The Next Fraud Unicorn Will Design Fraud Out of the System

Every fraud algorithm fails the same way: it waits for fraud to happen. In 2024, as companies raced to detect fraud faster, losses still soared 25% to $12.5 billion. AI-powered deepfakes and phishing hit 1 in 10 companies. The real problem isn't speed—it's strategy.

Most fraud solutions follow a familiar playbook: analyze transactions, flag suspicious activity, block bad actors. Established players like Stripe and Feedzai process billions of transactions and catch fraud in milliseconds. Yet fraudsters still win because detection is fundamentally reactive—by the time algorithms catch them, the money's gone.

Synthetic identity fraud is a prime example. These fabricated personas are engineered to look legitimate, operating undetected for years before executing schemes worth $15,000 per case on average. Since 2019, attempts have surged 184%. Traditional detection tools can't catch them because they're designed to pass every test.

Leading founders are now asking a bold question: What if fraud became structurally impossible?

The Signals We're Tracking

Through Blindspot, our proprietary platform tracking AI in fintech, we've identified three breakthrough approaches driving the shift from reactive to preventive fraud architecture.

• Behavioral Nudges: Embedded solutions that spot rushed payment and introduce strategic friction—like countdown timers, re-verification, or transaction previews—to prompt users before risky actions. The breakthrough isn't the friction itself, but how these nudges are now dynamically triggered by real-time risk signals.

• Dynamic Identity: Real-time identity graphs that continuously assess trustworthiness across device, location, and behavior patterns. Risk scores update in milliseconds, catching synthetic identity fraud at the moment of creation.

• Risk-Adaptive UX: User interfaces that adjust to risk signals. High-risk users see simplified flows with safeguards. Low-risk users get frictionless experiences.

The Builders

Companies across the industry are leading the charge with new strategies designed to stop fraud before it happens.

Established Leaders

Companies like Persona and Alloy have demonstrated the value of comprehensive identity platforms, achieving exponential growth while validating market demand for integrated, risk-aware onboarding. Their success underscores a broader industry shift from reactive fraud mitigation to proactive defense.

Scaling Innovators

In the growth stage, companies like Bureau are leveraging network effects to create scalable, shared defense systems. With a knowledge graph comprising over 500 million identities and a recent $30M Series B led by Sorenson Capital, Bureau enables real-time risk intelligence sharing across platforms. This model creates network effects that work against fraudsters: the more institutions that join Bureau's network, the stronger the defense becomes for everyone. When one bank identifies a synthetic identity, that intelligence instantly protects all connected institutions, making coordinated fraud operations less viable as the network grows.

Early-Stage Disruptors

Startups like Footprint are embedding fraud prevention directly into onboarding infrastructure. Their behavioral detection features—such as identifying typing hesitancy, preventing copy-paste in sensitive fields, and introducing well-placed friction—embody the "design fraud out" philosophy: making legitimate behavior effortless and fraudulent behavior costly. This approach captures fraud signals at the point of data entry rather than analyzing transactions after the fact. Backed by a $20M Series A led by QED and Index, Footprint is scaling an approach that integrates fraud prevention into the core user interface rather than attaching it as a separate security layer.

One to watch is Reken, a seed stage company with limited public information but strong signals. Founded by former Google Trust & Safety lead and veterans from Shape Security, Reken recently raised a $10M seed round led by Greycroft and FPV. Their layered approach requires fraudsters to succeed at both content generation (creating undetectable AI content) and source authentication (establishing trusted credentials) simultaneously. By combining zero-trust source verification with real-time content analysis, Reken forces attackers to solve multiple technical challenges rather than just one, significantly increasing the cost and complexity of successful attacks.

What Wins

The next fraud unicorn will score every interaction, not just transactions. They'll make synthetic identities costly to create, reward safe actions while slowing down suspicious ones, and share signals across platforms in real-time.

The companies that succeed will be the ones that fundamentally restructure how trust works online, making fraud economically unviable rather than just harder to detect.

Signal #001: Prevention First Fraud Architecture – Emerging