Trust score
One trust level per candidate, with the reasons behind it. It flags candidates for your review and never rejects anyone.
Each application gets one trust level — Low risk, Some risk or High risk — with a plain-words reason for every point that counts. You see it on the candidate page under Identity and trust, and as the reasons in the trust review queue (Vouch → Flagged applications).
The trust score flags for your review. It never rejects a candidate, never changes their status and never moves them in your pipeline.
What goes into it
| Part | Adds points when |
|---|---|
| Application checks | The application looked unusual when it came in (for example, many applications from one device or a throwaway email) |
| Basic identity check | The Basic check came back Some risk or High risk |
| Full ID check | The full ID check did not pass. A passed check lowers the score |
| Journey | The journey was taken through a VPN or proxy, or the location jumped between visits |
| References | Reference answers look unreliable (for example, the referee is on the candidate's own network) |
| Verify | The candidate's contact details could not be cross-checked |
| Resume | The resume integrity check found something worth a look (below) |
The score runs from 0 to 100. 25 or more is Some risk; 50 or more, or a failed full ID check, is High risk.
Resume integrity check
After a resume is scanned, InsightHire looks for things a person should double-check:
- Overlapping jobs — two full-time jobs overlapping for more than three months.
- Impossible dates — an end date before the start date, or a start date in the future.
- A more senior title than the employer record from Verify shows.
- Copied text — much of the resume matches another applicant's resume at your organization.
Each reason adds a little to the score, with a cap, so a resume alone can't make someone High risk. There's no on-screen switch for the resume check. If you want it off for your organization, ask InsightHire support.
Why there is no "AI-written" check
InsightHire does not try to detect whether a resume or answer was written with AI. Those detectors are unreliable and wrongly flag people who write in a second language or a formal style. Using an AI writing tool also isn't dishonest in itself. We check facts that can be wrong — dates, overlaps, titles and copied text — instead.
What to do with a flag
Open the candidate and read the reasons. Most have an innocent explanation: a contract job alongside a full-time one, a shared office network, a typo in a date. Ask the candidate, request a full ID check if you need one, then have an org admin record the outcome in the trust review queue with Confirm concern or Dismiss false positive.
