Dialogue AI Assist Assistant Editor User Guide

Introduction

Assistant Editors configure the assistants that support agents with Dialogue AI Assist in WebChat and on voice calls. Use the Deepdesk admin portal to configure assistant behavior, presentation, knowledge, and assignment settings.

Before changing assistant configuration, read the Dialogue AI Assist User Guide for Agents to understand how agents use assistant output during customer interactions. You can also use the tutorial or playground in Deepdesk to test behavior while you refine an assistant.

Accessing the Deepdesk admin portal

To access the Deepdesk admin portal, go to https://[account].deepdesk.com. The exact URL is unique for your organization. For the exact URL, contact your AnywhereNow representative.

To open the assistant area, select Agentic AI > Assistants.

Deepdesk admin portal areas for Assistant Editors

The Deepdesk admin portal contains several menu groups. The Assistant Editor tasks in this page focus on the areas used to configure assistants and their supporting resources:

• Agentic AI > Assistants: Create, configure, test, monitor, and assign assistants.

• Agentic AI > Settings: Configure assistant connectors, routes, and groups used by assistants.

• Admin Settings > Profile: Assign assistants to profiles for event-based, knowledge, and on-demand use.

• Admin Settings > Knowledge Base: Create and process knowledge sources that assistants can use.

Other menu groups, such as Recommendation Studio, are used for Editor tasks such as text suggestions, URL suggestions, pinned messages, style suggestions, and images. Assistant Editors do not maintain those reply suggestions in Recommendation Studio. Use the chat channel editor guide for those responsibilities.

Assistants

Assistants support agents by interpreting conversation context, producing relevant responses, calling configured tools, and using attached knowledge where available. An assistant can be used directly, assigned to specific conversation events, made available on demand, or exposed so another assistant can call it as a tool.

Assistants list

Agentic AI > Assistants shows the assistant inventory for your environment. Use Add assistant to create an assistant. Use search, All Labels, and the date range filter, for example Last 30 days, to find or filter assistants.

The page shows summary metrics for Requests, Tokens, and Cost. The assistant table contains these columns:

• Name

• Feedback score: The feedback score for assistant responses.

• Errors: The number of errors encountered.

• Requests: The total number of requests to the assistant.

• Model Cost: The total cost incurred by model usage.

• Token Usage: The total number of tokens processed.

Rows are paginated.

The assistant editor currently contains the following tabs:

• Insights

• Behavior

• Integration

• Run Control

• Feedback

• Activity

• Experiments

Insights

Use the Insights tab to review assistant usage and performance. The tab includes Download CSV, a date range filter such as Last 30 days, a summary text that includes agents, conversations, rating, and cost, and metric cards for Requests, Rating, and Cost. Requests is the number of requests received by the assistant. Rating is the average assistant rating. Cost is the assistant cost in USD. Charts can show metrics such as Requests.

Feedback

Use the Feedback tab to review user feedback for the assistant. The tab includes a date range filter and a table with these columns:

• Date

• Feedback score

• Feedback

• Assistant Response

• User

• Conversation ID

When no feedback is available, the table shows No items to display.

Activity

Use the Activity tab to review assistant runs. The tab includes Refresh, a date range filter such as Today, and a table with these columns:

• Start Time

• Event Type

• Output

• Conversation ID

• Status

• Duration

• Tokens

• User

When no activity is available, the table shows No items to display.

Experiments

Use the Experiments tab to add and review assistant experiments. Select Add experiment to create an experiment. When no experiments are available, the tab shows No experiments found.

Test your Assistant

Use the test panel to try an assistant before relying on it in production workflows. The side panel is titled with the assistant name and shows Start a conversation with [assistant name]. Use the input field with placeholder Type your message... to test the assistant response. Before you enable an assistant for the whole team, test it in Studio and ask a pilot agent to try it in WebChat or on calls.

Creating and naming assistants

When you create or edit an assistant, choose a clear name and a stable code. The name helps Assistant Editors recognize the assistant in the portal. The code is used to identify the assistant across assistant configuration and should remain stable after the assistant is used by assignments or other assistants.

Select Add assistant to open the create assistant popup. The popup contains Name, a prefilled Code, an edit button next to the code, an Instructions text area, and Save. Select the code edit button if you need to change the generated code. Selecting outside the popup closes it.

Behavior

Use the Behavior tab to configure how an assistant responds and which supporting resources it can use. The Behavior tab contains the following sections:

• Prompt

• Tools

• Presentation

• Knowledge

• Settings

Prompt

Use Prompt to define what the assistant must do, how it should interpret the available conversation context, when it should call tools, and how it should respond. Good prompts are specific about the task, the expected input, and the expected output. Include examples when the assistant must follow a particular structure or tone.

Use this section to guide how the assistant supports agents and customers, including communication style, tone, and context retention.

The prompt text area is labelled Instructions *. Describe the assistant's task, tool usage, and response instructions.

Handover summary and notification pattern

Use this pattern when an assistant should create a compact handover summary for an agent joining or taking over an ongoing interaction. If the assistant must proactively notify the agent, assign send_user_notification and explicitly instruct the assistant to use that tool instead of adding call-to-action buttons.

Copy

Example prompt

# Task
You are a helpful assistant supporting a customer service agent during an interaction.
Use the transcript and available metadata to create a compact handover summary for the agent.

# Output
Send a short notification that includes:
- the customer's reason for contact
- the most important facts
- actions already taken
- open questions or next steps

Do not include call-to-action buttons.
If the transcript does not contain enough information, say what is missing instead of guessing.

Sentiment insight pattern

Use this pattern when an assistant should classify the tone of the conversation and provide a short explanation. This pattern does not require tools.

Copy

Example prompt

# Task
Analyze the transcript and classify the customer's sentiment as positive, neutral, or negative.

# Output
Return one short sentence with the sentiment and the reason for the classification.

# Example
Input: The customer says the issue is resolved and thanks the agent for the help.
Output: Positive — the customer confirms the issue is resolved and expresses appreciation.

Verification reminder pattern

Use this pattern when an assistant should detect topics where a compliance or verification reminder is needed. Define the trigger criteria and the no-result behavior so the assistant does not create irrelevant reminders.

Copy

Example prompt

# Task
Review the transcript for situations where the agent must verify customer details before continuing.

# Trigger criteria
Return a reminder only when the conversation includes a request that requires verification.

# Output
If verification is required, return a concise reminder and explain why it applies.
If verification is not required, return "none".

Knowledge Base answer pattern

Use this pattern when an assistant must answer from a connected Knowledge Base. Attach the Knowledge Base from Behavior > Knowledge and add knowledge retrieval tools only when they are available and required for the assistant design.

Copy

Example prompt

# Task
Answer the agent's question using the connected knowledge source.

# Rules
Use the knowledge source as the source of truth.
If the answer is not available in the source, say that the answer is not available.
Include source information when it is available.
Keep the answer concise and agent-facing.

Microsoft Copilot-backed reply suggestion pattern

Use this pattern when an assistant has a Microsoft Copilot tool assigned and should draft an agent reply grounded in Copilot-backed sources. Configure Copilot tools from Agentic AI > Settings > Connectors.

Copy

Example prompt

# Task
Draft a reply that the agent can send to the customer.

# Rules
Answer as the agent.
Keep the reply concise and professional.
Use the assigned Copilot-backed source for grounding.
Do not mention internal tools or connector names to the customer.
Do not invent product details that are not present in the source.

Linked-assistant routing and orchestration pattern

Use this pattern when an assistant assigned to a conversation event must decide whether one or more linked assistants are relevant. Adjust the available assistants and the output requirements for your use case.

Copy

Example prompt

# Task
You are a helpful assistant supporting a customer service agent during an interaction.
Use the full transcript and assistant descriptions provided to determine if calling one of the available assistants is necessary.
Only call an assistant if it is highly relevant.

Available assistants:
--list of the assistants that may be called for this assignment.

If you lack the parameters required to call an assistant, skip it. Do not ask for missing parameters.
Call each assistant only once and provide a single response for each assistant.
Use only the assigned assistants; do not create new ones.
If an assistant has previously provided a comparable response, reuse the older one.
Exclude any responses from assistants that are "none", "{}", or equivalent.
If a response seems irrelevant, omit it.
If an assistant's response contains markdown, keep that intact.

If none of the assistants seems relevant, or none have relevant output, say: "[]".

Return the results as a list of objects with these keys:
`code`: The assistant code
`name`: The assistant name
`response`: The assistant response

# Example
[
{
"code": "how-to-assistant",
"name": "How-to Assistant",
"response": "The customer is experiencing issues with their Wi-Fi. They have tried resetting their router, but their devices still cannot connect."
},
{
"code": "sentiment-assistant",
"name": "Sentiment Assistant",
"response": "The sentiment is neutral."
}
]

Tools

Use Tools to configure the tools that are available to the assistant in your environment. The available tool list can differ by environment. Common tool patterns include calling a URL, calling a configured external API, or validating data with JSON logic.

Select Add to assign tools. If no tools are assigned, the section shows No tools assigned. Assigned tools appear by name and can be removed.

In Assign tools to assistants, choose the tools the assistant can use for its tasks. The assistant only works with the tools you select. Use Type to search to filter tools, then select Assign or Cancel.

Tools are grouped into these categories:

• System tools: Built-in tools visible in the environment.

• Assistants: Other assistants configured under Agentic AI > Assistants. Select an assistant here when this assistant should call it as a tool.

• Microssoft Copilot: Copilot tools configured through Agentic AI > Settings > Connectors.

• APIs: Tools generated from Assistant Connectors configured under Agentic AI > Settings > Connectors.

System tools include options for knowledge retrieval, API calls, notifications, response suggestions, decision-tree traversal, validation, and memory. Available tool examples include:

• answer_knowledge_assist_question: Answer a question based on the context.

• call_api: Make an API HTTP request.

• call_knowledge_assist: Make an HTTP call to Deepdesk knowledge assist.

• call_url: Make an HTTP request.

• get_current_time: Get the current datetime in ISO format with day of week.

• realtime_event: Send an event to a realtime endpoint with a payload.

• request_frontend_data: Request data from the frontend.

• retrieve_knowledge: Retrieve knowledge from the knowledge base.

• send_call_to_action: Send one or more call-to-action buttons to the user.

• send_response_chunk: Send a response chunk to the user.

• send_response_sources: Send response sources to the user.

• send_user_notification: Send one or more notifications to the user.

• store_generated_analytics: Store generated analytics events with non-deterministic data as key-value pairs.

• suggest_as_reply: Suggest one or more replies to the user.

• traverse_decision_tree: Traverse a decision tree for structured guidance.

• validate: Validate data against JSON logic.

• write_to_memory: Write data to memory.

If an assistant calls another assistant, include that assistant in the available tools and make sure the called assistant is configured in the Integration tab to be available as a tool.

Presentation

Use Presentation to configure how the assistant's responses appear in the Deepdesk widget. The fields are:

• Display Name: The name shown in the widget header. Leave this field empty to use the assistant's default name.

• Editable: Allows users to edit the assistant's response directly in the widget before using it.

Knowledge

Use Knowledge to attach a knowledge source to the assistant when this option is available in your environment. Use Select a base to connect the assistant to a Knowledge Base or decision tree.

Attach a Knowledge Base when the assistant must answer from controlled source material instead of relying only on the transcript and prompt. Decision tree entries can appear in the same dropdown when available in the environment.

Knowledge Bases are managed from Admin Settings > Knowledge Base. See Knowledge Bases for source and processing guidance.

Settings

Use Settings to configure runtime options for the assistant. The settings are:

• Model: Select the AI model used for this assistant. This field is required.

• Include transcript: Adds the voice transcript or chat conversation to the prompt.

• Run Silently: Executes the assistant in the background without visible messages.

• Reuse Thread: Includes previous interactions with this assistant for context.

• Error Handling Mode: Controls how the assistant handles errors during execution. Options are Retry on error and Stop on error.

The Settings tab also contains a Delete button for removing the assistant.

Integration

Use the Integration tab to control whether another assistant can call this assistant as a tool.

The Tool Interface toggle is assistant-specific. For example, it can read Allow other assistants to call E2E Summarizer as tool. Use this setting to define how other assistants can use this assistant as a tool, including a description and any required properties.

Enable Allow other assistants to call [assistant name] as tool to make the assistant available as a tool. When this setting is enabled, configure the following fields:

• Tool Description: Describe what the assistant does and when another assistant should call it. Other assistants use this description to decide when to invoke this assistant as a tool.

• Properties: Define the parameters that another assistant can pass when it calls this assistant. Add properties using the form or with a JSON schema.

Properties has Form and JSON modes. In Form mode, define input fields, types, and descriptions using JSON Schema. Required fields determine what must be provided. Select Add property to add a property. Property fields include Name, Type, Required, and Property description (optional). Supported types include String, Integer, Number, and Boolean. Use the delete icon to remove a property.

In JSON mode, enter the JSON schema directly. Use Prettify JSON to format the schema.

The tool description should be specific enough for another assistant to decide whether this assistant is useful for the current conversation. Use Properties for parameters that are useful or required as input for this assistant. For example, if the assistant must receive a conversation source ID, define it as a required property:

{
"type":"object",
"required":[
"source_id"
],
"properties":{
"source_id":{
"type":"string",
"description":"The source ID of the interaction."
}
}
}

Run Control

Use the Run Control tab to configure the conditions and permissions that control whether the assistant runs. The fields are:

• Allowed Groups: Restricts assistant availability to specific groups. Leave this field empty to allow all groups. The dropdown uses groups from Agentic AI > Settings > Groups.

• Run Condition: Defines the JSON logic that must evaluate to true before the assistant runs. If it evaluates to false, the assistant is skipped. Use Prettify JSON to format the JSON logic.

• Fallback Assistant: Defines the assistant to use as a backup when this assistant fails or cannot respond. The dropdown uses assistants from Agentic AI > Assistants.

Use Run Condition when an assistant should only run for specific conversation data. For example, to run the assistant only when the conversation has a matching source_id, use JSON logic like this:

{
"==":[
{
"var":"source_id"
},
"SOURCE_ID"
]
}

Triggers

The Triggers section in Run Control shows where the assistant is used:

• Event Based Assistant Assignments: Lists profiles where this assistant is triggered automatically for specific conversation events. The table can include Event Name, Event Type, and Profiles. Profile links open the related Admin Settings > Profile pages.

• On-Demand Assistant Assignments: Lists profiles where this assistant is available to be triggered on demand.

• Linked Assistants: Lists assistants that use this assistant as a tool.

When no trigger data is available, the section shows No triggers found.

Agentic AI settings

Use Agentic AI > Settings to configure shared assistant resources. The settings area contains Connectors, Routes, and Groups.

Connectors

The Connectors tab contains Assistant Connectors and Microsoft Copilot Studio Connection.

Assistant Connectors

Assistant Connectors define APIs that can become tools for assistants. Use a connector when an assistant must look up or update live data, credentials must stay out of the prompt, or several assistants use the same backend system. Typical connector targets include order systems, customer relationship management (CRM Customer Relationship Management, or CRM, is (usually) a software-based, data management method to deal with interactions with customers and potential customers.), ticketing, or knowledge systems exposed through an API.

The overview shows connector cards with the connector code, name, and base URL. Select Add API to create an API connector. The add form contains Code *, URL *, Headers, Prettify JSON, and Save.

Use the connector fields as follows:

• Code *: Defines the unique connector identifier. Assistants refer to this value as api_code when they use the call_api tool. Use short lowercase values, such as orders or crm-cases.

• URL *: Defines the base URL for API calls made through the connector.

• Headers: Defines default HTTP headers for calls through the connector. Use JSON and select Prettify JSON to format the value.

• Authentication: Defines how the connector authenticates to the API. Use None for APIs that do not require authentication, Auth Headers when the API expects static authentication headers, and OAuth2 when the API requires token-based authentication.

The API connector edit page contains Tools and Settings tabs.

In Tools, use Generate tools from OpenAPI spec to make API endpoints available as assistant tools. Provide a URL to openapi.json and select Import, or paste the OpenAPI specification into the text area with placeholder or paste your OpenAPI spec here. Use Prettify JSON and Save as needed. Generated tools inherit the connector configuration, so the assistant can use the generated endpoint tools without placing base URLs or credentials in the assistant prompt.

In Settings, configure Code *, URL *, Headers, Prettify JSON, and Save. In Authentication, select the authentication type for the assistant API. Authentication options are None, OAuth2, and Auth Headers. In OAuth2 mode, configure the token endpoint and client credentials required by the API, such as Token URL, Client ID, Client Secret, and optional Scope when those fields are available. In Auth Headers mode, configure Auth Headers, Show, Prettify JSON, and Save.

When deleting a connector, confirm the deletion for the selected connector code.

Microsoft Copilot Studio Connection

Microsoft Copilot Studio connectors make Copilot-backed tools available to assistants. Select Add Copilot Agent to create a connector. Copilot agent cards show the name and Direct Line token URL.

The add form contains Name *, Code *, Token URL *, Description *, and Save. The Copilot Agent edit settings page contains Name *, Token URL *, Description *, Save, and Delete.

Routes

Use the Routes tab to manage assistant routes. Routes select or activate assistants from conversation metadata. Use routes when different queues, topics, languages, or other metadata values need different assistants within the same environment.

Routes evaluate JSON Logic in Tags. The rule is matched against metadata values such as queue, topic, locale, or values sent from a Dialogue Studio Dialogue AI Assist node. For example, a route can match a claims queue with { "==": [{ "var": "queue" }, "claims"] }, a Dutch locale with { "==": [{ "var": "locale" }, "nl-NL"] }, or a retention topic with { "==": [{ "var": "topic" }, "retention"] }.

Select + Add assistant route to open the drawer. The route table contains Assistant Code, Assistant Name, and Tags. When no routes are available, it shows No items to display.

In the add assistant route drawer, add the basic route information. The drawer contains Assistant Code *, Tags *, Prettify JSON, Save, and a close control.

Use the following guidance to decide between Routes and profile-level event assignments:

Routes and Profile configuration both use JSON Logic, but they are different settings. Routes select assistants from metadata. Profile configuration assigns the profile by evaluating received key-value tags.

For queue-based routing, create one route for each assistant. For example, route a claims assistant with { "==": [{ "var": "queue" }, "claims"] } and a sales assistant with { "==": [{ "var": "queue" }, "sales"] }. The metadata key must match the value received by Deepdesk, for example from a Dialogue Studio Dialogue AI Assist node tag.

Groups

Use the Groups tab to define groups that can be selected in Run Control > Allowed Groups. Groups limit who can see or run an assistant. Use groups for permission and rollout control; use profiles or routes to determine which conversation receives which assistant.

The tab contains Add Group, search, and a table with Name and In use for. Rows include Edit and Delete actions and are paginated.

The add group drawer contains Name *, Add User, Save, and a close control.

Common group patterns include:

• Pilot rollout: Create a group such as pilot and add early users before making a new assistant available to all users.

• Department or skill access: Create groups such as sales, tier-2, or voice-agents when an assistant is relevant to a specific agent population.

• Compliance access: Create a group such as compliance-officers when an assistant supports regulated or sensitive workflows.

• Open to all: Leave Allowed Groups empty on the assistant when every user can access it.

Using APIs and tools with assistants

APIs allow assistants to interact with other systems, retrieve or send data, or trigger actions. Configure external APIs and authentication details in the Deepdesk admin portal, then make the relevant tool available in the assistant Behavior > Tools section.

Assistant API configuration

Code

Choose a code for the API. This is how you identify it later across assistants. The code must be unique for each API.

URL

Configure the base URL required for each call.

Headers

Configure the headers required for each call. Headers are stored encrypted, so tokens can be added here.

The available authentication methods can include:

• No Auth

• OAuth2, with Token URL, Client ID, Client Secret, and optional Scope.

• Auth Headers, with the authentication headers required by the API.

Tool prompt patterns

Some tools require the prompt to include specific instruction patterns. Confirm that the tool is available in your environment before using it in an assistant.

call_url

If the assistant must call a URL, add call_url in Tools and include instructions like this in Prompt:

Call the call_url tool:
- url: ##http://some.url.com/api/path
- method: ##the action to perform on the API server, for example GET
- headers: {"Authorization": "******"} ##if you do not need headers, delete this line
- payload: ##any additional parameters to include in the request

call_api

If the assistant must call a configured external API, add call_api in Tools and include instructions like this in Prompt:

Call the call_api function:
- path: ##the endpoint path, optionally with query parameters
- api_code: ##the unique code of the API to call
- method: ##the action to perform on the API server, for example GET
- headers: {"Content-Type": "application/json"} ##if you do not need headers, delete this line
- payload: ##any additional parameters to include in the request

Example

Copy

Example prompt

Call the call_api function:
- path: 1.0/interactions/{customer_id}/save_summary
- api_code: crm_interactions
- method: POST
- payload: {"interaction_summary": "{summary}"}

validate

If the assistant must validate a set of data as part of its prompt, add validate in Tools and include instructions like this in Prompt:

Call validate function with the following logic and data:
- logic: ##JSON logic
- data: ##JSON data

Example

Copy

Example prompt

Call validate function with the following logic and data:
- logic: {"==": [{"var": "result"}, "200 OK"]}
- data: {"result": <call_api response>}

Knowledge Bases

Knowledge Bases can provide controlled source material for assistants. Use them when an assistant must answer from your own content, such as public web pages, PDF documents, or documents uploaded through an API.

Knowledge Bases are managed from Admin Settings > Knowledge Base. After a Knowledge Base is created and processed, attach it to an assistant from Behavior > Knowledge when that option is available in your environment.

Knowledge Base edit pages use a route pattern similar to /admin/settings/knowledge-base/assistant/[code].

Knowledge Base name and code

Choose a name and code for the Knowledge Base. The name helps administrators recognize it in the portal and is not visible to agents. The code identifies it in configuration and should be unique. The Name field is required.

Knowledge Base answering behavior

Configure guidance for how answers should be generated from the Knowledge Base. Use clear wording that tells the assistant to answer from the provided source material and to avoid guessing when the source does not contain the answer.

Use the provided context to answer the user's question.
If the answer is not available in the provided context, say that you do not know.
Do not make up an answer.

I do not know.

Source processing

Source processing scans and analyzes the online sources or PDFs added to the Knowledge Base. This processing extracts and comprehends information from the sources so the Knowledge Base can generate relevant responses. The duration depends on the size and complexity of the sources. After processing is complete, process the sources again when source content changes. Large content updates can take time to process, so plan ahead for product launches or policy changes.

The page shows processing status and last job details. Select Start Processing when sources are added or changed so the Knowledge Base uses the latest content.

Knowledge Base sources

Knowledge Base sources are organized in Online Sources, PDF Sources, and API tabs.

Online sources

Add URLs of publicly available sources, such as a website or community page. Add one URL per line in the text area labelled urls, then select Save.

PDF sources

PDF sources are accessible to Deepdesk users and cannot be shared or viewed by external parties, such as customers. Drag PDFs into the window to upload.

The PDF table contains a Name column, row actions to open or delete a PDF, pagination, and Add PDF.

API sources

API sources are documents uploaded through the API. The API table contains Title, Url, and Last Crawled. When no API sources are available, it shows No rows.

Deleting a Knowledge Base

Deleting a Knowledge Base permanently removes its configurations, settings, and sources. Save any required information elsewhere before you delete it.

Assistant assignments

Assistant assignment visibility is available under Triggers. Use Triggers to review where assistants are used and how they are made available. The Triggers area includes:

• Event Based Assistant Assignments

• On-Demand Assistant Assignments

• Linked Assistants

Profile settings

Use Admin Settings > Profile to configure a profile. The page heading shows Profile settings and the profile name, for example profile-michael. Profile settings combine assistant assignments with the Profile configuration rules that determine when the profile is assigned to a conversation.

For voice conversations that use Dialogue Studio Dialogue AI Assist nodes, the node sends key-value tags as conversation context. Profile configuration evaluates those received values with JSON Logic and assigns the profile when the rule matches. Configure tags in the Dialogue AI Assist Nodes; configure the matching JSON Logic rule in the Deepdesk profile. The profile page evaluates incoming values and does not configure the tags sent by Dialogue Studio.

Profile settings contain the following areas:

• Knowledge assistant

• Event Based Assistants

• On-Demand Assistants

• Profile configuration

• Recommendation Engine

• Delete Profile

Knowledge assistant

Use Knowledge assistant in profile settings when a profile needs a dedicated knowledge-backed assistant. The dropdown placeholder is Select a knowledge assistant. Configure the underlying Knowledge Base from Admin Settings > Knowledge Base, then attach it to the relevant assistant from Behavior > Knowledge where available.

Event Based Assistants

Event Based Assistants rows combine an assistant selection with an event selection. Assistants can be assigned to these events:

• Conversation accepted: Triggered when an agent accepts a conversation.

• Conversation new message: Triggered upon receiving a new message from either the customer or the agent.

• Conversation ended: Triggered when the conversation ends, either through closure, transfer, or customer departure.

Use event-based assignments when the assistant should run automatically at a specific moment in the conversation. For example, assign a summarizing assistant to Conversation ended when the assistant should prepare a summary after the interaction ends. Enable Include transcript in the assistant Settings when the assistant requires the conversation transcript. Your AnywhereNow representative may provide recommended assistant setups for your organization.

On-Demand Assistants

On-Demand Assistants are available for agents to start when needed. The table contains Assistant Name, Button Label, and Pinned. Select Add Assistant, choose an assistant, optionally enter Button Label, configure Pinned, then select Add.

In profile settings, on-demand assignments include the following fields:

• Pinned: Keeps the assistant prominently available to the agent when supported by the client.

• Button Label: Defines the label shown to the agent for the on-demand assistant.

Assistant setup examples

The following examples show compact configuration patterns for common assistant scenarios. Adjust names, prompts, routes, and groups to match your organization's process and available connectors.

Order status assistant

Use this pattern when agents need an on-demand assistant that retrieves order information from an external API.

Verification reminder assistant

Use this pattern when agents need a reminder or checklist when they accept a conversation.

Knowledge search assistant

Use this pattern when an assistant should answer from controlled source material.

Next-best-action or retention assistant

Use this pattern when an assistant should coach agents during a conversation based on transcript, queue, or topic context.

Profile configuration

Profile configuration specifies the rules that trigger the assignment of a profile to a conversation. Use the JSON Logic editor to configure the rule. The editor includes View, Edit, JSON, and Save controls.

For example, if a Dialogue Studio Dialogue AI Assist node sends a tag with key profile and value support, the profile can match that value with this JSON Logic rule:

{
"==": [
{
"var": "profile"
},
"support"
]
}

When the profile configuration is published, the page shows This profile configuration is published.

Recommendation Engine

Recommendation Engine shows whether recommendation engine functionality is enabled for the profile. Contact support to enable or disable it.

Delete Profile

Delete Profile permanently removes all settings associated with the profile.