Last week marked our sixth Demo Day, this one focused on Fighting Third-Party Fraud. Ten vendors stepped up to show how their solutions tackle account takeover (ATO), business email compromise (BEC), and synthetic identity fraud. Each had 15 minutes to prove their case, followed by a live Q&A with an audience of fraud, risk, and security leaders.

Across the sessions, a consistent theme emerged: the biggest shift in the fraud prevention market isn’t in the tactics fraudsters use, but how enterprises are buying solutions. Detection is expected; what matters now is whether a tool can keep the business running without stalling growth or turning away good customers. Buyers want assurance that fraud prevention supports stability by keeping customers moving, revenue intact, and trust unbroken when fraud inevitably spikes.

What is third-party fraud?

For readers outside the space, third-party fraud happens when criminals exploit someone else’s identity to gain access. Unlike first-party fraud, where the individual misrepresents themselves, third-party fraud relies on stolen or fabricated credentials to impersonate a trusted user.

Classic examples include:

• Account takeover (ATO): hijacking legitimate accounts, often through phishing or stolen credentials
• Business email compromise (BEC): impersonating executives or vendors to redirect payments
• Synthetic identity fraud: blending real and fake data to create convincing personas
  

In 2024, consumers reported losing $12.5 billion to fraud, a 25% jump year-over-year and the highest on record. Account takeover attacks alone rose nearly 67% in the past two years as fraudsters leaned on phishing, social engineering, and increasingly AI-driven methods.

As Miguel Navarro, Head of Applied Emerging Technologies at KeyBank, put it: “Think about deepfakes like carbon monoxide — you may think you can detect it, but honestly, it’s untraceable without tools.” That risk is no longer theoretical; it’s already showing up in contact centers and HR pipelines.

Walking the friction tightrope

Every fraud solution has to walk a tightrope: protect the business without slowing customers down. In this Demo Day, that balance was explored in the Q&A, with audience questions focusing on onboarding delays, false positives, and manual review trade-offs. What happens when onboarding drags? How are false positives handled? Where do manual reviews fit?

Miguel also added:“…a tool might be a thousand times more effective, but if it’s too complex for teams to adopt, it’s effectively useless.”

Providers responded with different approaches. Several leaned on behavioral and device-based analytics to make authentication seamless, layering signals like keystroke patterns and device intelligence so genuine users pass in the background. Others showed risk-based orchestration, combining machine learning models and workflows so only high-risk activity triggers extra checks.

Protecting customers from themselves

One theme that stood out was how solutions are evolving to address social engineering. As Mzu Rusi, VP of Product Development at Entersekt, explained: “It’s not enough to protect customers from outsiders — sometimes we have to protect them from themselves when they’re being socially engineered to approve fraud.”

That means fraud platforms are no longer judged only on blocking malicious logins. They’re also expected to intervene in context, analyzing signals like whether the user is on a call while approving a transfer, or whether a new recipient account shows signs of mule activity.

Human touch as a deterrent

Technology was the backbone of every demo, but Proof emphasized how human interaction remains a powerful fraud defense. Lauren Furey, Principal Product Manager, shared how stepping up to a live identity verification can shut down takeover attempts while preserving trust: “The deterrence of getting a fraudster in front of a human with these tools is enormous. Strong proofing doesn’t have to feel heavy, and customers leave reassured rather than abandoned.”

This balance — minimal friction for real customers, targeted intervention for fraudsters — ran through the day.

From fraud loss to balance sheet risk

Fraud was reframed as a balance sheet problem, not just a technology one. As Sunil Madhu, CEO & Founder of Instnt, put it: “Fraud is inevitable. Fraud loss is not. For the first time, businesses can transfer that risk off their balance sheet through fraud loss insurance.”

That comment landed because it spoke to CFO and board-level concerns. Fraud is no longer just an operational hit; it’s a financial exposure that can be shifted, managed, and priced. But shifting fraud into financial terms doesn’t reduce the pressure on prevention teams — it only raises the bar for the technology that keeps fraud within acceptable limits.

How detection is evolving

On stage, several demos highlighted identity and device scoring as the new baseline, layering biometrics, transaction history, and tokenization to judge risk in milliseconds. Others pushed detection even earlier in the journey, using pre-submit screening to catch bad actors before they hit submit.

Machine learning also played a central role in the demos. Several providers showed how adaptive models can cut down false positives while continuously improving through feedback loops. Phil Gordon, Head of Solution Consulting at Callsign, described it as creating a kind of “digital DNA”: “Every customer develops a digital DNA — how they type, swipe, or move through a session. That lets us tell genuine users apart from bots, malware, or account takeover attempts in milliseconds.”

That theme carried into the fight against synthetic identities. Alex Tonello, SVP Global Partnerships at Trustfull, explained how fraudsters engineer personas to slip through traditional checks: “Synthetic fraudsters build identities with new emails, new phone numbers, no history. By checking hundreds of open-source signals at scale, we see right through that façade.”

Others extended the conversation to fraud at the network level. Artem Popov, Solutions Engineer at Sumsub, noted: “Fraudsters reuse documents, devices, and identities across hundreds of attempts. By linking those together, you expose entire networks — not just single bad actors.”

The boardroom shift

Fraud used to be a line item in operations, managed quietly by fraud prevention teams and written off as the cost of doing business. That’s no longer the case. The scale of losses, reputational damage, and operational disruption means fraud has moved up the agenda and onto boardrooms.

Executives now face a harder challenge: choosing tools that don’t just stop fraud, but that protect business continuity. They want proof that investments in prevention will keep revenue flowing when attacks spike, not just reduce fraud losses on a spreadsheet. Boards are asking whether controls are strong enough to protect customer trust, whether onboarding processes can scale without breaking, and whether the business can keep moving if a wave of account takeovers hits overnight.

They are right to pay attention. Fraud and continuity now rank among the top five enterprise risks. Technology shifts like Apple and Google restricting access to device data are making established defenses less reliable, reframing fraud not only as a security issue but as a continuity problem.

Watch the Recording

Did you miss our Third-Party Fraud Demo Day? You can still catch the full replay of vendor demos and expert insights:
Watch the Third-Party Fraud Demo Day recording here

Key Takeaways

• Liminal’s sixth Demo Day spotlighted 10 vendors tackling third-party fraud.
• Global fraud losses are nearing $1 trillion annually, with ATO alone costing banks $6,000–$13,000 per incident.
• Audience Q&A revealed that the hardest problems are manual reviews, onboarding delays, and false positives.
• Leading vendors balance speed, scale, and user experience, reducing both fraud losses and abandonment.
• Fraud prevention has shifted from a back-office function to a board-level resilience strategy.