Enabling document-based interactions in chat

Edit online

Enable the chat to receive and process user-uploaded documents, allowing the agent to generate answers based on the document’s content. This feature makes the agent more context-aware and helpful across various use cases by leveraging the full document for accurate responses. Use this option when your agent needs to analyze or extract information from files that users provide.

Use case examples:

  • HR agents reviewing resumes or policy PDFs

  • Finance agents extracting data from invoices

  • Support agents analyzing logs or screenshots

Before you begin

Before you enable document upload in chat:

  • Ensure that you are signed in as an Admin or Builder.

  • Decide how many citations you want to show in the chat. When users upload documents, they can review citations in the chat to verify accuracy. Citations show the sources of data, models, or algorithms that the agent uses to generate outputs or make predictions.

Note: Document processing is currently supported only in embedded and web chat channels. Other digital channels are not supported.

Steps to enable document upload in Orchestrate Chat

To enable the document upload feature, complete the following steps:

  1. Go to the agent configuration page.

  2. Navigate to the Behavior section.

  3. Toggle Chat with documents to On.

  4. In Citations shown in chat, choose how many references the agent must include in its response.

  5. Click Deploy.

  6. After deployment, users can see the File Upload icon File Upload icon in the Orchestrate Chat.

After you enable uploading documents, see how users can use it in the chat conversations. See Uploading the document in a conversation for more details.

Note:

If this feature is turned off, users can only upload files through tools that support file upload.

What to do next

  • Test the agent in chat to ensure expected responses.

  • Review uploaded document behavior for accuracy.

  • Adjust instructions or guidelines based on test results.