How to make the agent include images in its reply?

In the fixed script settings, only text responses can be configured. When images are uploaded to the knowledge base, they are automatically parsed, but you still cannot use the uploaded images directly as responses to customer inquiries.

We recommend uploading the image to the knowledge base first, then proceeding as follows:

  1. Upload files to the knowledge base, or edit an existing knowledge base chunk and add the image.
  2. You can turn off “Parse Image Elements,” as shown below:
  3. Add an image description: In the chunk, edit the section where the image is located, then right-click to edit and add a description, such as “Product Price Table.” This allows you to manually specify the meaning of the image.

Once these tasks are completed, return to [AI Skills] or follow the “Condition Adherence” guidance, and use a prompt stating: “Retrieve the knowledge base and return the Product Price Table image.”

Please try this approach first. If you encounter any issues, feel free to provide further details. Thank you.

The PDF I uploaded is being automatically split apart. How can I keep it as a single image?

Does “a complete sheet” refer to a single PowerPoint slide? If convenient, please upload it as an image to the knowledge base.