---
title: "Monitoring & Diagnosis"
lang: "en"
canonical: "https://winin.ai/en/blog/topics/measurement/"
alternate: "https://winin.ai/zh/blog/topics/measurement/"
---

# Monitoring & Diagnosis

Has the recommendation rate really changed?

Turn isolated answers into a baseline, separating mentions, recommendations, citations and change.

## Start with this guide

[How to monitor AI recommendations: from question set to comparable retest](https://winin.ai/en/guides/same-conditions-retest/)

Same-conditions retest means measuring again with the same questions under the same conditions, so that before and after can be compared, while statin

## Explore specific questions

### [ChatGPT brand monitoring tools — how to choose](https://winin.ai/en/answers/chatgpt-brand-monitoring-tools/)

Many “ChatGPT brand monitoring / AI visibility” tools exist, ranging from mention-rate dashboards to operating platforms with retest. Check whether th

### [Brand visibility in ChatGPT explained](https://winin.ai/en/answers/chatgpt-brand-visibility-explained/)

Brand visibility in ChatGPT means: under a given prompt and sampling setup, whether your brand/product is mentioned, shortlisted, described accurately

### [What is the risk of consumer-brand absence in AI recommendations?](https://winin.ai/en/answers/consumer-brand-absence-risk/)

More consumer decisions shift from “search it” to “ask it.” Persistent absence on category prompts means you lose before the consideration set forms ,

### [How do I know if I’m missing from AI category recommendations?](https://winin.ai/en/answers/detect-category-recommendation-absence/)

Detect absence with fixed category prompts + multi-model sampling + archived comparison , not occasional manual chats. Prompts should mirror real buye

### [What Formal Baseline means for AI recommendation / brand visibility measurement](https://winin.ai/en/answers/formal-baseline-ai-recommendation/)

A Formal Baseline is the operating-grade, retestable scope for measuring AI recommendation and brand visibility—custom prompts, model sets, and cadenc

### [What are the four bad outcomes for brands in AI answers?](https://winin.ai/en/answers/four-bad-outcomes-ai-answers/)

Winin’s homepage illustrates four outcomes for the same pre-purchase question:

### [What are generative AI brand visibility risks?](https://winin.ai/en/answers/generative-ai-brand-visibility-risk/)

Primary risks include absence (missing from category answers), wrong (stale prices/specs under your name), mentioned but not recommended (name without

### [How to choose AI brand monitoring tools](https://winin.ai/en/answers/how-to-choose-ai-brand-monitoring-tools/)

Don’t stop at “can it show mention rate.” Verify: (1) model coverage for the CN/Global surfaces you care about; (2) drill-down to raw answers; (3) sam

### [Is mention rate enough for GEO?](https://winin.ai/en/answers/is-mention-rate-enough-for-geo/)

No. Mention rate only answers whether your brand is named in an answer. Operating-grade GEO also separates shortlist (are you considered), accurate & 

### [How do you measure ChatGPT brand visibility?](https://winin.ai/en/answers/measure-chatgpt-brand-visibility/)

Measure with fixed prompts + scheduled sampling + structured labeling + same-conditions retest . At minimum score: mentioned, shortlisted, accurate, r

### [Why does AI mention my brand without recommending it?](https://winin.ai/en/answers/mentioned-but-no-recommendation-reason/)

This is classic mentioned-not-recommended : Reach or name-dropping occurred, but Accurate & Recommended did not—models may narrate you while awarding 

### [Monitor-only vs optimization loop — which do you need?](https://winin.ai/en/answers/monitor-only-vs-optimization-loop/)

It depends on the job. If you only need “were we mentioned this week,” monitor-only may suffice. If you must answer “why not shortlisted, which fact t

### [What do Reach, Shortlist, Accurate & Recommended mean?](https://winin.ai/en/answers/reach-shortlist-accurate-recommended/)

These are Winin’s public AI recommendation stage gates : a brand can be discovered, considered, then accurately recommended. Homepage summary: Reach s

### [Reach · Shortlist · Accurate & Recommended](https://winin.ai/en/glossary/reach-shortlist-recommended/)

Definition : Three stage gates for brands inside generative answers.