# Why do AI agents fail even when the underlying model is good? — retained record

> Retained reddit record from reddit:AI_Agents.

Canonical: https://highsignal.app/data/records/4f70d4e0ae6228bc

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:AI_Agents)  published 2026-10-09 09:45 ingested 2026-10-10 02:34

[original source ↗](https://www.reddit.com/r/AI_Agents/comments/1x1h3p9/why_do_ai_agents_fail_even_when_the_underlying/)

## retained text



```
A pattern I've noticed in enterprise AI projects:

The model isn't necessarily the main problem.

The failure often happens around it.

Wrong context.

Missing information.

Bad tool descriptions.

Too many permissions.

Unclear state.

No evaluation.

No fallback.

So an agent architecture starts looking less like:

**LLM → answer**

and more like:

**Context → LLM → Tools → Validation → Execution → Observability → Human escalation**

I'm curious what failure mode has caused the most trouble in your agent projects.

**Model quality, context, tools, permissions, or something else?**
```



document fetched 2026-10-10 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1x1h3p9/why_do_ai_agents_fail_even_when_the_underlying
