Pega BluePrint - Which LLM are being used in the backend for ingesting documents

Blueprint and App Design
1

Hi All,

Do we know Which LLM are being used in the backend for ingesting documents , which are uploaded in blueprint wizard “upload supporting assets”

I had few observations, when we upload a JAVA code in wizard for business rule , its not able to get business rule out of it and same thing works better when i write a pseudocode equivalent of same Java code

If we know which LLM being used for each of below categories in Blueprint , we can better leverage the blueprint

Open to suggestions from the Pega Blueprint community

image
image1561×586 52 KB

2

According to the document “Pega Blueprint Security and Privacy” (version 3 of this file is available to view and download as of today)

  • Blueprint leverages multiple LLMs under the hood – including Anthropic models on AWS, Google Gemini, and OpenAI on MS Azure
  • Blueprint leverages LLM’s based on use case & performance. Primarily Claude models running on AWS Bedrock
  • All LLM’s are continually governed, performance tested and leverage provider best practices for content filtering
2 Likes
3

This helps Thanks so much , I went through the document link you shared. I am more interested in knowing specifically which LLM being used when we ingest documents in each of this categories , intention to ask this question is to have better output from Pega Blueprint. For ex legacy application media(Kindly refer screenshot attached) will use Antrhopic Clause LLM etc, Looking forward for your response