assigns specific roles to the currently authenticated user in your bot flow. You can provide roles in various formats, offering flexibility for different use cases:

• String or List of Strings: Directly list the role names you want to assign (e.g., ["admin", "editor"]).
• Dynamic Roles from Set Variables: Define roles dynamically using the Set Variables section and the Code Extension feature. This allows you to assign roles based on conditions or external data.
• Other String Type Expressions: Any expression that evaluates to a string containing a valid role name can be used.

string / List<string> / [[Entity]].EntityField / @Variable

Parameters

This internal action accepts as parameter the role names you want to assign. The format depends on the chosen input method:

• String: A single role name (e.g., "Helpdesk Agent").



• List of Strings: An array containing multiple role names (e.g., ["Employee","Manager"]).



• Dynamic roles specified in [[Role]].Name set in the Set Variables section. You can use any entity list that stores roles, an entity that references the [[Role]] system entity.

Example: Imagine you have a field named [[ChatUser]].AddRoles, which is an entity list referencing the system entity [[Role]]. You can dynamically assign roles based on user data using the Set Variables section.

This internal action prevents a user from being placed in the waiting queue. You should use this action in the Live Chat agent disconnected special flow. Use this action to manage user access effectively.

{
    "BlockIntervalMinutes": <integer>,
    "BlockReason": "<string>",
    "SaveToChatActivityData": <boolean>
}

Parameters

Example: This example would block the user for 60 minutes, log the reason "Abusive behavior," and record the block event in the chat activity data.

{
  "BlockIntervalMinutes": 60,
  "BlockReason": "Abusive behavior",
  "SaveToChatActivityData": true
}

For additional information on how to block or unblock chat users, see Block Chat Users.

This internal action allows bot authors to provide helpdesk agents with the ability to cancel their break requests, enhancing agent flexibility and improve overall chat support efficiency.

The action is used in the CancelHelpdeskAgentBreakRequest flow, part of the Live Chat Advanced solution template available for download from the Solution Library.

If the agent has requested a break while having ongoing conversations and decides to cancel the break, the agent can do so by clicking the Cancel button.

This internal action allows bot authors to provide helpdesk agents with the ability to log out of Live Chat.

The action is utilized in the CancelHelpdeskAgentLogOutRequest flow, part of the Live Chat Advanced solution template available for download from the Solution Library.

If an agent clicks to log out while having ongoing conversations or when engaged in a conversation and automatic allocation is active, a log out request is initiated and the agent will be logged out after all live sessions have ended.

This internal action switches the current conversation to a new one by changing the ConversationId. Depending on the parameter value, the context can be retained (either partially or fully) or dropped entirely.

Parameters

The action supports an integer for context retention that can have one of the following values:

• 0: No context is saved, including authentication status. In this case, [[ChatUser]].IsAuthenticated becomes false.
• 1: No context is saved, except for the authentication status. [[ChatUser]].IsAuthenticated retains its previous value.
• 2: All context is kept, meaning the new conversation will have the previous context.

This internal action changes the queue that the helpdesk client is awaiting on to the one of the current agent.

The action is intended for internal use in integrations with Mediatel.

This internal action closes an open activity. By default, an activity (indicated by [[ChatUser]].ActivityId) automatically closes after 1 hour of inactivity. Use this action to close an open activity before the 1-hour threshold.

No parameters are required.

The ConnectAsHelpdeskAgent internal action connects a user with the Helpdesk Agent role to the live chat agent queue, providing a list of helpdesk clients currently waiting. This action ensures proper routing and client assignment in a live chat environment.

This action applies to single Live Chat / route-to-human.

This action is crucial for effective agent-to-client assignment in live chat sessions.

Parameters

No parameters are required.

The internal action returns the list of clients waiting for helpdesk support in [[HelpdeskClientList]].

This internal action places the user in the waiting queue for an agent and provides status updates regarding agent availability and working hours. This action is particularly useful for handling requests outside of standard working hours.

Scope: This action applies to Live Chat / route-to-human.

Key Details

• Queue placement: Adds the user to the agent waiting list.
• Working hours status: Sets [[HelpdeskClient]].IsOutsideWorkingHours to true if the request is outside working hours.
• Working hours information:
• Populates [[HelpdeskClient]].WorkingHoursBeginAt and [[HelpdeskClient]].WorkingHoursEndAt with the daily working hours.
• If agents are outside their schedule, the bot can share working hours with the user for clarity.

Behavior with Multiple Time Slots

When multiple working slots exist in a day, the next available slot will be provided to the user.

If the day’s working hours are continuous, the full range will be shown.

Example: If working hours are 8:00–12:00 and 13:00–18:00, and a user attempts to connect at 19:00, the bot will notify them of the day’s full working hours range (08:00 to 18:00), excluding the lunch break.

{
    "QueueCode":"string"/[[<Entity>]].<QueueCodefield>/@queue_code,
    "PriorityLevel":"@priority",
    "AutomaticAllocation": true,
    "SendConversationTranscript": true,
    "ContactCenterParameters":{},
    "ForBotResponseValidation": false,
    "ResponseEntity":"[[Entity]].AgentResponse",
    "InfoToValidate": {
        "UserQuestion":"<[[Entity]].UserQuestion/@UserQuestion>",
        "BotResponse":"<[[Entity]].BotResponse/@BotResponse >",
        "MaxWaitTimeForAgentResponse": 30
    }
}

Parameters

This internal action connects the agent with the specified user.

This action automatically leverages existing contexts from the [[HelpdeskClient]] and [[ChatUser]] entities to initiate a live chat session.

This internal action connects the agent to a specified user in a multichat setup.

This action automatically leverages existing contexts from the [[HelpdeskClient]] and [[ChatUser]] entities to initiate a live chat session.

No parameters are required.

On web channels (direct line) only!

You can also use this internal action to send a backchannel event — such as triggering a blink or opening a pop-up — along with the message "ConnectedToHelpdeskAgent", notifying the client application that the agent has connected successfully.

{
    "EventName":"string",
    "PayloadEntity":"string"
}

Example: Defining a backchannel event with an optional payload entity:

• EventName is set to OpenChatPopup, triggering a pop-up in the client application.
• PayloadEntity is ChatSessionData, sending additional serialized entity data.

{
    "EventName": "OpenChatPopup",
    "PayloadEntity": "ChatSessionData"
}

The ConnectToPastConversation action allows agents to reconnect with users in past conversations across all supported channels, including WhatsApp, Facebook, Telegram, Viber, Google Chat, Instagram, Slack, and Twilio SMS, provided the following conditions are met:

• The user is not currently engaged in an active conversation with the bot.
• The time since the last user interaction with the bot is within the allowed limit for the specific channel. For instance, agents can reconnect to past live chat conversations within a 24-hour window from the last interaction, depending on channel restrictions.

For more information, see Multichat for Helpdesk Agents.

Determines the language of the intent provided in parameters block and returns the result in @DetectedLanguage.

{
 "Intent":"@phrase/[[Entity]].Intent"
}

This internal action disconnects the user from the agents queue. It applies to single Live Chat / route-to-human.

No parameters are required.

This internal action ends a break for the current help desk agent.

No parameters are required.

The action is utilized in the EndHelpdeskAgentBreak flow, part of the Live Chat Advanced solution template available for download from the Solution Library.

While in break, a pop-up displays on screen with a break timer (showing the time consumed in break). If the agent clicks the I’m back button, the break ends and the agent becomes available again.

This internal action ends an active live conversation and logs the chat activity in the system for accurate record-keeping.

No parameters are required.

This internal action identifies and extracts text related to non-deterministic entities, such as Boolean, Date, DateTime, Decimal, Email, Integer, Location, Organization, Percentage, and Person. This action utilizes the same ML service function as intent recognition, but unlike intent recognition, the extracted text does not depend on a specific training phrase pattern.

Use this action to detect and retrieve sensitive data from user input, such as personally identifiable information (PII) like addresses, emails, or names. Extracting these entities allows you to safeguard user privacy by either removing or anonymizing sensitive information from [[Intent]].Text before any further processing. This is especially important before transmitting data to external systems like ChatGPT or other external services.

This internal action allows you to filter an entity list based on specified conditions. You can apply multiple conditions to narrow down the list as needed, using accepted operators such as Equality, Inequality, LessThan, LessThanOrEqual, GreaterThan, and GreaterThanOrEqual.

Parameters

To filter the list, provide a JSON structure that specifies the entity list name and the conditions to apply. Each condition includes:

• Expression: The entity field to evaluate, given as "string/[[Entity]].Field".
• Operator: The operator to apply (e.g., GreaterThanOrEqual).
• ValueExpression: The value to compare against the entity field.

{
    "EntityListName": "EntityList",
    "Conditions": [
        {
            "Expression": "[[Entity]].Field",
            "Operator": "OperatorType",
            "ValueExpression": "value"
        },
        {
            "Expression": "[[Entity]].Field",
            "Operator": "OperatorType",
            "ValueExpression": "value"
        }
    ]
}

Example: Filter the EmployeeList to include only employees aged between 18 and 99

{
    "EntityListName": "EmployeeList",
    "Conditions": [
        {
            "Expression": "[[Employee]].Age",
            "Operator": "GreaterThanOrEqual",
            "ValueExpression": "18"
        },
        {
            "Expression": "[[Employee]].Age",
            "Operator": "LessThanOrEqual",
            "ValueExpression": "99"
        }
    ]
}

This internal action generates a new 4-digit random authentication code and stores it in [[ChatUser]].AuthenticationCode.

No parameters are required.

Use this action for multi-factor authentication, allowing you to send a code as a second factor for chat user authentication.

This internal action loads the [[AlertInfo]] system entity, which is used in the "Alerts" presentation flow. This action retrieves and prepares alert information for display within the flow, ensuring the latest alerts are available for users.

This internal action saves the conversation transcript in [[ConversationInfo]].TranscriptTxt.

This action allows bot authors to save and send conversation transcripts to external systems through DRUID connectors (e.g., email, CRM). For setup information, refer to Sending the conversation transcript to external systems.

Parameters

{
    "ActivityId":"[[ChatUser]]. ActivityId",
    "ConversationId":"[[ConversationInfo]].ConversationId",
    "StartDate":"YYYY-MM-DD",
    "EndDate":"YYYY-MM-DDThh:mm:ssTZD",
    "TimeoutMilliseconds":3000
}

You can filter a conversation transcript by the following parameters:

This internal action retrieves the intent specified by [[Intent]].Text. It captures the exact intent communicated by the user in the chat and allows you to manually trigger a prediction using NLU (Natural Language Understanding) or Knowledge Base (KB) models.

Parameters

{
    "Model": "string", 
    "LanguageCode": "string", 
    "KBDataSourceId":"string",
    "BotId":"string",
    "CorrelationId":"string",
    "IsChildIntent":"string",
    "ParentFlowId":"string",
    "UrlMetadataFilter":"string",
    "KBIncludeExcludedNodes":"string"
}

{
"Model": "KB", 
"LanguageCode": "en-US", 
"KBDataSourceId":"79adf7d0-1894-4deb-bdff-f6d5fbad31f8"
}

{
    "Model":"KB",
    "UrlMetadataFilter":"*useful-information*"
} 

{
    "Model":"KB",
    "UrlMetadataFilter":"<root Folder>/<SubFolders>/file_name.xlsx"
}

{
    "Model":"KB",
    "UrlMetadataFilter":"*file_name.xlsx*"
}

{
    "Model":"KB",
    "UrlMetadataFilter":"subFolderName*"
}

This internal action retrieves the intent specified by [[Intent]].Text considering the Megabot feature. Use it to capture the exact intent communicated to the user in the chat and manually execute a prediction against the NLU/KB models of megabots and child bots.

The order in which predictions are made depends on where the internal action is used:

• When used in a flow on the megabot: The prediction first checks the NLU/KB models of the megabot, then the models of its child bots.
• When used in flows on child bots: The prediction follows this sequence:
1. The child bot where the flow is located.
2. The megabot to which the child bot is connected.
3. All other child bots of the megabot.

Parameters

This internal action encompasses the same parameters as the internal action GetIntentMatching.

This internal action retrieves the list of currently online Helpdesk Agents and stores it in [[HelpdeskAgentList]]. This action provides helpdesk managers with flexible options for checking agent availability, whether for all agents or specific queues.

Parameters

By default, the action retrieves all online agents. To filter the results to agents from a specific queue, provide the queue code in one of the following formats: [[key]].queueCode or @.queueCode.

This internal action analyzes the unweighted sentiment of the specified intent and stores the sentiment score in the @UnweightedSentiment variable as a string decimal (e.g., "-0.3"). This allows for easy interpretation and further processing of the sentiment analysis results.

Parameters

{
    "Intent":"@phrase/[[Entity]].Intent"
}

This internal action retrieves the list of clients currently waiting in the helpdesk queue and stores it in [[HelpdeskClientList]], providing a clear view of all clients waiting to be picked up by a helpdesk agent. This action is particularly useful when used within the ConnectToHelpdeskClientMultichat flow.

Parameters

By default, the action retrieves all clients in the waiting queue. To check the list of clients waiting in a specific queue, provide the queue code in one of the following formats: [[key]].queueCode or @.queueCode.

The IsInRole action checks whether the current user has any of the specified roles associated with their account. It returns the result as a string, either "true" or "false". This action allows you to control the flows based on user roles, enabling personalized experiences and permissions management.

Parameters

You can specify one or more roles to check against:

• Single role: ["User"]
• Multiple roles: ["Employee", "Manager"]

Usage

You can use the result of this action to set up conditions to perform subsequent steps by using one of the following expressions:

• To check if the result is "true":
Copy

@InternalActionResult == true.toString()

• To check if the result is "true" in a case-insensitive manner:
Copy

@InternalActionResult.toLowerCase() == "true"

This internal action lets you create knowledge base (KB) data sources directly within your conversational flows. Use it to create KB data sources from files or SharePoint site libraries, enabling access to relevant information in your conversations.

Parameters

{
  "Name":"[[Entity]].Field",
  "Type":"string",
  "Language":"en-US",
  "SourceFile":"[[Entity]].FileProperty",
  "SharepointTenantId":"string",
  "SharepointClientId":“string",
  "SharepointClientSecret":"string",
  "Uri":"string”,
  "MinThreshold": "string",
  "TargetThreshold": "string",
  “StoreType”:”string”,
  “Host”:”string”,
  “UserName”:”string”,
  “Password”:”string”,
  “DisableCertificat”:”string”,
  “Port”:”string”,
  “CrawlIntegrationId”:”string”,
  “ExtractintegrationId”:”string”,
  “Icon”:”string”,
  “SourceType”:”string”
}

The action returns the following data source information that you can use to match articles within that specific data source:

• [[KBOperation]].CreatedDataSource.Id
• [[KBOperation]].CreatedDataSource.Name
• [[KBOperation]].CreatedDataSource.CreatedOn

For more information, see KB Management at Conversation Time.

This internal action allows you to delete a specific KB data source directly within your conversational flows.

Parameters

{
  "DataSourceId":"[[Entity]].Field"
}

This internal action allows you to edit a specific KB data source directly within your conversational flows.

Parameters

{
  "DataSourceId":"[[Entity]].Field",
  "SourceFile":"[[Entity]].FileProperty",
  "Name":"string",
  "Uri":"string",
  "SharepointTenantId":"string",
  "SharepointClientId":"string",
  "SharepointClientSecret":"string",
  "MinThreshold":"string",
  "TargetThreshold":"string",
  "StoreType":"string",
  "Host":"string",
  "UserName":"string",
  "Password":"string",
  "DisableCertificatValidation":"string",
  "Port""string"
}

Each data source type has unique settings you can adjust to keep your knowledge base (KB) data accurate and up-to-date. Here’s what you can edit for each data source type:

• For a data source of type File (Excel or PDF), you can:
• Edit the name
• Upload a new source file
• Update extraction thresholds
• For a data source of type SharePoint, you can:
• Edit SharePoint credentials (Tenant ID, Client ID, and Client Secret)
• Update extraction thresholds
• For a data source of type Shared Drive (FTP/SFTP), you can:
• Edit the host and port
• Update FTP/SFTP login credentials (ID and password)
• Update extraction thresholds

This internal action extracts the content of a document directly within your conversational flows and returns it as a collection of strings. This action accepts a document from a [[File]] entity field and provides output in the following format:

• For unstructured documents: content is returned in [[KBOperation]].ExtractedDocument[i].Content.
• For structured documents: content is divided into [[KBOperation]].ExtractedDocument[i].Question and [[KBOperation]].ExtractedDocument[i].Answer.

Parameters

{
  "SourceFile":"[[Entity]].FileProperty",
  "DocumentType":"structured" //optional, default value = unstructured
}

This internal action crawls a specified website directly within your conversational flows and extracts unstructured content as paragraphs. The extracted content is returned as a collection of strings in [[KBOperation]].ExtractedDocument[i].Question and [[KBOperation]].ExtractedDocument[i].Answer.

Parameters

{
  "Url": "",
  "CrawlHttpRequests":"true|false"
}

This KB-specific internal action trains the knowledge base directly from within your conversational flows after adding or updating a specific data source.

{
  "DataSourceId":"[[Entity]].Field"
}

This internal action allows you to:

• Filter an indexed entity and return all similar records into an entity list.
• Perform extended Named Entity Recognition (NER) on entity candidates identified by the Flow Engine.

Parameters

{ 
"Index": "entity_index", 
"Value": "@variable | [[Intent]].Entities[0].Text",
"MinScoreThreshold": integer,
"MaxLimitForOutput": 25
} 

Basic Lookup Example

In this example, the action populates [[LocationList]] with records similar to @Location, using a similarity score threshold of 30%.

{
  "Index":"Location",
  "Value":"@Location",
  "MinScoreThreshold":30,
  "MaxLimitForOutput":20
}

Extended Lookup Example

This example applies extended NER to entity candidates found by the Flow Engine in the CourseTopic index, with a minimum similarity score threshold of 40%.

{
  "Index":"CourseTopic",
  "Value":"[[Intent]].Entities[0].Text",
  "MinScoreThreshold":40,
  "MaxLimitForOutput":20
}

Marks all previous alert messages as read, clearing any unread status for past alerts.

No parameters are required.

This action removes the specified roles from the current user.

string / List<string> / [[Entity]].EntityField / @Variable

Parameters

This internal action accepts as parameter the role names you want to remove. The format depends on the chosen input method:

• String: A single role name (e.g., "Helpdesk Agent").
• List of Strings: An array containing multiple role names (e.g., ["Employee","Manager"]).
• Dynamic roles specified in [[Role]].Name set in the Set Variables section. You can use any entity list that stores roles, an entity that references the [[Role]] system entity.

Example: Imagine you have a field named [[ChatUser]].RemoveFromRoles, which is an entity list referencing the system entity [[Role]]. You can dynamically assign roles based on user data using the Set Variables section.

This internal action requests a break for the current helpdesk agent once all active conversations are completed. No parameters are required.

This internal action requests a logout for the current helpdesk agent once all active conversations have concluded. No parameters are required.

This internal action sets [[ChatUser]].IsAuthenticated to true, indicating that the user is authenticated and can access flows that require authentication. No parameters are required.

This internal action to match an existing profile or provision a new user as an authenticated user in the Druid Portal. This action is designed for the Authentication Flow.

The action uses [[ChatUser]].MasterId and [[ChatUser]].DomainUserName to search the DRUID Users database for a matching user profile. If it finds a match, it replaces the current anonymous user profile with the identified user’s profile. If it doesn’t find a match, it provisions a new user profile.

The table below lists the [[ChatUser]] attributes expected as input by the SaveIdentifiedUser action. In the Set Variables section of the step where you call the internal action, provide the [[ChatUser]] attributes listed in the table below.

The system applies the following criteria for matching the user profile:

1. Searches for MasterId. If found, it verifies the DomainUserName. If they match, it links the profile.
2. If MasterId exists but DomainUserName does not match, it updates DomainUserName and other profile details (FirstName, LastName, Email, DomainUserName).
3. If MasterId is not found, it creates a new user profile.

Parameters

{
    "UpdateProfile": true,
    "MatchBy": "Email/MasterId"
}

Example: Identify the user by Email and update/provision the user profile

In this example, the following [[ChatUser]] fields are mandatory: FirstName, LastName and MasterId.

{
    "UpdateProfile": true,
    "MatchBy": "Email"
}

Example: Identify the user by MasterId and update/provision user profile

In this example, the following [[ChatUser]] fields are mandatory: FirstName, LastName and Email.

{
    "UpdateProfile": true,
    "MatchBy": "MasterId"
}

Error codes

The following error codes may occur in [[SystemError]].Code:

This internal action sets [[ChatUser]].IsAuthenticated to false, marking the user as unauthenticated. The user will no longer have access to flows that require authentication. No parameters are required.

This internal action signs in the current user as a helpdesk agent, enabling them to join live chat sessions and take over customer inquiries from the available agent pool. No parameters are required.

The action applies to Multi-chat (route-to-human).

This internal action starts a break for the current help desk agent. No parameters are required.

This action applies to Multi-chat (route-to-human).

This internal action takes the specified helpdesk agents offline based on the specified parameter.

The action applies to Multi-chat (route-to-human).

Parameters

This internal action takes the current user off the helpdesk waiting queue. No parameters are required.

The action applies to Live chat / route-to-human.

This internal action transfers the client to another queue or helpdesk agent. Use either the code of a live chat queue, or the AgentId of an agent, to transfer the call to a specific agent (e.g., [[HelpdeskAgent]].AgentId).

Parameters

{
    "QueueCode":"string"/[[<Entity>]].<QueueCodefield>/@queue_code,
    "PriorityLevel":"@priority",
    "AutomaticAllocation": boolean,
    "DisplayConnectButton": boolean,
    "SendConversationTranscript": boolean,
    "ContactCenterParameters":{}
}

Example: In this example:

• The call will be transferred to the queue "Support".
• The priority level for the call is defined by the variable @priority.
• The system automatically assigns the call to the next available agent.

{
    "queueCode":"Support",
    "PriorityLevel":"@priority",
    "AutomaticAllocation" : true
}

The UpdateIntentMatching internal action refines intent matching outcomes by incorporating post-processing resolutions. The KB Response Flow primarily uses this action to enhance intent detection, particularly when a Large Language Model (LLM) provides the final answer and might have a different interpretation of the initial intent match.

This action ensures that the Evaluation Tool accurately logs LLM-generated responses, including those initially categorized as ‘Intent Not Found’.

By using UpdateIntentMatching, authors can override the Flow Engine's default scoring and precisely record whether a relevant intent was ultimately found.

Key Benefits

• Facilitates manual corrections and feedback loops to refine intent recognition.
• Reduces false negatives by correctly recording when a response is provided, even after an initial "intent not found".
• Optimizes the handling of GPT-generated responses, particularly in "Intent Not Found" scenarios.
• Facilitates Knowledge Base refinement. Logs the source of KB articles, aiding in the evaluation and improvement of KB content.

When and how to use UpdateIntentMatching

Implement this internal action to handle key scenarios where the initial intent matching needs correction. Add UpdateIntentMatching as a Post Internal Action. The parameters you set depend on the situation.

• Potentially relevant matches found but not used by LLM. When the KB Engine retrieves articles that the LLM considers irrelevant, use UpdateIntentMatching to log these KB articles in the Evaluation Tool. This allows for subsequent manual review and correction, and it records the source bot (if the predict is done from a megabot) and article for better tracking. You can use the provided JSON structure, populating the KBQ* and KBBot* fields with information from the [[Intent]] context.
Copy

{
  "MatchFound": false,
  "FlowId": "00000000-0000-0000-0000-000000000000",
  "FlowScore": 0,
  "KBQ1Id": "[[Intent]].KBQ1.Id",
  "KBQ1Name": "[[Intent]].KBQ1.Name",
  "KBQ1Score": "[[Intent]].KBQ1.Score",
  "KBBot1Id": "[[Intent]].KBQ1.BotId",
  "KBBot1Name": "[[Intent]].KBQ1.BotName",
  "KBQ2Id": "[[Intent]].KBQ2.Id",
  "KBQ2Name": "[[Intent]].KBQ2.Name",
  "KBQ2Score": "[[Intent]].KBQ2.Score",
  "KBBot2Id": "[[Intent]].KBQ2.BotId",
  "KBBot2Name": "[[Intent]].KBQ2.BotName",
  "KBQ3Id": "[[Intent]].KBQ3.Id",
  "KBQ3Name": "[[Intent]].KBQ3.Name",
  "KBQ3Score": "[[Intent]].KBQ3.Score",
  "KBBot3Id": "[[Intent]].KBQ3.BotId",
  "KBBot3Name": "[[Intent]].KBQ3.BotName",
  

• False negative scenario. Implement this internal action within your ‘Intent not found’ flow that uses LLM. If the Flow Engine executes the ‘Intent not found’ flow (due to an initial lack of intent matches), but the LLM within that flow successfully generates an answer, the final output will be a positive match (an intent is found because the LLM addressed the user's intent). However, by default, the response is logged as 'not found' in the Evaluation Tool. To correct this false negative misclassification, use the UpdateIntentMatching internal action to log the triggered flow's information in the Evaluation Tool. This information is accessible via [[Intent]].Match. Additionally, set the MatchFound to true.
Copy

{
  "MatchFound": true,
  "FlowId": "[[Intent]].Match.Id",
  "FlowName": "Match.Name",
  "FlowScore": null,
  "FlowBotId": "[[Intent]].Match.BotId"
}