Foundational public methodology for the first open public ranking of brand
visibility in AI search results (ChatGPT, Perplexity, Gemini, Claude).
This release establishes the framework — no rankings have been computed
or published yet. First scan cycle: late May 2026 (private validation).
First public ranking publication target: August 2026, after 3 validation
cycles.
Includes:
- methodology.json: machine-readable formulas, weights, policies
- README.md: human-readable overview + open/closed boundary
- CHANGELOG.md: versioning policy + v1.0.0 release notes
- taxonomy.md: tier system + 11 PL pilot categories
- LICENSE: MIT
- .gitignore: closed operational data (exact prompts, anti-gaming thresholds)
- prompts/README.md: 6-stage prompt curation process
- prompts/example-swiece-sojowe-pl.md: illustrative framework for first category
Strategic principles:
- Algorithm-first, no advisory board
- Open methodology + closed exact prompts (Goodhart's Law defense)
- No retroactive changes (FIDE 2024 lesson)
- No pay-to-play, hard rule (Moody's / Forbes 30 Under 30 lessons)
- Subjective opinion disclaimer (Gartner v. NetScout 2020 First Amendment shield)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
