---
title: "Discussion on \"Where RAG Fails: Understanding the Limitations of Retrieval-Augmented Generation\" | Hashnode"
description: "Discussion on \"Where RAG Fails: Understanding the Limitations of Retrieval-Augmented Generation\". Discover why Retrieval-Augmented Generation (RAG) fails. Explore chunking issues, retrieval gaps, and when RAG isn't the right solution."
url: "https://hashnode.com/posts/where-rag-fails-understanding-the-limitations-of-retrieval-augmented-generation/6a5d0b1909c5af167ab31f43"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Where RAG Fails: Understanding the Limitations of Retrieval-Augmented Generation](https://darshan-pawar.hashnode.dev/where-rag-fails-understanding-the-limitations-of-retrieval-augmented-generation?utm_source=hashnode&utm_medium=feed)

Discover why Retrieval-Augmented Generation (RAG) fails. Explore chunking issues, retrieval gaps, and when RAG isn't the right solution.

Jul 19 · 4 min read · darshan-pawar.hashnode.dev  
Read the article

-   [#ai](https://hashnode.com/tag/ai)
-   [#rag](https://hashnode.com/tag/rag)
-   [#llm](https://hashnode.com/tag/llm)
-   [#chaicode](https://hashnode.com/tag/chaicode)

## Written by

[Darshan Pawar](https://hashnode.com/@darshanpawar)

@darshanpawar

  
[View profile](https://hashnode.com/@darshanpawar)

Stop thinking. Start executing.

## Responses 1

Markdown supported · ⌘↵ to post

[Ahmet Özel](https://hashnode.com/@ahmetozel)

[Jul 21](https://hashnode.com/posts/where-rag-fails-understanding-the-limitations-of-retrieval-augmented-generation/6a5d0b1909c5af167ab31f43/comment/6a5f0a495fe50ce983807475)

Nice breakdown — parent-child ingestion plus a reranker are exactly the two upgrades that move the needle first.

One "where RAG fails" case those two don't fix, worth adding: when the answer simply isn't in the corpus, the model still confidently stitches one together from the nearest chunks. Precision fixes don't catch that — you need a groundedness check and permission for the system to say "not in the docs."

And on rerank-top-25: the reranker can only reorder what first-stage retrieval already pulled. If recall missed the right chunk, no cross-encoder recovers it — so first-stage recall@k is the real ceiling, and it's worth measuring separately from final answer quality.