We ran 1,300+ proctored sessions.
Here's what production data shows.
Aggregated results from live recruiter-led assessments — not projections. Skill depth, session integrity, and verified TalentProof scorecards at scale.
Totals exclude test-user accounts and matching guest sessions.
1,300+
Proctored sessions
11m
Average assessment time
30
Paper Tigers (5% of scored)
520
Real experts verified
The Problem
Recruiters and hiring teams across 7 published roles received hundreds of applications — polished resumes, strong keywords, confident cover letters. After phone screens and interviews, the same pattern kept showing up:
“Candidates who looked great on paper couldn't answer basic technical questions in the live interview.”
Engineering and recruiter time was being spent on people with the right keywords but not the right skills — and with AI-assisted take-homes, resume screening alone couldn't tell the difference.
The Approach
Applicants on published recruiter-led roles completed TalentProof's AI-proctored assessment — 859 applications across 1,703 role views. Each candidate:
Completed a role-specific assessment
AI-generated questions tailored to the actual job requirements, validated by a second AI for accuracy and relevance.
Was monitored by AI proctoring
Tab switches, copy-paste events, face presence, and answer timing were tracked in real time — no browser extensions required.
Received a verified TalentProof scorecard
Every completed session produced a score combining skill depth and session integrity — a single, shareable result the hiring team could trust.
Average assessment time: 11 minutes per candidate. 48% of all sessions reached completion;553+ verified scorecards issued to date.
What AI Proctoring Actually Caught
Every number below is from production proctored sessions. Without proctoring, these signals would have been invisible to hiring teams.
Copy-paste events
Large text blocks pasted into answer fields — a strong signal of AI-generated or pre-written responses.
Tab switches detected
Candidates leaving the assessment window during proctored sessions — often looking up answers or consulting external tools.
Session flag rate
24 of 1,395 sessions flagged for integrity review — surfaced before recruiters invest interview time.
Paper Tigers caught
High skill scores with low authenticity — candidates who knew the material but didn't produce their answers honestly.
The integrity funnel filtered 35% of candidates (282) before they reached a trustworthy shortlist — signals that resume screening alone would have missed entirely.
The Paper Tiger Problem
The most dangerous finding wasn't the obvious cheaters — it was the 30 candidates we classify as “Paper Tigers” (5% of all scored candidates).
Skill score
High (top quartile)
Session integrity
Low (flagged)
These candidates gave technically correct answers — but proctoring data showed they likely didn't produce those answers themselves. High paste counts, suspicious timing patterns, or identity inconsistencies all pointed to the same conclusion.
Without proctoring, Paper Tigers advance to interviews looking like top performers on a traditional assessment — wasting engineering hours and interview slots on candidates who crumble under live scrutiny.
The Outcome
Instead of screening 859 applications blind, hiring teams reviewed 554 verified scorecards — and prioritized 520 real experts with high skill and high authenticity.
64%
Application-to-scorecard rate
554 of 859 applicants completed and received a verified scorecard on published roles.
11m
Per candidate cost
No recruiter time required for initial screening. AI generated, proctored, and scored every assessment automatically.
520
Real experts identified
High skill + high authenticity — each with a shareable verified TalentProof scorecard the hiring manager could review before the call.
Spotlight: Front End Developer — 295 applications, 208 completed, 71% completion rate.
Recruiter Time Saved Per Role
Modeled recruiter active time on a role: skim the top of the applicant pile and phone-screen a shortlist today vs async intake, scorecard review, and fewer calls with TalentProof. We do not assume reading every resume (that upper bound is ~12.5 hrs on our avg role — not typical).
Small active req (~10 applicants)
~0.8 hrs
Today ~2.4 hrs → with TalentProof ~1.6 hrs (skim + phone model).
100-applicant example
2–3 hrs
Today ~5.8 hrs (30 skims + 8 calls) → ~3.4 hrs with scorecards + 5 calls.
Avg published role (~123 applicants)
~3 hrs
Today ~7.1 hrs → with TalentProof ~4.1 hrs (same model).
Modeled recruiter active time only — excludes avoided bad submittals, Paper Tiger live screens, and client rework. In-app pipeline estimates use a separate admin formula (auto-screen + verified signals).
What This Means for Your Hiring
Every candidate gets the same fair test
AI generates role-specific questions validated by a second AI. No bias from resume formatting, school names, or employer brands.
Cheating is visible, not hidden
You see exactly what happened during the session — tab switches, paste events, timing anomalies — and decide what matters for your role.
Your team interviews fewer, better people
Engineering and hiring manager time is the most expensive part of hiring. Stop spending it on candidates who can't do the work.
Setup takes minutes, not weeks
No ATS integration required. Create a role, send a link, review scorecards. Your first assessment can be live in 5 minutes.
See It With Your Own Candidates
Your first role is free. Create an assessment, send it to your applicants, and see the same integrity data in your own pipeline.
No credit card required. First role free.