Five categories of fake candidate detection tool
Tools that claim to detect fake candidates cluster into five categories. They solve different problems at different stages of the funnel, and the right answer for most teams is two or three working in combination, not one tool trying to do everything.
- Intake-stage browser extensions: fast, per-applicant checks on email, phone, and identity that run from the recruiter's toolbar. Verif_Hire, TurboCheck, Endorsed, Jobright, and Tofu live here.
- Interview-stage proctoring: real-time deepfake, proxy, and AI-assist detection during live video interviews. Sherlock, InterviewGuard, Pindrop Pulse, Talview, and Polygraf live here.
- Full identity verification services: government-ID matched to a live selfie, with liveness detection. Persona, Onfido, Veriff, Sumsub, Socure, and Proof live here. Best-fit for offer stage.
- Background-check providers: SSN trace, criminal, employment, and education verification under the FCRA. Checkr, Sterling, and HireRight live here. Best-fit after offer acceptance.
- ATS-native fraud controls: rules built into the applicant tracking system itself. Ashby, Greenhouse, and (more thinly) Lever and Workday live here.
Intake-stage browser extensions
Browser extensions are the only category that can realistically run at top-of-funnel volume. The marginal cost per check is near zero, the recruiter never leaves their ATS or LinkedIn tab, and the check takes seconds. This is where Verif_Hire competes, and it is the most crowded shelf in the category right now.
Verif_Hire
Verif_Hire is an intake-verification chrome extension. It runs from the recruiter's toolbar, takes about 20 seconds per applicant, and returns one of three verdicts: all information verified, review recommended, or inconsistency detected. It is positioned as a recruiter-driven tool: no engineering integration, no ATS plugin required, no sales call to start. Free plan is one verification per weekday; paid Pro runs $49/mo for unlimited verifications. Verif_Hire is explicitly not a background check, not an offer-stage ID service, and not an interview-stage proctor. It sits at the top of the funnel and pairs with one of those other categories.
TurboCheck
TurboCheck markets a five-second fake candidate check that cross-references identity against 40+ U.S. data sources including LinkedIn, Facebook, Twitter, GitHub, and Atlassian. It also runs a video-call selfie check with facial biometrics for liveness, partnered with authID for the underlying identity assurance, and integrates with Bullhorn through their marketplace. Pricing is volume-tiered, with discounts above $300/month of service. The TurboCheck wedge is "intake plus a lightweight liveness layer," sitting between pure intake and full IDV.
Endorsed (FraudShield)
Endorsed is an AI recruiting platform whose FraudShield product detects fake candidates inside the ATS. They tag a risk score on every applicant in Greenhouse, Ashby, Workday, and Lever, with a chrome extension as a thin client. CEO David Head's prior product context is Braintrust and Laskie, both marketplace plays, and Endorsed has positioned itself enterprise-up: bigger ATS integrations, sales-led, fewer self-serve recruiters. The competitive overlap with Verif_Hire is the chrome extension surface.
Jobright Fraud Detection
Jobright's fraud detection is a free chrome extension that flags fake applicants inside Greenhouse, Lever, Workday, and other ATSs. It cross-references resumes against LinkedIn, public records, phone, and email across 230+ signals, and rides the brand halo of Jobright's main B2C product (an AI job-search copilot used by 2M+ U.S. job seekers). The pitch is "free, lots of signals, ATS-embedded." The tradeoff is that it is a side product for them rather than a core focus, and the data flow is one-way: signals only, no remediation workflow on top.
Tofu
Tofu's chrome extension lets recruiters run applicant fraud detection with one click on a full job, a particular pipeline stage, or an individual applicant, directly from inside the ATS. Tofu is the lightest of the intake-stage tools by surface area — extension only, ATS-embedded, narrow scope.
What this category catches and misses
Catches: high-risk email domains, contact information that fails basic sanity and footprint checks, cross-source identity mismatches, resume-derived contact inconsistencies. Misses: proxy interviewee fraud that starts after intake, offer-stage ID fraud, the rare fraudster using a legitimate long-held personal email and real mobile under a borrowed identity. Pair with one of the other categories to close those holes.
Interview-stage proctoring
This category did not exist as a clean shelf two years ago. Cheap consumer deepfake tools, the KnowBe4 incident, and the AI-assist tools (Cluely, Yoodli, InterviewCoder, Pickle) created enough demand for purpose-built proctors to monitor live interviews. Five vendors are doing the bulk of the noise here.
Sherlock AI
Sherlock AI applies a multimodal adversarial ML approach to interview fraud, combining device activity, audio environment, and candidate behavioral signals into a single classifier. They focus on "human authenticity" rather than pure media-deepfake detection, on the theory that the underlying signal of a proxy or AI-assisted candidate is observable across modalities even when individual deepfake artifacts are subtle.
InterviewGuard
InterviewGuard markets real-time detection of every category of interview fraud: AI assistants, proxy candidates, deepfakes, and remote desktop takeovers. Positioning is "single tool covers every interview-stage threat," which is rare in the category — most peers focus on one or two of those vectors.
Pindrop Pulse
Pindrop is a long-standing voice fraud company (originally fintech call-center anti-fraud) that pivoted Pulse for Meetings into interview-stage detection. It integrates directly into Zoom, Microsoft Teams, and Webex sessions and alerts recruiters the instant it detects synthetic audio or video. Strongest signal stack of the group on the audio side specifically, given Pindrop's heritage.
Talview
Talview embeds deepfake interview detection inside its broader video-interviewing platform. It analyzes micro-expressions, facial-texture inconsistencies, eye-movement patterns, and depth cues to flag synthetic or manipulated video feeds. Best-fit for teams already running structured async or live video interviews on Talview's platform; less appealing as a standalone overlay.
Polygraf AI
Polygraf transcribes live interview responses and detects when candidates are using specific AI-assist tools (Cluely, Yoodli, InterviewCoder, Pickle). The wedge is brand-specific: rather than trying to detect "any AI assistance," it identifies the named tools by their behavioral fingerprints. That makes it narrower but more actionable when those tools are the threat.
Full identity verification services
This is the mature category that predates the fake-candidate panic by about a decade. The names are well known: Persona, Onfido, Veriff, Sumsub, Socure, and Proof. They verify a candidate's government-issued ID against a live selfie, run liveness detection to prevent static-photo spoofing, and increasingly add device, network, and behavioral signals on top.
Persona
Persona launched a dedicated Candidate Verification SKU in 2026 that integrates with Ashby, Greenhouse, and Workday. Government ID plus a live selfie, plus device, behavioral, and network signals to detect anomalies. The new SKU is the most direct enterprise competitor in this category to anything that calls itself "candidate verification."
Onfido
Onfido offers identity verification using document and biometric checks. Mature SDK-style capture experiences, ~2,500 documents supported, decision times in the 30–60 second range with manual review on harder cases.
Veriff
Veriff supports 13,500+ identity documents and delivers verification decisions in roughly 6 seconds on average. They lean hard on AI fraud prevention features (Risk Labels, FaceBlock, CrossLinks) and have an in-house fraud team adapting to new threats. Best-fit when document coverage breadth and decision speed both matter.
Sumsub
Sumsub is the compliance-heavy KYC option of the group. Strong fit when the same hiring flow needs to satisfy financial-services or crypto-grade verification, weaker fit for general-purpose recruiting use cases that don't need that depth.
Socure
Socure runs passive identity checks before pipeline entry, positioning itself one step earlier in the funnel than the document-and-selfie group. Closest spiritual neighbor to the intake-extension category, but at enterprise pricing and with an enterprise sales motion.
Proof
Proof is an identity layer for HR, with a dedicated candidate-fraud use case. Recruiters embed identity verification links into their hiring workflows to filter out bots, deepfakes, and impersonators early. Different enough in delivery model from the SDK group above that it sometimes shows up in shortlists alongside Persona rather than as a direct alternative.
What this category catches and costs
Catches proxy interviewee fraud at offer stage. A candidate who cleared five interviews but can't produce their ID on video is the offer you don't extend. The economics: typically $2–$8 per check and 3–5 minutes of candidate time. That cost shape is fine at offer, wrong at intake. Use them where they're strong.
Background-check providers
Checkr, Sterling, HireRight, and similar providers (GoodHire, Verified First, First Advantage, Certn) verify SSN, criminal history, employment, and education. These are mature services with clear FCRA compliance paths. For a category-wide look at lower-cost and faster options, see alternatives to traditional background checks.
Where they help on fraud
Detection of identity fraud where the fraudster is using another real person's SSN. Employment verification sometimes surfaces fabricated roles, though it depends on how thoroughly the vendor contacts the former employer.
Where they fall short on fraud
They're post-offer, so they don't protect against the time and calendar cost of fraudulent candidates who waste interview loops. They also don't catch proxy interviewees, because the person being background-checked is the persona on paper, not the person who will show up on day one.
The right mental model
Background checks are a compliance and hire-quality tool. Treat them as such. They aren't a fake-candidate-detection tool in the real-time, top-of-funnel sense.
ATS-native fraud controls
The native picture is uneven by ATS. Ashby launched what it described as the first ATS-integrated fraud detection system in 2026 and has the deepest native feature set in this group. Greenhouse integrates with IPQS and surfaces fraud signals directly to recruiters inside the platform, automatically flagging suspicious applications. Lever and Workday have nothing equivalent native at the time of writing — they rely on partners and third-party plugins.
The practical recommendation: turn on whatever fraud controls your ATS offers (they're typically free), and supplement with a dedicated intake-verification tool for the signals your ATS doesn't cover. If you're on Lever or Workday, the supplement matters more, because there is no native baseline.
How to choose
A simple decision tree:
- If you're screening more than 50 applicants a week: install an intake-verification browser extension. This is the highest-leverage change. Verif_Hire, TurboCheck, Endorsed, Jobright, or Tofu depending on your ATS, your budget, and whether you want self-serve or sales-led.
- If you do live remote video interviews for technical roles: add an interview-stage proctor. Sherlock, InterviewGuard, Pindrop, Talview, or Polygraf depending on whether your dominant threat is AI-assist, deepfake video, deepfake audio, or proxy.
- If you hire for sensitive remote roles or anywhere proxy fraud is plausible: add a full IDV service at offer stage (Persona, Veriff, Onfido, Sumsub, Socure, Proof). This closes the proxy-interviewee hole.
- If your hires have regulated access or touch sensitive data: keep or add a background-check provider. This isn't optional; it's a compliance requirement in many industries.
- If your ATS offers fraud controls: enable them. They cost nothing. Lever and Workday users: this step is moot, skip to step 1.
A well-protected hiring team usually runs an intake-stage tool plus one or two of the others, with the intake tool doing the bulk of the per-applicant work and the others closing specific holes downstream.
FAQ
Can one tool do everything?
Not well. The tradeoffs between top-of-funnel speed and offer-stage rigor pull in different directions. Vendors that claim to do both tend to do neither particularly well.
How much should a hiring team expect to spend?
An intake-verification extension for a team of five recruiters is typically in the low hundreds of dollars a month. Interview-stage proctors are usually per-interview or seat-based and run higher per-use. Offer-stage ID verification is priced per check ($2–$8). Background checks vary widely by package. The total fraud-protection spend for a small company should sit under $1,000/month and well under 1% of recruiting opex at most stages.
What about AI-powered fraud detection?
Several vendors market "AI" as the headline capability. In practice, the reliable part of fraud detection is still the rules-based signal stack described in our red flags guide. AI plays a supporting role (detecting AI-generated résumés, anomalous behavioral patterns), but it's not a replacement for the boring, high-yield checks on email, phone, and identity.
How do I pick between Verif_Hire, TurboCheck, Endorsed, Jobright, and Tofu?
The honest split: Verif_Hire and Jobright are the self-serve recruiter-driven options (no sales call required, fastest to install). Endorsed and TurboCheck are the sales-led enterprise options with broader ATS integration and bigger feature surface. Tofu is the narrowest, embedded directly inside ATS workflows. If your team is small and wants to install something today, start with Verif_Hire (free plan) or Jobright (free extension). If you're enterprise and want a procurement-friendly contract, look at Endorsed and TurboCheck.