Configure Generative AI chat control to search using Generative AI

Configure the following settings to perform search using Generative AI.

Setting

Description

Header label

Enter the header label to be displayed at runtime. (Default: Generative AI Chat)

Search type

To execute the search based on Generative AI, select Generative AI.

Provider

Select the Generative AI provider to use, such as OpenAI, AI Agent, Custom LLM, or Azure OpenAI. (Default: Tungsten)

Alternatively, select  to provide an inline value, or select a form control, a form variable, or a global variable.

The Provider list includes the AI providers configured in Integration > Generative AI. See Integrate Generative AI with TotalAgility.

If a Generative AI provider is not already configured, a warning message appears when you add the Generative AI chat control to a form.

When you select an AI Agent provider, lists for additionally input and output variables appear. You can set a static value or select a dynamic variable (form control, form variable, global variable, and Copilot Insights) for the input and output variables.

Model

The AI model used by the component depends on the type of provider you select. Providers are listed under the names given when they were configured, so a name such as "ChatGPT" may belong to any type. To check the type of a provider, open it in Integration > Generative AI and look at Type. The name of the property and the options it offers vary by provider type:

Provider type	Model property
Tungsten	The property is labelled Model and a list is displayed, set to Default. Do one of the following: Keep Default to use the model configured in System > System settings > System > Tungsten AI > Model. Select a specific model to use for this component only. Select  to provide an inline value, a String-compatible variable or a server variable that contains the model name.
OpenAI Compatible	The property is labelled Override model and no list is displayed. Leave it empty to use the model configured at the integration level, or enter the model name to override it. Select  to provide the value as a String-compatible process or server variable instead.
Azure OpenAI	Not available. The model is part of the URL configured for that provider.
Custom LLM or AI Agent	Not available.

When you provide an inline value for the Provider, the model cannot be selected from a list, so you must provide an inline value for the model as well. If either value is a variable, it is resolved at runtime.
A provider is configured with a standard model and a vision model. The model you specify here overrides the vision model when the component sends an image, and the standard model otherwise.
If the configured model has been removed or is otherwise unavailable, the error is raised at runtime, not when you design or publish the component. An invalid model name is not validated: the request is sent as entered.
On upgrading or importing from an earlier version, the Model of a component using the Tungsten provider is set to Default. For any other provider, the component continues to use the model configured on that provider's integration.

Source

Select one of the following options.

Option	Description
None (default)	The response from the AI provider is based on the input provided (plain text or variables).
Image	On the Document list, select a form control (Web capture control), form variable, or global variable of type String, and on the Mime type list, select a file type. The following mime types are supported: GIF, JPEG, and PNG. Mime type is not supported for form variables. When using images, make sure that your selected provider supports the use of images. control When you select an AI Agent or Custom LLM as the provider, at runtime: Document Content variable uses the content of the document as a base64 string. Document variable uses the instance ID of the document.
Text	The OCR text of the document is returned in a more logical reading order, resembling how a narrator would read it aloud. Instead of being presented in the "top-left to the bottom-right" order returned by OCR, the text is arranged more clearly and coherently. If the document contains layout text, such as a semi-structured document (for example, invoice or purchase order) or a fixed form, the document text can provide better results compared to using a Custom LLM provider. On the Document list, select the form control, form variable, or global variable. The response is generated based on the text passed as input. When you select an AI Agent or Custom LLM as the provider, at runtime: Input Prompt variable: Adds content of the image at the end of the input prompt. Document variable: Uses the instance ID of the document. The following file types are supported in the Transformation Designer for text source format in TotalAgility at runtime: TIFF (.tif), JPEG (.jpg), PNG (*.png), and PDF. When you set the source format as Text, the standard model configuration is used.
Inline image	Select the attachment icon îˆ¦ available at runtime to upload an image and ask a question about it. The thumbnail of the uploaded image is displayed. The response is generated based on the image provided at runtime. When you select an agent or Custom LLM as the provider, the Document Content variable uses the document content as a base64 string at runtime.

Persist conversation

If selected, the context persists during conversation. (Default: Clear)

Persisting the conversation increases the token usage that may result in exceeding the token limit. When the prompt exceeds the token limit, the earlier conversation history is automatically removed to manage the token limit.

Use response profile

If selected, enables adding a response profile, so that at runtime, your responses display in a specific word count or style. (Default: Clear)

Response profile

Provide an inline value, such as "Summarize the response in 20 words or less." or select a global variable, form variable, or form control.

Use seed

If selected, lets you dynamically set "Use seed" using a form variable (Boolean) or form control. (Default: Clear)

The seed value increases the likelihood of obtaining more consistent results from the AI provider across repeated requests, though not guaranteed.

Seed

Select a form variable or form control, or select a numeric value. (Default: 0 and Maximum: 32,767)

Include case data

If selected, lets you include case data for analysis when using Generative AI. (Default: Clear)

When you include case data, you are prompted to enhance your experience by providing descriptions for both the process and activities to achieve optimal results.

The "Include case data" option is not available when using a Custom LLM provider.

For Case identifier, select a variable (global variable, form variable, or form control) or provide an inline value for Case ref or Case ID.
A list of case data is displayed. By default, Key Case Details, History (full), Variables, and Notes options are selected. You can select the items from the list as needed.

At runtime, the case details are included in the information available to search and in the response.

Override default message

If selected, enables the Welcome message and Input help text boxes where you can enter a customized welcome message and help text to be displayed in the Chat control at runtime. (Default: Clear)

Option

Description

Welcome message

You can provide a static or dynamic value (form control, form variable, or global variable) or a combination of both.

For example, if you enter "Good day!" as the message, and a form variable called Name that contains the value "John Doe", the default "Hello <username>, how can I help you today?" is replaced with "Good day! John Doe" at runtime.

Input help text

Enter the help text to be displayed in the Chat window at runtime.

Response format

Select either option for response format.

HTML (default): Returns the response with HTML formatting, such as bullet points, headings, and bold text, to present structured content clearly and enhance readability.
None: Returns the response as plain text with no formatting.

On upgrading TotalAgility, the Response format is set to None.
This setting is not available when using Custom LLM or AI Agent providers.

Use MCP servers

If selected, enables you to add MCP servers to be executed at runtime. (Default: Clear)

Select .

The New MCP server dialog box appears.
From the list of configured MCP servers, select the required servers.

When MCP servers are added, all tools exposed by the selected servers become available for use in the chat control.

You can dynamically set what MCP server tools are available for use within a chat by selecting Override tools. When this option is enabled:
- The MCP server configuration is not modified.
- Only the tools selected for the chat control are available for use.
Optionally, use the Search tools box and the dropdown beside it to find specific tools by name and filter them. See Configure MCP server connection.

You can override MCP tools only at design time and not at runtime.
Select Save.

Show tool progress

If selected, displays MCP tool execution progress to indicate that a request is actively being processed. (Default: Clear)

This property is available only when Use MCP servers is enabled.

When an MCP tool is executed, the tool name is displayed during response generation and removed automatically after the response is generated.

Audit conversation

If selected, enables auditing of AI interactions (default: Clear). The audit log records:

Prompt-response pairs while maintaining user anonymity to comply with data protection standards.
The provider and the actual model used for the interaction.
Each tool executed during a chat interaction when MCP servers are used with a chat control.

When you submit a prompt in a Generative AI chat control on a form using Generative AI as the search type, an entry is recorded in the audit log.

When you view the Generative AI chat control at runtime, the Copilot chat window appears with the following:

- "This conversation is being audited" message.
  
  This message appears if the "Audit conversation" setting is enabled.
- Learn more link. Select the link to know more about Tungsten Automation's commitment to responsible AI and compliance.
- A welcome message, "Hello <username>, how can I help you today?".
  
  You can modify the welcome message and input help text, customize the appearance of the chat window, such as the header color of the Copilot chat window, set an icon for Copilot, and more. See Generative AI chat control style.
When you enter any input, the response is streamed before it is displayed. You can select the  button to cancel the request. If you add an image or text as part of the input, such as uploading a car registration image and requesting the AI provider to get the registration number, the response is returned accordingly.
When you configure the "Generative AI chat" control in a Same page action, Prompt and Conversation buttons appear at runtime. When the user submits a prompt in the Chat window, the corresponding properties are updated at runtime.
- Prompt contains the text the user entered in the chat.
- Conversation contains the conversation history based on the "Persist conversation" setting. If "Persist conversation" is enabled, the full conversation history, including all prompts and responses from the current session, is displayed; otherwise, only the most recent response from the AI provider is displayed. The conversation content is formatted based on the "Response format" setting (HTML or plain text).
To clear the current conversation, including the context (if "Persist conversation" is selected in Search using Generative AI) and start a new one, select  in the top right corner of the Chat box.
If MCP servers are used:
- You can perform MCP-related functions. For example, you can enter a prompt to upload AI-generated content or a document to SharePoint or Google Drive.
- An MCP server exposes one or more tools that an AI assistant can use. When an AI assistant needs to use an MCP tool that requires permission, a Tool Permission Required prompt is displayed. The prompt identifies the tool the AI assistant wants to use and allows you to decide whether to grant or deny access. See MCP tool approvals for more information.