What they actually interview on, with the grounding shown.
Company Intelligence researches an employer from its own careers and about pages, classifies every section of the result as researched, from the job description, or estimated, and caches it so the next person asking about the same company does not pay for it again.
What is Company Intelligence?
Company Intelligence is a researched profile of an employer: what they interview on, per round, where they actually hire, how their process runs, and what they say about themselves. It is built by fetching the company's own site rather than a paid search vendor, every section is labelled with how well grounded it is, and profiles are cached and versioned so an improvement to the pipeline reaches companies already on file.
Named mechanisms, not adjectives.
Every one of these is implemented and can be pointed at. On a product whose pitch is that claims get checked, the description of it cannot be the one place that is not.
Grounding computed, not claimed
Each section is labelled researched, from the job description, or estimated, and the label is derived from what was actually found, not accepted from the model.
Fetched from the source
Research pulls the company's own careers and about pages and records what was fetched, including the case where something came back that was not this company.
Real structured data where it exists
Address, employee count and the real external site come from the structured data companies publish, not from a model's memory.
Cached and versioned
A profile is reused across everyone asking about the same company for thirty days, and a pipeline version gate means a better pipeline re-researches companies already on file.
Pages that stand themselves up
When the job pipeline sees an employer it has not met, a company page is created automatically from real structured data, so every role links to a real company.
In practice.
- Know what a company tests for before you apply, with the confidence shown.
- Every role on the board links to a company page, and the page tells you where it got its facts.
- Track dream companies and see your fit recomputed as your record grows.
Against Glassdoor and company review sites.
| insiderOne | Glassdoor and company review sites | |
|---|---|---|
| Source | The company's own pages and structured data, fetched directly | Anonymous reviews |
| Honesty about what is known | Every section labelled researched, from the job description, or estimated | No distinction |
| Focus | What they interview on and where they hire | Ratings and complaints |
Questions, answered.
Where does the information come from?
The company's own careers and about pages, fetched directly, plus the structured data it publishes. Each section of the profile is labelled with how well grounded it is, and the label is computed from what was found.
How current is a profile?
Profiles are cached for thirty days and re-researched when the pipeline improves, so a company on file does not fall behind a better method.