Integrity Score
How accurately AI describes your brand.
What the Integrity Score measures
Your Integrity Score (0–100) measures how accurately AI describes your brand in a scan. A higher score means AI is getting your brand facts right. A lower score means AI is stating things that contradict your truth record or can't be verified.
Why it matters
An AI model that mentions your brand but gets your pricing wrong, attributes a competitor's feature to you, or states the wrong founding year isn't helping you — it's creating misinformation.
The Integrity Score measures this problem in a number you can track over time. Use it to see whether your efforts to fix AI inaccuracies are actually working.
What a good score looks like
| Score | What it means |
|---|---|
| 80–100 | AI is describing your brand accurately |
| 60–79 | Minor inaccuracies — check your Actions for what to fix |
| 40–59 | Significant inaccuracies detected — this needs attention |
| 0–39 | Major inaccuracies — AI is getting important facts wrong |
First-scan Integrity Scores are often lower than 60. This is normal. The score improves as you add more facts to your truth record and as your website content becomes clearer and more authoritative. Track the trend over several weeks rather than reacting to a single scan result.
What affects your Integrity Score
Your Integrity Score is based on five factors combined into a single number:
Presence — Whether your brand was detected in the response at all.
Prominence — How prominently your brand was mentioned when detected.
Accuracy — How many of the checked statements matched your truth record. This is the most heavily weighted factor.
Citation — Whether AI included links to your official pages in the response.
Verdict weight — The overall verdict contributes a weight to the final score. An Accurate verdict adds more than an Unverified verdict.
Integrity Score per scan vs. overall
Every individual scan result has its own Integrity Score — shown in Scan Summary and in each Audit Ledger entry.
Your Overview shows an overall Integrity Score based on recent scans. This is a rolling average that smooths out single-scan fluctuations.
Use the per-scan score to evaluate a specific response. Use the overall score in Overview to measure your trend over time.
How to improve your Integrity Score
Add more facts to Brand Hub. The accuracy factor is the most heavily weighted component. The more specific, checkable facts you have in your truth record, the more precisely AIVIS.space can evaluate AI responses.
Publish clearer information on your website. AI learns from published content. If your website has clear, structured information about your pricing, features, and other key facts, AI is more likely to state those facts correctly.
Fix the highest-priority Actions first. Go to Actions and start with Critical and High priority Accuracy Fix actions. These represent the inaccuracies with the biggest impact on your score.
Add forbidden claims. If AI repeatedly states something your brand doesn't do or offer, add it as a forbidden claim in Brand Hub's Verdict Engine section. Every time AI states it, AIVIS.space will flag it immediately.
Why scores vary across AI models
ChatGPT, Claude, and Gemini are trained differently. The same query may produce an Accurate verdict in one model and a Hallucinated verdict in another.
Check AI Model Visibility to see your Integrity Score broken down per model. If one model consistently produces lower scores, focus your attention on fixing the specific inaccuracies that model is producing.
Integrity Score vs. Visibility Score
| Integrity Score | Visibility Score | |
|---|---|---|
| Measures | How accurately AI describes you | How often AI mentions you |
| What it tells you | Whether AI gets your facts right | Whether AI includes you in the conversation |
| What improves it | Better facts in Brand Hub + clearer website content | More authoritative content + broader topic coverage |
| Main risk | Low score = AI misinforming people about your brand | Low score = AI not mentioning your brand at all |