The CV Is Dead.
ChatGPT broke traditional candidate screening. CandiScan verifies candidates through adaptive video pre-interviews — then gives you a report with evidence, not a mystery score.
Defensible decisions. In 30 minutes.
71%
of HR professionals have encountered false candidate information
Source: Checkr 2025
23 hrs
spent screening per single hire on average
Source: Testlify
€17K
average cost of one bad hire
Source: CareerBuilder
" Monday morning. 347 new applications for your senior developer role. Every CV reads like it was written by the same person. Because it probably was.
CandiScan caught three candidates with fabricated project experience in the first 3 days. The evidence reports made the decision easy — we could show exactly why we passed on them.
Reijo S.
CTO, Fentrica
How it works
Three steps. Thirty minutes. Defensible decisions.
Upload a CV
Upload a candidate's CV or forward the application email. Our system extracts every verifiable claim — job titles, timelines, technical skills, project responsibilities — and identifies where fabrication typically hides.
Takes 30 seconds. Works with any CV format.
Candidate completes a pre-interview
We conduct a 30-minute video interview with the candidate, with questions generated from THEIR specific CV. When they're vague about their "team of 12 at BigCorp", we follow up. When timelines don't add up, we probe deeper. Questions adapt real-time based on their answers.
Same rigorous process for every candidate. No bias, no shortcuts.
Review evidence, make decisions
You get a report with color-coded flags. Each flag shows the exact question, the candidate's response, and why it raised concerns. No mystery scores. No black boxes. One-click to override any flag you disagree with.
Defensible decisions. Average review time: 5 minutes per candidate.
What you'll actually see
Every flag comes with the full story. No mystery scores.
QUESTION ASKED
"You listed leading a team of 12 at Acme Corp from 2019–2021. Can you walk me through a specific project?"
CANDIDATE RESPONSE
"It was more of a collaborative environment... I wouldn't say I was directly managing them..."
WHY FLAGGED
CV states "Led team of 12 engineers" but candidate describes a peer collaboration role. Significant discrepancy in responsibility level.
QUESTION ASKED
"Your CV says you 'architected the payment system' at FinCo. Walk me through the key design decisions you made."
CANDIDATE RESPONSE
"I was involved in the project... my manager made most of the architecture calls, but I implemented parts of it."
WHY FLAGGED
CV claims sole ownership of system architecture, but candidate attributes key decisions to their manager. Role was implementation, not design.
QUESTION ASKED
"You mentioned leading the migration to microservices. What specific services did you break out?"
CANDIDATE RESPONSE
"It was a team effort, really. We followed standard practices... you know, the usual coordination stuff."
WHY FLAGGED
Candidate avoids specifics across 3 follow-up questions. No concrete details provided despite claiming a leadership role.
OBSERVATION
9 long pauses of 10+ seconds detected. Eye gaze suggests looking at a secondary screen.
POSSIBLE INTERPRETATIONS
You decide how to weigh this. We just report what we observed.
These are just 4 examples. The system detects 65+ different flag types across varying severities — every one with reviewable evidence. You make the decision.
Other tools give you a score.
We give you a case file.
What you see
"Score: 34/100"
"I think this looks suspicious"
"Claimed 'led team of 12' but described a peer role. Video at 14:32."
Time per candidate
Instant (unreliable)
45 minutes
5-min evidence review
Bias risk
Up to 61% false positives (Stanford HAI)
Inconsistent criteria
Same process for everyone. You interpret.
Defensibility
"The algorithm said no"
"My gut feeling"
Evidence for legal, leadership, or the candidate
| Black-Box AI | Manual Screening | CandiScan | |
|---|---|---|---|
| What you see | "Score: 34/100" | "I think this looks suspicious" | "Claimed 'led team of 12' but described a peer role. Video at 14:32." |
| Time per candidate | Instant (unreliable) | 45 minutes | 5-min evidence review |
| Bias risk | Up to 61% false positives Source: Stanford HAI | Inconsistent criteria | Same process for everyone. You interpret. |
| Defensibility | "The algorithm said no" | "My gut feeling" | Evidence for legal, leadership, or the candidate |
"Won't candidates hate this?"
We don't force anyone through AI screening. Instead, candidates who want faster consideration choose to complete a 30-minute verification interview. Verified candidates move to priority review within 48 hours.
It's positioned as a benefit, not a barrier. And strong candidates? They're the first to opt in.
CandiScan Verification
via Acme Corp hiring team
Hi Sarah,
Acme Corp is reviewing applications for Senior Developer and wants to fast-track qualified candidates.
Complete a 30-minute video verification at your convenience — verified candidates move to priority review within 48 hours.
This is optional. Your application is being reviewed regardless.
But... Is It Too Good to Be True?
Fair questions. Real answers.
"AI screening tools are inaccurate, biased, and I'll get sued!"
"True talent doesn't want to be pre-screened"
"The CVs I get don't look fake"
"We only hire 5–10 people a year — is it worth it?"
"Our candidates are senior/executive — they won't tolerate this"
"We already do reference checks"
"What about GDPR and data privacy?"
We were spending two full days per hire just on phone screenings. CandiScan cut that to 5 minutes per candidate — and the evidence reports are better than anything we were getting manually.
Brian Loughnane
CEO, StoreSafe
COMPLIANCE ISN'T A FEATURE. IT'S A DEADLINE.
The EU AI Act hits in
6 months
Recruitment AI is classified as high-risk. Non-compliance means fines up to 3% of global revenue. Here's what that means for your hiring stack.
GDPR
Every piece of candidate data — CV uploads, interview recordings, evidence reports — requires informed consent, encrypted storage, and deletion on request. CandiScan was architected for this. Consent flows, E2E encryption, and one-click data erasure are built into the core, not bolted on.
EU AI Act — High-Risk Classification
Recruitment AI must provide human oversight, explainable outputs, and documented risk assessments. Black-box scoring is explicitly prohibited. CandiScan was built after the regulation was published — evidence-based flags with human override were the founding design decision, not a retrofit.
EEOC & Anti-Discrimination Law
AI screening tools that profile communication styles have a 61% false positive rate for non-native English speakers (Stanford HAI). CandiScan doesn't profile language, accents, facial expressions, or "personality." We verify factual claims from the CV. Same questions, same process, every candidate.
Most recruitment AI tools were built before these regulations existed. CandiScan was built because they exist.
Simple, transparent pricing
Per role. No per-seat charges, no annual contracts.
FREE TRIAL
- Up to 100 candidate interviews*
- AI or human-conducted pre-interviews
- Full evidence reports
- 30-day access
*100 AI-powered or 20 human-powered
AI VETTING
- Up to 100 candidate interviews
- Adaptive AI-powered pre-interviews
- Full evidence reports + all flag categories
- Interviews in 20+ languages
Additional interviews:
€4 each
HUMAN VETTING
- Up to 100 candidate interviews
- Human-conducted pre-interviews
- Full evidence reports + all flag categories
- Interviews in English, Estonian, Russian, Georgian
Additional interviews:
€9 each
AI Vetting languages: English, Spanish, French, German, Portuguese, Italian, Dutch, Polish, Swedish, Danish, Norwegian, Finnish, Czech, Romanian, Hungarian, Turkish, Arabic, Hindi, Mandarin Chinese, Japanese, and growing
A bad hire costs an average of 30% of annual salary. For a €60K role, that's €18K wasted. CandiScan pays for itself if it helps you avoid just one.
Screen Your First Role Free
No credit card. No sales tactics. Start screening as early as tomorrow.