Key Benefits
A signal, not a verdict.
Why it works, and how it saves agencies.
The problem
What bad hires actually cost.
Stealth copilots are common enough now that most agencies have almost certainly already placed a candidate who used one — they just didn't have the evidence to catch it before the offer went out.
Zero-friction by design
Nothing for your candidate to install.
Most proctoring and cheat-detection tools force candidates into a restricted environment — a lockdown browser, a separate testing portal, a webcam-monitoring app they have to download before they can even start. That friction alone puts good candidates off, and it still doesn't catch a live copilot running quietly on a second screen.
Candit needs none of it. You join meetings exactly like you already do — same link, same video call tool, nothing for the candidate to open, agree to, or install. Detection is built in and runs the whole time, completely in the background.
Traditional proctoring
- Lockdown browser or dedicated testing portal
- Webcam/screen-monitoring app to install
- A separate step outside the actual interview
- Still misses a copilot on a second device
Candit
- Joins the meeting like any other participant
- Nothing for the candidate to install or open
- Runs the whole interview, completely in the background
- Reads the transcript itself — no device to miss
Reading the score
Where the real threshold sits.
- Under 90%
- Not a clear match yet — a good cue for your recruiter to ask one more question before moving them forward.
- 90% and above
- A high-confidence match against a live AI copilot — roughly what a bad-hire refund would've cost your agency.
More benefits
What else you get.
- 01Real-time, not a post-call report
The flag lands while the interview is still happening — early enough for your recruiter to ask a follow-up, not read about it after the candidate's already gone.
- 02Per speaker, with evidence
Every flag is tied to the exact line that triggered it and who said it — never a single blended score with no way to check the reasoning.
- 03Scored per candidate, too
Every scorecard also carries an authenticity score across the whole interview. Below 70% is worth a look; above that, most agencies don't think twice.
- 04Tested weekly against real copilots
Checked against the same stealth copilots candidates actually use, every week — not tuned once at launch and left to drift.
- 05One tool, not two
Detection is built into the same session you're already recording notes in — nothing extra to buy, install, or hand your candidates.
Why not just any notetaker
Notes are the baseline, not the whole job.
A plain notetaker transcribes and summarizes — useful, but it has nothing to say about whether an answer was actually the candidate's own. Bolting a separate proctoring tool on top brings back all the friction above, for a second subscription. Candit is the one tool that already does both: real meeting notes, and the only detection built to catch what those other tools can't.
Catch it while the interview is still on.
AI Detectability is included on Plus, Pro and Team. Start free, upgrade when you need it.