Every review and report, readable by AI
When someone asks an AI assistant whether a company is reliable, the answer should come from verified reviews and official records, with dates. OpenDossier publishes every company profile, its ratings, its review summaries and its report summaries in the forms assistants read and the tools they call, and tells companies what the assistants actually say.
| What is published | Where | What is not |
|---|---|---|
| Profile, ratings, review counts and dimension scores | HTML with schema.org Organization and AggregateRating; profile.md and profile.json per company; llms.txt | Reviewer identities, ever |
| Review text of the reviews shown free on the profile | Same, with dates and verification type | Reviews behind the give-to-get wall (summarised as themes only) |
| Company responses | Next to each review | Private messages |
| Report summaries: key findings, open-question count, coverage | Profile and machine twins | Full report detail (sold to buyers; available to licensed API and MCP users) |
| Official-record claims with the evidence chain | Dossier pages and dossier.md / dossier.json | Company-supplied private documents |
How assistants reach it
Crawl and search
robots.txt explicitly allows GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot and CCBot; the sitemap and IndexNow are kept current; every profile has a Markdown alternate.
Tools
The OpenDossier MCP server and OpenAPI let assistants and agents query live profiles and receive sources, dates and coverage with every answer.
Monitoring
For companies on Profile Pro and above: a fixed set of questions asked monthly of the major assistants, the answers and citations recorded, errors flagged, share of voice against chosen competitors.
Correction loop, honestly
An error is corrected by publishing better evidence and re-asking. Nothing accurate is hidden, nothing is seeded under another name, no reviews are generated.