Before you publish an AI accuracy claim, check whether its test result supports the number in the sentence. In our synthetic Nimbus example, a website promises 98% accuracy for AI content detection. The supplied test records 53% on a mixed human/AI text sample. ClaimLens flags that mismatch under its configured rules. The useful check goes beyond comparing two numbers: which task and inputs did the test cover, and does the wording promise more? Keep those details with the claim so the next reviewer can assess its support. This demo checks supplied records. Validating the tests and judging the final wording still needs human review. #AIClaims #MarketingCompliance #AIGovernance