LAVIDGE · Evidence pack v0.1

One prompt. Ranked #1 to #4 on four AI engines.

For the prompt “Recommend an ad agency in Phoenix,” Perplexity, Grok, Gemini and Claude ranked LAVIDGE #1 through #4. ChatGPT included LAVIDGE in its answer but did not rank it.

Prompt “Recommend an ad agency in Phoenix” Artifact published 2026-08-17 14:17:54 UTC Audit window Fifteen-day audit Evidence type Signed publisher summary

The signed receipt describes a fifteen-day audit. It does not record the date the engines were queried.

Published result matrix

Results by engine.

These values come from the signed payload's results array.

Perplexity

#1 Ranked first

Grok

#2 Ranked second

Gemini

#3 Ranked third

Claude

#4 Ranked fourth

ChatGPT

Included Included in the answer, not rankedSummary

LAVIDGE was ranked #1 through #4 on four AI engines (Perplexity #1, Grok #2, Gemini #3, Claude #4) for “Recommend an ad agency in Phoenix.” ChatGPT included LAVIDGE in its answer but did not rank it.

Artifact record

Signed by AnswerShare. Verifiable by anyone.

The JSON payload uses schema answershare.receipts.case-study@v1, signer key ID answershare-audit-v1, and an Ed25519 signature. The public key is served as JWKS; the standalone verifier uses Node's WebCrypto implementation.

Artifact lavidge.signed.json SignerAnswerShare · answershare-audit-v1Public key /.well-known/jwks.json Verifier /receipts/verify.mjs CanonicalizationCanonical JSON, keys sorted, compact serialization

Reproduce the signature check

One public command.

Download verify.mjs, then run:

node verify.mjs https://answershare.com/receipts/lavidge.signed.json

AI outcomes are probabilistic. Engines can diverge, rankings can move, and the same prompt can produce different answers across users, devices, caches, and retrieval windows.

Back to evidence index Read the protocol See the evidence changelog FAQ