Debugging Pega GenAI Connectors (Pega Infinity '25)

General
1

Hi this is urgent. Your answer would be helpful.

I have a GenAI Connector to extract the Form number from a PDF document. So in the connector prompt after explaining what and how it needs to be extracted, I have mentioned the content of the PDF in property {.formContent} which is a rich text property containing the PDF content in base 64 format.

The issue is when a second document is sent via the connector, it will not extract that document’s Form number and still shows the Form number of the first document. And I’ve found that it’s the same {.formContent} that’s been kept in the GenAI.

Do we have any refresh mechanism in the GenAI connectors or how can we refresh these properties that are been called inside the prompts?

P.S.: We are already removing the temp page that’s used to set the value to .formContent in an earlier step

2

Could you please provide more context?

Are you extracting the Form Number from a PDF document? I’m little confused since you mentioned about base64 format. Could you please share the screenshot of Gen AI connect or more insights?

4

Yes Form Number is extracted from a PDF document.

===========================================

It’s the PDF content that’s been converted to base 64 format for GenAI connector to read and extract the Form number.

Below is a glimpse of GenAI prompt.

  • AB89000 etc are the sample Form numbers given for GenAI to identify.
  • {.formContent} is how this base64 converted PDF content if referred in the prompt.

image
image1369×499 62.5 KB