We have several decision tables having more than 100 rows and having 10+ columns each. The requirement is to ingest the contents of these decision tables into Pega Knowledge Buddy.
We have tried out the following options for getting the decision table content in plain text but those did not work as expected.
As export of the decision table rule as excel and then publishing appropriate knowledge article in KM with the excel as attachment and subsequent ingestion into knowledge buddy - When user asks questions through Knowledge Buddy, it is not able to interpret the tabular contents properly and does not give expected result.
Using Pega Gen AI connect to summarize the row content from the XML of the decision table rule - Pega Gen AI connect fails to differentiate between rows and repeats the same result for each row, also it is not able to take into consideration more than a few rows and columns so not all rows and columns could be taken into account.
Please advise -
how to make this solution feasible to work with minimal human effort and maximum automation for ingestion of the documents as there are plenty of such voluminous decision tables which hold the business logic?
how to make Knowledge buddy interpret the content properly and provide the correct contextual answer?
Knowledge Buddy supports ingestion of various types of text-based content through different methods:
Supported Content Types:
Web content
SharePoint documents
PDF files
Text-formatted attachments
Microsoft Office files
Content from Knowledge Hubs
Ingestion Methods:
REST API service for text-based content
Direct integration with Pega Knowledge for automatic ingestion upon article publication
Knowledge Buddy gadget for seamless content management
Pega Knowledge Loader application for SharePoint content
Regarding your specific question about Decision Management table contents: While Knowledge Buddy can ingest any text-based content through its REST API service, there isn’t explicit documentation about direct export capabilities from Decision Management tables for Knowledge Buddy ingestion. However, since Knowledge Buddy processes text-based content, you would theoretically be able to ingest the data if it can be exported in a supported text format.
For your specific use case, I’d recommend:
First verify if your Decision Management table can be exported to a supported text-based format (such as CSV or structured text)
If exportable, ensure the text maintains the necessary structure and context
This is a GenAI-powered answer. All generated answers require validation against the provided references.