AI Perception Risk

See when AI leaves you out or gets your business wrong.

AI Perception is not judged by one favorable answer. It shows whether AI recognizes your business, describes it accurately, has evidence to trust it, and can include it when buyers ask relevant questions.

AI Perception is included in Pro and agency plans for verified domains. The free Snapshot checks public website clarity and does not include AI Perception.

Left out or described incorrectlySee the supporting evidenceGuidance, not prediction
Illustrative product viewNorthstar Supply · Pro
AI PerceptionHow do AI systems understand, trust, and surface this business?Selected scan · each result keeps its source and observation date
Score69Workable, with gaps
Direct answerStrong understanding after reviewing evidence; weak recognition without it.
AAwareness31%
UUnderstanding88%
TTrust74%
RReach67%

Illustrative Northstar example, not customer results. AI Perception is included in Pro and agency plans for verified domains.

The practical risk

A business can be missing, misunderstood, or accurately represented.

Those are different states with different causes. Start with what the answer did, then inspect the evidence that may explain it.
Observed answer stateWhat it meansWhat to inspect
Left out

The business fits the question but is not named.

Check whether evidence is easy to find, the business is clearly identified, and credible sources support it.
Described incorrectly

The business appears with the wrong category, services, or audience.

Find conflicting descriptions and confusion with another business.
Accurately represented

The answer reflects the business facts supported by the available evidence.

Keep the facts current and continue observing relevant contexts.

Four questions behind AI Perception

Recognition and understanding are not the same thing.

An AI system may describe a business accurately after reviewing its website while still failing to recognize it on its own. Gemmetric shows those results separately.
AAwareness31%

Do AI systems recognize the business on their own?

Based on observed AI responses in this review.

UUnderstanding88%

Can they describe the business accurately after reviewing its evidence?

Based on observed AI responses in this review.

TTrust74%

Is the description supported by credible public evidence?

Based on how well public sources support the answer.

RReach67%

Which relevant buyer questions could include the business?

Based on the buyer questions the available evidence can support.

Illustrative Northstar scores, not customer results. Each result keeps its source, method, and observation date.

Awareness and Understanding come from observed AI responses. Trust looks at how well the answer is supported. Reach looks at the relevant buyer questions where the available evidence supports inclusion. Each result keeps its source and observation date.

What the score can and cannot do

Use the breakdown to locate the gap, not to forecast the next answer.

A score can summarize the selected evidence. It cannot guarantee mentions, citations, recommendations, rankings, leads, or revenue.
01

Start with the explanation

Which part of the result is holding the business back?

Read the plain-language summary before treating the score as the conclusion.
02

Check the supporting evidence

Which AI response or public source supports each part of the result?

Keep the source, date, and method visible.
03

Choose one useful improvement

Work on the clearest weakness instead of chasing the overall score.

Website clarity, outside proof, and business consistency may require different work.
04

Check again

Review comparable questions and evidence after the change is live.

Look for a repeated pattern, not one favorable response.

Choose the symptom

Was the business left out or described incorrectly?

A missing mention and a wrong description are different problems. Start with what you actually saw in the answer.
Not mentioned

The business is missing from the answer.

Check whether it fits the question, is easy to identify, and has enough visible support.

Understand a missing mention →
Mentioned incorrectly

The business appears with the wrong category or capabilities.

Look for conflicting descriptions, vague definitions, or confusion with another business.

Understand a wrong description →

AI Perception and citation guides

Choose the guide that matches what you saw.

Learn why a business was left out, why it was described incorrectly, how citations work, or what to improve next.

AI Perception questions

Short answers about what AI Perception means.

What the review covers, what the free Snapshot includes, and why one answer never proves the whole pattern.
What is AI Perception Risk?

AI Perception Risk is the possibility that an AI system leaves a relevant business out, describes it inaccurately, or represents it differently across similar questions.

What does Gemmetric mean by AI Perception?

AI Perception looks at four questions: Does AI recognize the business? Can it describe the business accurately? Is the answer supported by credible evidence? And can the business appear for relevant buyer questions?

Does a lower AI Perception score mean every AI system gets the business wrong?

No. The score is a guide based on the evidence, AI systems, methods, and dates used for that review. Individual answers can still vary.

Does the free Snapshot include AI Perception?

No. The free Snapshot checks a sample of public website pages for GEO clarity. AI Perception is included in Pro and agency plans for verified domains.

Start with the website evidence AI can inspect.

Run a free Snapshot to check the clarity of your public website. When you need AI Perception, deeper evidence, and saved history, compare the managed plans.