Verdicts
What Accurate, Hallucinated, Competitor bias, and other verdicts mean.
What a verdict is
A verdict is the overall result of a single AI response check. Every time AIVIS.space scans your brand, it produces one verdict per AI response — based on comparing what AI said against your truth record.
The verdict tells you whether AI described your brand accurately, got something wrong, or couldn't be fully evaluated.
The six verdicts
Accurate
AI described your brand correctly. The statements AIVIS.space checked matched the verified facts in your truth record.
What this means for you: No action required for this response. Continue monitoring — an Accurate verdict today doesn't guarantee accuracy in the next scan.
Integrity Score: Highest range. An Accurate verdict typically produces an Integrity Score of 70–100.
Hallucinated
AI stated something about your brand that contradicts your truth record or cannot be verified from any source.
Examples:
- AI states your Starter plan costs $99/month when your truth record says $39/month
- AI says you were founded in 2010 when your truth record says 2018
- AI says you offer a feature your truth record lists as a forbidden claim
What this means for you: This is the most serious verdict. AIVIS.space generates an Accuracy Fix action with the exact inaccurate statement and what the correct information should be.
Integrity Score: Lowest range. A Hallucinated verdict typically produces an Integrity Score below 50.
Competitor bias
AI recommended a competitor in your place when someone asked a question where your brand should appear.
Example: Someone asks "what's the best accounting software for freelancers?" and AI recommends a competitor without mentioning your brand — even though your brand directly serves freelancers.
What this means for you: AIVIS.space generates a Competitive Threat action. This is an AI visibility issue, not a factual inaccuracy. Publishing clearer content about the topic may help AI mention your brand in future responses.
Integrity Score: Variable — depends on whether AI mentioned your brand at all and how accurately.
Mixed signals
Some statements about your brand were accurate and some were not. AI got part of it right but also included at least one inaccuracy or unverifiable claim.
What this means for you: Check the evidence in the Audit Ledger (Growth and above) to see which specific statements were accurate and which were not. AIVIS.space generates Accuracy Fix actions for the inaccurate parts.
Integrity Score: Middle range, typically 40–70.
Unverified
AIVIS.space couldn't confirm or deny what AI said. This usually means the topic AI discussed isn't covered in your truth record.
Example: AI says your product "integrates well with enterprise workflows" but your truth record doesn't include any information about enterprise integrations. AIVIS.space can't confirm this is true or flag it as false.
What this means for you: This verdict signals a gap in your truth record. Open Brand Hub → Verdict Engine and add facts that cover the topic AI discussed. AIVIS.space generates a Truth Gap action to guide you.
Integrity Score: Middle range.
Insufficient evidence
Your truth record doesn't have enough information for AIVIS.space to produce a meaningful evaluation of this AI response.
What this means for you: Add more facts to the Verdict Engine section in Brand Hub. A minimum of 5 specific, checkable facts is needed to evaluate most AI responses accurately. AIVIS.space generates a Truth Gap action pointing you to the relevant section.
Integrity Score: Lower range.
How verdicts are produced
AIVIS.space reads the AI response and identifies every statement that can be checked against your truth record. It then compares those statements fact by fact and determines the overall verdict based on what it found.
The Audit Ledger (Growth and above) shows the full evidence trail for every verdict — the exact statements found, the facts they were compared to, and the comparison results.
Verdicts and Integrity Score
A verdict is a category. An Integrity Score is a number. They're related but different.
- An Accurate verdict with a high Integrity Score (90+) means AI was both correct and comprehensive
- An Accurate verdict with a lower Integrity Score (70–79) means AI was correct but didn't cover many of your facts
- A Hallucinated verdict can still have a partial Integrity Score if some parts of the response were accurate
Use the verdict to understand the category of result. Use the Integrity Score to track improvement over time.
Normal verdict patterns
On your first few scans, Unverified and Insufficient evidence verdicts are common. AI covers a lot of ground in its responses, and a new truth record may not have facts for every topic AI brings up.
As you add more facts, Unverified verdicts should decrease. You'll see more Accurate verdicts and, if AI is getting things wrong, more Hallucinated verdicts.
A mix of verdicts across models is also normal. ChatGPT, Claude, and Gemini have different training data and may produce different verdicts for the same query.