What should an AI visibility export contain?

A visibility score is easy to export. An audit trail is harder—and much more useful.

For every metric, I think a portable AI-visibility export should include:

• the exact prompt and benchmark version

• provider, model, locale, parameters, and acquisition time

• the complete raw answer

• raw citation URLs in their returned order

• resolved and canonical URLs without overwriting the raw values

• explicit timeout, block, parsing, and resolution failures

• metric eligibility plus numerator and denominator

• parser, normalizer, matcher, and scoring versions

Scores should be projections over this evidence, not the only thing that survives outside the dashboard.

A team should be able to load the export into a warehouse, replay a newer parser against preserved answers, recalculate the metrics, and explain every material change.

I wrote the fuller data-contract proposal here:

https://corank-ai-search.hashnode.dev/a-minimal-data-contract-for-ai-visibility-exports

At Corank, we are applying the same evidence-first approach to AI-search visibility:

https://corank.ai

For teams that consume these reports: would you rather receive a set of related tables, newline-delimited JSON, or both?