Keep fake candidatesout of your pipeline.
Candidate fraud detection inside your ATS. Every application checked, every signal explained. Your team makes the call.
Candidate profiles worldwide will be fake by 2028.
Gartner, 2025Of companies say they have already hired a fake candidate at least once.
GetReal Security, 2025Reported losses from employment fraud in the US in 2024, up from $90M in 2020.
FTC, 2025US companies unknowingly hired IT workers linked to North Korea through one scheme.
US Department of Justice, 2024Your ATS sees applicants.
It doesn't see the pattern.
AI writes the resume. A borrowed identity signs it. The same template lands in twenty pipelines at once. And a real applicant hides instructions for your screening software in white text. By the time someone notices, the interview is booked and the budget is gone.
CVs that read perfectly
Generated in minutes, tuned to your job ad, clean on the surface. Keyword matching loves them. Recruiters lose hours before the first inconsistency shows up.
Identities that don't belong to the applicant
Real names, borrowed LinkedIn profiles, a stand-in on the video call. What you hire is not who you interviewed. The cost shows up in access rights, not in the recruiting budget.
Applications that come in batches
One operation, many profiles. Shared phone blocks, cloned skills sections, the same file properties across "different" people. A single CV looks fine. The cluster does not.
Every application checked.
Every signal explained.
TalentShield connects to your ATS and reviews each application as it arrives. Recruiters get an Application Integrity Score, a plain-language badge and the reasons behind it. Nothing gets rejected automatically.
Connect your ATS
Native integration or API. No new screen to learn, no change to your process. Recruiters keep working where they already work.
Each application is checked against 70+ signals
Document integrity, timeline consistency, identity and contact data, cross-application patterns. Deterministic rules where the question is closed, AI where it is open. Results in minutes, not days.
Recruiters see the score, the badge and the reasons
Application Integrity Score from 100 down. One signal means "have a look". A cluster means "look before you book the interview". The verdict is always yours and it is always recorded as yours.
Signals with reasons.
Never verdicts.
Most tools try to score candidates. We help recruiters verify applications. Think of it like a lab result. One value off the reference range is a note for the doctor, not a diagnosis. TalentShield measures how consistent an application is with itself and with public information. Interpreting it is a human job.
Document integrity
What the file says about itself and whether that matches what the candidate says. Hidden content, authorship, generation traces, templates.
Timeline consistency
Roles that overlap, dates that drift between the CV and the profile, gaps that are explained and gaps that are not.
Identity and contact data
Whether the name, contact details, location and public professional profile point at one real person.
Cross-application patterns
Signals no single CV can show: the same phone block, the same skills section, the same file fingerprint across many applicants.
Every signal carries a weight, and signals that cluster across categories count more than any single one. The result sorts your pipeline so recruiters know where to start. It never rejects anyone. What candidate fraud detection covers, from AI resumes to prompt injection.
Lives where recruiters already work.
TalentShield runs inside your ATS. Results show up next to the candidate, verification starts when an application arrives, and nobody has to remember to open another tool.
- Automatic checks per job, with per-job on and off switches
- Badge and score synced back to the candidate record in your ATS
- Candidate data stays in your ATS, TalentShield links back to it
- API for ATS vendors and job boards that want a trust layer of their own
ATSs we support today or connect through the API
Something else? We integrate through the API. Tell us in the demo form.
One click from the ATS to the TalentShield dashboard.
Day to day, recruiters stay in the ATS and have everything they need there. When someone wants the details, the dashboard has them: full reports, every signal with its evidence, decision history, and an audit trail your compliance team will actually like.
- Full verification reportsEvery signal, the evidence behind it, downloadable as PDF
- Decision historyWho reviewed what and when, recorded as a person's verdict
- Audit trailSignal version, detected signals, score and timestamp per check
- Jobs and candidates overviewVerification status per job, filters, re-verify when the library updates
Compliance isn't a feature
we added later.
Recruitment is a high-risk area under the EU AI Act. We designed for that review from day one, not after it. TalentShield is built to pass the review of your legal team, your security team and your works council.
- Human in the loop
- No automated decisionsTalentShield produces information for a recruiter to review. It does not reject, rank out or filter anyone on its own. Every recorded verdict is a person's.
- Explainability
- Every signal has a reasonRecruiters see which checks passed, which raised a flag and why. Candidates can get a meaningful explanation, as the AI Act requires.
- Data residency
- Hosted and processed in the EUApplication data is stored on EU infrastructure. AI processing runs under EU data residency, with no retention and no training on your data by model providers.
- GDPR
- Processor, not controllerYou stay the controller of candidate data. TalentShield acts as your processor under a Data Processing Agreement with a named sub-processor list.
- Audit trail
- We log what we didSignal version, detected signals, score and timestamp for every verification. Ready for a DPIA, a candidate request or an auditor.
- Minimisation
- Only what the check needsDeterministic checks run locally. Open questions go to the model with identifying data reduced to what the specific check needs. Configurable per customer.
Security & Architecture Overview, DPA and AI Governance Statement are ready for your review. Ask for the compliance pack in the demo form.
Built by recruiters,
for recruiters.
TalentShield didn't start as a search for a SaaS idea. It started in our own pipeline. We run Team Up, a technical recruitment agency, and after eighteen years of hiring for demanding clients we got hit by fake candidates on our own jobs. So we built a process for ourselves first, on our own roles, with our own team.
Turning that process into a product took an engineer who would treat it as a security problem, not a feature request. That's Bartek. He spent years building secure digital products for US companies, and he owns the architecture, the code and every integration. The three of us are equal co-founders.
"If it doesn't work in our own hiring workflow, it doesn't ship."
TalentShield runs on every application Team Up receives today. When a new fraud pattern shows up in our pipeline, it's in the signal library the same week.
Made for teams that hire at scale
and can't afford to guess.
You hire across six countries. You need recruiters to spend their time on people, not on checking whether a PDF is real.
- A second pair of eyes on every application, without slowing the team down
- Signals sorted by priority, so screening starts where it matters
- Works in the ATS your recruiters already live in
One fake candidate sent to a client costs more than a year of the tool. Your name is on every CV you forward.
- Every candidate checked before it reaches the client
- "Recruitment secured by TalentShield" as part of your offer
- Cross-application patterns caught across all your jobs, not one at a time
Your customers keep asking what you do about fake applicants. You want an answer that doesn't take a year to build.
- API trust layer you can embed in your own product
- Explainable signals, so your customers' legal teams say yes
- A differentiator that shows up on the candidate card, not in a slide deck
You hired someone who wasn't who they said they were. The question from the board was: what did you check?
- A documented check on every application from day one
- Audit trail of what was verified, when and with which signal version
- Patterns from the incident added to the signal library, so it doesn't repeat
Questions we get from
TA leads, CTOs and legal.
No. It shows which signals it found in an application, how strong they are and what the evidence is. A recruiter reviews the evidence and records their own conclusion. There is no automatic rejection and no "fake" label produced by the system.
Usually, yes. Most patterns are visible in the application itself: file properties that don't match the person, dates that don't add up, contact data that can't be verified, a profile that differs from the CV, or the same template showing up across many applicants. TalentShield checks these the moment the application arrives.
ATS screening answers "does this CV match the job?". TalentShield answers "is this application consistent with itself and with reality?". Different question, different checks. TalentShield sits next to your ATS screening and adds the layer it was never designed to provide.
It is designed for exactly that review. Human-in-the-loop by design, explainable signals, EU hosting and EU data residency for AI processing, a Data Processing Agreement with a named sub-processor list, and an audit trail of what the system did. We describe how it works precisely rather than stamping "compliant" on it. The full compliance pack is available on request.
It stays yours. Application data lives in your ATS. TalentShield processes it on your instructions, stores what the verification report needs on EU infrastructure, and links back to the ATS record instead of copying it around. Model providers see reduced data, keep nothing and don't train on it.
With a supported ATS, verification runs on your first job within days of signing. We start with one or two live jobs, review results together weekly, and tune the setup to your pipeline. You see real signals on real applicants, not a demo dataset.
No, and we say so on every report. A single signal is a note to look closer. Honest people have messy CVs, gaps and odd file names. What separates a fabricated application is a cluster of signals across different categories, and that is what the score reflects.
Fake candidates 101,
written by people who hire.
What a fake candidate is, who is behind them, and what to do about it from Monday. Field notes from our own pipeline, with sources.
Five things fake applications get wrong, and the one thing they get right
Fabricated applications are built to pass keyword filters, not to survive a careful look. Five checks any recruiter can do in a minute, and why the sixth one is the trap.
What is a fake candidate? A plain-language guide for recruiters
Fake candidates are not one thing. From AI-polished CVs to stolen identities run by state operators: what the term covers, how big the problem is, and why one signal never settles it.
Who is behind fake candidates? The operator, the farm and the ChatGPT candidate
Three profiles account for most fake candidates in tech hiring: state-backed operators using stolen identities, laptop farms holding several jobs at once, and real people with AI-written CVs. What each wants and how each shows up.
See it on your own pipeline.
Leave your email and the ATS you use. We come back within one business day to book a 30-minute demo, on your own applicants if you want.
- Real signals on your real applicants, not a sample dataset
- Weekly review with the founders, who still recruit every day
- Compliance pack for your legal and security teams from day one
Thanks, we've got it.
One of us will reply within one business day. In the meantime, feel free to email contact@talentshield.app.