---
title: "Discussion on \"Before Adding an LLM, Define What a Good Answer Looks Like\" | Hashnode"
description: "Discussion on \"Before Adding an LLM, Define What a Good Answer Looks Like\". Explore evaluation-first AI design through an SRE playbook finder: define useful answers, test failure cases, and decide when to abstain."
url: "https://hashnode.com/posts/evaluation-first-before-adding-an-llm/6af317e8ecd7fc3e101da6e6"
---

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

Discussion

# [Before Adding an LLM, Define What a Good Answer Looks Like](https://blog.codapus.studio/evaluation-first-before-adding-an-llm?utm_source=hashnode&utm_medium=feed)

Explore evaluation-first AI design through an SRE playbook finder: define useful answers, test failure cases, and decide when to abstain.

Nov 3 · 6 min read · blog.codapus.studio  
Read the article

-   [#ai-engineering](https://hashnode.com/tag/ai-engineering)
-   [#llm](https://hashnode.com/tag/llm)
-   [#site-reliability-engineering](https://hashnode.com/tag/site-reliability-engineering)
-   [#software-engineering](https://hashnode.com/tag/software-engineering)
-   [#testing](https://hashnode.com/tag/testing)

## Written by

[Sotiris Zacharopoulos](https://hashnode.com/@sz-codapus-studio)

@sz-codapus-studio · Dublin, Ireland

  
[View profile](https://hashnode.com/@sz-codapus-studio)

## Responses

Markdown supported · ⌘↵ to post

No responses yet.
