How the Greenhouse TA team uses Real Talent to spot fraud earlier and protect recruiter time

Summary: Greenhouse VP of Talent Planning & Acquisition, Ariana Moon, shares how the Greenhouse TA team used Real Talent™ to surface fraud earlier, reduce manual screening and protect recruiter time. She explains how human-led fraud signals helped restore confidence in high-volume hiring without sacrificing judgment.
‍

Hiring teams everywhere are feeling the same pressure: application volume is up and candidate fraud is no longer an edge case.

At Greenhouse, we felt this shift internally as well. As our Talent Acquisition team opened roles, we started seeing both an increased rush of applications as well as more candidates who looked legitimate on paper but weren’t qualified or acting in good faith. Our recruiters were doing a significant amount of manual verification. LinkedIn checks. Validating companies. More volume at the top of the recruiting funnel made it harder to spend time with genuine candidates.

In addition, we had little confidence in our ability to clearly distinguish fraudulent candidates until we were already on the phone with them. In these cases, we’d loop in our security team to double-check IP addresses, devices or activity patterns before deciding how to proceed, but this wasn’t efficient or scalable. Even a small number of fraudulent candidates reaching the screening stage translated into meaningful time loss – time that should have been spent engaging real candidates and collaborating with hiring managers.

In order to address this problem, we started using Real Talent  in our day-to-day hiring workflow to identify potential fraud earlier, reduce time spent on low-quality applications and free up recruiters to focus on real candidates.

‍

What changed once fraud signals were visible?

One of our biggest upgrades was piloting fraud detection directly inside Greenhouse, which surfaces candidate fraud signals during application review. Identifying potential risk earlier in the funnel gave recruiters clear, reviewable signals that helped them have more confidence in who they chose to interview. This made the process more efficient, especially for high-volume roles.

Like many teams navigating increased application volume, we believe strongly that the decision to reject a candidate should be a human one, not made automatically by a system. Real Talent supported that approach by surfacing a spectrum of fraud signals without taking action on our behalf. Recruiters could prioritize candidate review, apply judgment and decide how to proceed based on the context of the role and their team’s tolerance for risk.

Using Real Talent, we adopted a simple mental model to guide our decision-making:

• A single signal was a prompt to look closer
• Multiple high-risk signals together were a strong indicator of fraud

This approach proved invaluable. It allowed our team to move quickly on low-risk candidates, spend time where nuance was required and escalate only when necessary. Real Talent gave our recruiters the flexibility to balance speed, quality and risk in a way that felt practical and trustworthy.

The initial numbers were astounding. In piloting fraud detection for a Machine Learning job search, we conducted a manual review of applicants flagged for fraud to better understand how well the signals were performing.
‍

What we found was:

• More than 35% of applicants were flagged as having high-risk signals
• 91% of the high-risk candidates further reviewed by our security team were confirmed as fraudulent
• Candidates flagged only by weaker signals were more nuanced. Among those reviewed, 26% were confirmed fraud, while many fell into gray-area cases that required human judgment

This reinforced an important principle to us: while strong signals are highly reliable, weaker signals work best as warnings. This was the balance we wanted in a human-led hiring process.

‍

What this looks like in practice

One example best outlines the tangible impact of fraud detection.

During our pilot of the functionality, one of my recruiters had five screens scheduled for a role and ran the fraud report across those candidates. Four out of five were flagged as high risk. That said, the recruiter kept the interviews to validate what the system was showing. Out of the five:

• Three candidates no-showed
• One candidate showed suspicious behavior during the interview
• One candidate had a strong resume but could not meet the bar in live conversation

Those interviews were 30 minutes each, representing roughly an hour of recruiter time that could have been spent engaging stronger candidates or moving real applicants forward.

As we continued piloting Real Talent, we began to see reassuring patterns. Fraud was being surfaced earlier, interviews were becoming more intentional and recruiters were regaining confidence in who they chose to schedule.

Most importantly, teams began to trust the hiring process again. In some cases, recruiters who had previously avoided publicly posting their jobs because the inbound risk felt too high began to rely on the pipeline again.

‍

How we think about Real Talent as essential to our hiring process

Fraud detection is non-negotiable for us today because of how we’ve seen it help our recruiters save time and prioritize their ability to meet with real candidates. And the impact compounds when we use fraud detection as part of a layered approach with the rest of the Real Talent functionalities:

• Fraud and spam detection reduce volume and risk at the top of the funnel
• Talent Matching helps prioritize the most relevant candidates based on recruiter-defined criteria
• Identity verification adds confidence before the most time-intensive stages of the process

When combined, Real Talent delivers triple layers of trust. Each layer strengthens the next, helping recruiters focus their attention where it belongs – on candidates who are both real and qualified.

The truth of the matter is, volume of applications is not just an efficiency problem. It’s a trust problem. With Real Talent, we’ve gained the ability to identify fraud earlier in the process, thereby protecting recruiter time and increasing confidence without automating away important hiring decisions. Through surfacing better and more timely information to recruiters, Real Talent ultimately helps keep hiring human.
‍

Real Talent is now available. Request a demo to see how teams use it to reduce fraud at the top of the funnel and make confident, human-led hiring decisions.