guavaHow can a tech staffing firm catch candidate fraud with AI?
Use AI to flag what deserves a second look: work histories that don’t line up, identity and location inconsistencies, and duplicates across your ATS. Test interviews with live, job-related follow-ups. A recruiter checks every flag, and nothing is rejected automatically. Guava’s AI Rollout builds this in 3–6 weeks for $25,000.
What to flag
- Work histories Dates, employers and titles that differ between the résumé, the LinkedIn profile and what the candidate says on the screen.
- Identity and location An address, phone number or time zone that doesn’t fit what the candidate told you, and details that change between the interview and onboarding.
- Duplicates in your ATS The same résumé, photo, phone or email turning up under different names.
- Changes at onboarding A late change of address or payment platform. The FBI specifically asks hiring teams to watch for these.
Why this reached staffing desks
Gartner predicts that by 2028, one in four candidate profiles worldwide will be fake. The FBI has warned of North Korean IT workers using AI and face-swapping technology in video interviews, and it tells employers to verify that third-party staffing firms run robust hiring practices and audit them. So your clients may ask you to show yours.
AI-assisted interviews
Don’t flag how an answer sounds. AI-text detectors have misclassified non-native English writing as AI-generated, so a style flag can put real candidates under suspicion. Ask every candidate the same job-related follow-ups live, then compare what they can show with what the résumé claims.
Where AI stops
The FBI’s October 2026 notice says these workers can’t be identified reliably with any single indicator, and that aggressive controls carry unusually high risks of false positives. That’s why a flag is only a reason to look closer. The check itself is human: a live follow-up question, specific questions about location and education, as much of the hiring and onboarding in person as you can manage, and identity verification. Never flag on name, accent or national origin, and read the AI hiring rules before you automate anything.
What it can be worth
The sample AI Assessment for a fictional staffing firm with 45 internal staff gives these flags a verdict of Pilot: 10 hours a week back for $300 a month in AI tools. Those numbers are illustrative, not client results. Auto-rejecting candidates gets Don’t build.
How Guava does it
- AI Assessment $4,000, 2 weeks. Most staffing firms start here. Ranks fraud flags against screening, submittals and outreach. Credited in full toward the Rollout.
- AI Rollout $25,000, 3–6 weeks. Sets up the flags with measured baselines, then trains your recruiters and account managers on checking them.
- Ongoing Optimization & Governance $4,000 a month, cancel anytime. Keeps the flags working as tactics change, with access reviews and a quarterly review.
Fixed fees, agreed in writing before work starts, and you own everything built. See the engagements
Questions staffing owners ask
Can AI reliably detect a deepfake interview?
I wouldn’t rely on it alone. Treat any detection as a flag, then check with a live follow-up question or an in-person step. The FBI says no single indicator is reliable.
Will this reject candidates automatically?
No. A recruiter checks every flag, and nothing is rejected automatically. I don’t recommend auto-rejecting candidates.
Could fraud flags discriminate against real candidates?
They could, so build them on facts that don’t add up, never on name, accent, national origin or how someone writes, and check the results for bias. Your lawyer should review the process.
Do we need a new ATS or a fraud tool?
Not necessarily. The Assessment looks at what your ATS and interview tools already give you, and says whether a dedicated tool is worth buying.
More questions staffing owners ask
Talk it through.
Twenty minutes on what’s reaching your desk. Prefer email? romeo@workwithguava.com