### **Key Takeaways**

- AI should support talent strategy, not become the strategy itself
- TA teams should focus on solving business outcomes rather than automatically filling roles
- Internal mobility should be evaluated before opening new headcount
- Many hiring problems are actually workflow or process problems
- AI and tooling can create capacity, but they cannot replace judgment and contextual decision-making
- Slowing down early in the hiring process often leads to better organizational decisions
- TA leaders can play a significant role in org design, not just recruiting execution
- Hiring success should be measured by business impact, not simply by filling roles

### **Transcript**

*Note: transcript is auto-generated, there may be typos.*

At Dave, finding high-performance talent and fighting fraud are the same job. My name's Tom Chapman. I'm the Director of Talent Acquisition at Dave. And I'm here to tell you a little bit about our processes and how we fight fraud. Today I'll be covering three things: who is responsible for fighting fraud, where in your processes fraud can slip through, and what a small remote-first fintech has done about it.

Before we jump into that, I think it's really important to look at what we're up against. Ashby's Talent Trend report shows that applications per hire are up 182% since 2021. Has everyone else seen that in real time? Yeah. In April this year, we opened three roles. We had 7,500 applicants in two to three days. It was absolutely insane and very hard to deal with. Only 19% of hiring managers right now are extremely confident that their processes would catch a fake candidate — and that's getting harder and harder to spot. And 35% say someone other than the listed applicant has participated in a virtual interview. Lastly, by 2028, Gartner predicts that one in four candidates will be fake, which is insane to think about. So the scale of the problem is clear.

But before we talk about solutions, I think it's really important that we align on whose problem it actually is to solve. Because from my point of view, it shouldn't just be talent acquisition. And I don't think it should be IT and InfoSec's problem either. If a fraudulent candidate gets through your hiring process, it's a company problem. The consequences will land everywhere. At Dave, every team understands that. And we're all committed to fighting fraud. We're a fintech — we serve over 12 million customers. Fraud isn't abstract for us. It's part of how we operate day to day. Fraud and financial risk can emerge quickly, and when it does, the impact isn't isolated. It can affect customers, the platform, and the business pretty fast. Because of that, risk isn't owned by just one team. Everyone across the company is expected to recognize and call something out if it doesn't look right.

So we're not operating in isolation in TA. We're part of a broader system designed to protect the company. And in that context, a fraudulent hire isn't just a bad hire — it's a potential security risk. For TA at Dave, fraud isn't new either. We've been fully remote since 2020. So we've been dealing with things like location fraud, where candidates misrepresent where they're based. We'd also see credentials that didn't check out, experience that didn't quite add up. Back in the day, it was much easier to spot — a quick LinkedIn check or a phone call and things would unravel pretty quickly. The key difference right now is that before, those were edge cases. AI has changed the nature of fraud. It's changed the scale and how easy it is to pull off.

So where does fraud slip through? You can't control the sophistication of bad actors, but you can control the environment that your team is working in. And in my experience, that's where the real risk lives — not in the threat itself, but in the conditions that make it easy to miss. We all know what a stretched recruitment team looks like: too many open roles to manage effectively, volume over quality targets, out of control inbound candidates, and pressure to fill things quickly — forcing decisions made on momentum rather than judgment. When that's the environment you're working in, fraud doesn't have to be sophisticated. It just needs you to be distracted or overwhelmed and the door to be open. Our response wasn't a new tool or a new process or an overhaul of policies. It was building a system that gives recruiters the room to do their jobs properly.

That starts with how we manage our rec load. Not fewer roles necessarily, but opening positions when the TA team is ready. We stagger openings based on where current open positions are in the process and how ready the new position is to open. Our capacity plan matters and overloading capacity has a negative impact on quality. So this might feel a bit counterintuitive — proactively trying to slow things down. But we're on track to fill double the amount of roles that we filled in 2025, and not by adding headcount to TA, but by being intentional with sequencing. So for me, pressure is the real risk factor. The solution is designing a system that reduces pressure from the start. My team is targeted to take less than 15 recruiter screens per hire. Any more than that in our environment signals misalignment in the search. If your team has space to think, they will notice more, they will make better decisions, they will ask more questions. They'll rely less on surface-level signals. More time to think means fewer misses. The hard part here is holding the line when pressure from the business to hire fast kicks in. And I think that's exactly where it matters most.

When I started working on this session, I really had to rack my brains on what we actually did at Dave to build a fraud-resistant process — because it definitely wasn't something fancy. I realized for us it was about getting back to basics. And I kept coming back to four words: capacity, clarity, consistency, and confidence — which is pretty convenient. So let's double click on the Four C's. We start with capacity. How much work can your team do before quality starts getting impacted? How many roles can your team handle at the same stage before they start getting overwhelmed? Are they stretched to the point where they're making fast decisions instead of good ones? That's where signals get missed.

Then we move to clarity. This can show up in a few different ways, but it starts with a headcount plan — are you able to support the needs of the business based on your capacity model? This is also extremely important for us when it comes to opening specific positions. We ask for what success looks like before a rec is even approved. What KPIs or OKRs will this person impact when they join? If that's fuzzy, everything downstream becomes more complex. And complexity impacts your capacity plan. Then we move to consistency. How much variance are your recruiters and hiring teams dealing with? When your processes differ by team, by hiring manager, by recruiter, cognitive load goes up and assessment quality goes down. Consistency in process is one of the most effective ways to improve signal detection and reduce fraud risk. The goal here isn't rigidity — it's reliability in how your signals are gathered and evaluated. The less unnecessary variation there is, the easier it is to spot when something feels off.

And finally, confidence. Not just in your recruiters — though that is very helpful — but really through your process. It's about building a process thorough enough that when you make a decision, you trust it. So ask yourself: what aligns? What contradicts? Where are the gaps? Confidence earned this way isn't just a feeling — it's a conclusion. And if confidence wavers after everything you've been through, go back to clarity. Redefine what a confident decision would look like.

Here's what this looks like for us end to end. It starts earlier than you might think, with role design and calibration at approval. Before a role even opens, hiring teams define what success and what good looks like. TA is brought in at that stage to provide feedback on profiles ahead of approval. This gives recruiters a precise bar to assess against — and fraudulent candidates to fail against. Then channel strategy: referrals and targeted sourcing are activated deliberately, often before we open roles externally. These candidates provide a strong early signal and typically present lower fraud exposure. Controlled postings: we open external postings for a defined window, not indefinitely. Closing a role creates a natural reset point — we can review talent, recalibrate, and repost if needed, rather than just accumulating noise.

Application questions — I used to be so against this earlier in my career, and I can't quite believe I'm stood here suggesting that people do this. I used to think it was a friction point that would deter strong candidates from applying. Now I see it as a signal of engagement in a very, very saturated talent market. Targeted questions signal genuine engagement. AI-generated responses have a texture to them — and they're not treated in isolation. Someone uses AI, they're not just automatically dispositioned. We take it in context, not in isolation. We also use Ashby's fraud detection. This is a fantastic piece of kit for us. Ashby flags identity inconsistencies automatically. Our recruiters can then assess the full picture — the resume, application responses, digital footprint, and fraud signals against pre-aligned criteria. My team doesn't have any volume targets. That means quality is always coming first. They aren't speaking to a candidate just to hit a target.

We have the recruiter screen — always a two-way conversation. Recruiter pitching the role and assessing candidates against that predefined criteria. This is an opportunity to look at depth of experience and test potential fit in real time. Structured interviews, consistent stages, custom scorecards. Interviewers assessed against the predefined criteria or focus area created before the role is even opened. This gives hiring teams time to align on what success looks like. Consistency in process surfaces inconsistency in candidates. And this has helped more than you might think. Sometimes our interviewers will run their questions through an LLM before conducting interviews. And we've had candidates come through providing the exact same answers verbatim. It's that type of prep from our hiring teams that alleviates the pressure on TA to catch every potential risk. And finally, the last step for us is a formal identity and background check with our friends at Checkr. This is the backstop to a process that's already done quite a lot of the work. This recently helped us avoid hiring a candidate that hadn't been forthcoming about their location. Turns out they weren't where they said they were.

So here's the thing about candidate fraud. It becomes easier to detect as a byproduct of a well-designed hiring system. So slow down, get specific about what good looks like, build with intention. Do that well, and fraud doesn't need to be the headline. Quality will be. Thanks very much.

Tom Chapman from Dave shares how his team navigates candidate fraud and strengthens recruiting judgment in fintech hiring.

Getting Back to Basics to Fight Candidate Fraud

Speaker

Recommended Sessions

What TA Can Steal from Growth Hackers

Background Checks Without Bottlenecks

Product

Resources

Company

Developers