Skip to content
Understanding RAG: Why Your Business Needs Better AI Responses episode artwork

Oct 31, 2025 29 min

Understanding RAG: Why Your Business Needs Better AI Responses

from The Web Talk Show Hosted by Armando J. Perez-Carreno

AI chatbots can make up answers when they don't have the right data. Armando explains how retrieval augmented generation (RAG) grounds AI answers in your own business knowledge base so you get fewer hallucinations and more reliable responses. In this episode: If you give an AI a document that's too big for its context window, it can start losing data and making things up. Plain semantic search works well for general questions but can miss exact part numbers or serial numbers, so hybrid search adds full text search to catch them. Adding context to each chunk, like the section heading it sits under, keeps the AI from saying tennis elbow is covered when it's actually listed under excluded conditions. A reranker such as Cohere sorts the search results so the LLM gets only the few pieces that best match the question. In the numbers Armando shared, failed retrievals dropped from 5.7% with regular embeddings to 1.9% with all the steps combined. Chapters 00:00 What RAG means for your business 01:57 How AI tools added web search 03:08 Grounding AI in your business data 04:17 Context windows and hallucinations 07:20 How vector stores work 09:19 Where semantic search misses 11:36 Building your own vector store 17:43 Hybrid search for exact terms 20:04 Adding context to each chunk 23:12 Error rates and reranking Episode page: https://webtalk.show/episodes/036-understanding-rag-why-your-business-needs-better-ai-response/ Hosted by Armando J. Perez-Carreno. Learn more at https://perezcarreno.com

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

Understanding RAG: Why Your Business Needs Better AI Responses

0:00 29:45

No transcript yet.

Requests help us decide what to transcribe next.

Frequently asked questions

How long is this episode of The Web Talk Show?

This episode is 29 minutes long.

When was this The Web Talk Show episode published?

This episode was published on October 31, 2025.

Can I download this The Web Talk Show episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.