Topic 5: RAG in Product Management & Business Analysis
RAG in Product Management & Business Analysis
Product Managers and Business Analysts spend a lot of time working with information. Customer research sits in one place, requirements in another, policies somewhere else, and important decisions may be buried across documents and systems.
RAG can help by making this knowledge easier to find, connect and use with AI.
Instead of manually searching through every document, a RAG-enabled system can retrieve relevant information and give it to an LLM as context for answering your question.
Explore customer and user research
PMs and BAs often have large amounts of qualitative information from interviews, surveys, support conversations, feedback and research reports.
RAG can help you ask questions across this information, such as:
“What problems are customers mentioning most often during onboarding?”
“What reasons are customers giving for abandoning the process?”
The system can retrieve relevant pieces of research and use them to help the LLM identify patterns, themes and insights.
This does not replace proper research analysis, but it can make exploring a large body of research much faster.
Work with requirements
Requirements can be spread across BRDs, PRDs, user stories, acceptance criteria, business rules and previous project documentation.
A BA could ask:
“What existing business rules could affect this new requirement?”
A PM could ask:
“Do we already have requirements related to this feature?”
RAG can retrieve relevant information from approved requirements and project documents, giving the LLM better context for helping you analyse the problem.
Navigate product knowledge
As products grow, so does the amount of information around them.
You may have roadmaps, specifications, release documentation, user guides, decision logs and knowledge articles across different systems.
A RAG-enabled product assistant could help teams ask questions such as:
“What functionality currently exists for account recovery?”
“What decisions have previously been made about this feature?”
This can make product knowledge easier to access, particularly for people who are new to the team or unfamiliar with the history of the product.
Understand policies and processes
Many product decisions are influenced by organisational policies, business processes, regulations and operating rules.
Instead of manually searching through several documents, a BA could ask:
“Which policies apply to this customer verification process?”
RAG can retrieve relevant sections from trusted sources and give them to the LLM to help explain what may need to be considered.
For regulated or high-impact decisions, the original policy or source should still be checked.
Support product decisions
RAG becomes particularly interesting when relevant information exists across several different sources.
Imagine a PM exploring whether to improve a product feature. Useful context might exist across:
Customer research + Product analytics + Requirements + Previous decisions + Product documentation
A RAG-enabled assistant can help retrieve relevant information from these sources and bring it together as context for exploring the decision.
The AI can help you understand the information, compare options or identify questions that need further investigation.
The decision itself still belongs to the product team.
Improve access to internal knowledge
One of the broader uses of RAG is creating internal AI assistants that help employees find organisational knowledge.
Instead of asking:
“Where is that document?”
or
“Who knows how this process works?”
people can ask the question directly.
For PMs and BAs, this could make it easier to understand existing products, processes, requirements and previous decisions without spending as much time searching across multiple systems.
Keep the human in the loop
RAG gives an LLM better information, but it does not automatically make the answer correct.
The system might retrieve incomplete information. A source might be outdated. The LLM might misunderstand what it receives.
Before using an important output, check:
- whether the right sources were used
- whether the information is current
- whether important context is missing
- whether the AI has made assumptions
- whether the conclusion is supported by the source information
This becomes even more important when the output affects customers, compliance, investment or important product decisions.
DSM Tip
RAG is most useful when your work depends on finding the right information from many trusted sources.
Use it to reduce the time spent searching and bring relevant knowledge into your work faster. Then apply your own product or business judgement to decide what that information actually means.
