Incognia

Your food order is being delivered. Someone is spoofing it.

A driver on a gig delivery platform opens the app in São Paulo. The platform sees the login. It sees the device. It sees the GPS coordinates. What it doesn't see - unless it has Incognia - is that the GPS coordinates are fake, the device was reset three times this week, and the account was created with a cloned identity. The platform thinks this is a regular driver. Incognia's system flagged it before the order was even accepted.

This is what Incognia does. Not with facial recognition. Not with a government ID scan. Not with a two-factor authentication prompt that frustrates the driver who actually owns that phone. With location signals. With device behavioral patterns. With an AI system that has processed enough legitimate-versus-fraudulent behavior to tell the difference in milliseconds.

The company sits quietly at the infrastructure layer of some of the world's largest gig economy platforms, financial services apps, and consumer marketplaces. You probably haven't heard of them. The fraudsters wish they hadn't either.

Fraud prevention was broken long before anyone admitted it.

The standard playbook for fraud detection goes something like this: watch for red flags after they appear. Block the transaction. Investigate. Apologize to the legitimate users you blocked by accident. Repeat.

The industry built elaborate systems around this reactive approach - rules engines, velocity checks, IP blacklists, email verification chains. The problem is that fraudsters read the same papers as the fraud teams. They adapted. They started using VPNs. They bought device farms. They cloned legitimate-looking accounts. They learned to spoof GPS coordinates so convincingly that the fraud system thought they were standing outside your house when they were actually running a script from a warehouse in another country.

Meanwhile, the legitimate user - the actual grandmother ordering groceries, the actual freelancer logging into Upwork from their regular coffee shop - kept getting friction. More verification steps. More "are you sure you're human" prompts. The fraud prevention industry, despite spending billions, was simultaneously failing to stop fraud and making life harder for real people.

The insight Incognia bet on: your location behavior is more uniquely you than your password, your email, or your face. Not your address - your patterns. The Wi-Fi networks your phone encounters. The GPS drift signatures of your specific neighborhood. The timing and sequence of the places you actually go. A fraudster can fake your credentials. Faking six months of nuanced location behavioral data is a different problem entirely.

André Ferraz had already solved location. The fraud problem found him.

Before Incognia, André Ferraz built In Loco, an ad-tech company in Recife, Brazil that used location intelligence for mobile advertising. It grew. It got noticed. Magalu - one of Latin America's largest retail conglomerates - acquired it. Ferraz had built and sold a company by his late twenties.

He could have stopped there. Instead, he noticed something while building the ad-tech stack: the same location signals that made In Loco's ad targeting unusually precise could identify individual users with a level of certainty that no existing identity system had attempted. ACM, IEEE, and Microsoft Research had all recognized the underlying technology's accuracy. The question was whether that recognition translated into a product someone would actually pay for.

It did. Incognia launched in 2020, specifically targeting the fraud use case - a market with real, measurable pain, where the ROI of a better solution was direct and quantifiable. Ferraz and his team went after gig economy platforms first, where fraud is endemic, the cost is direct, and the decision cycle is short. Series A came in June 2022 from Point72 Ventures. Series B came in January 2024, $31 million led by Bessemer Venture Partners.

The numbers above come directly from Incognia's published case studies and press materials. In fraud prevention, a 0.0001% false positive rate is the kind of thing that makes competitors nervous.

From Recife to nearly a billion devices.

A fraud detection system that knows your neighborhood better than your landlord.

Incognia's core product ingests signals from GPS, Wi-Fi networks, and Bluetooth - all without capturing your name, email, phone number, or any other directly personal identifier. It builds a behavioral fingerprint of how, where, and when you use your device. That fingerprint gets compared against what's expected every time you log in, initiate a transaction, or trigger a risk check.

The system flags anomalies that traditional fraud tools miss. Someone logging in from a device that was factory-reset three days ago and has never established a location behavioral pattern. A driver claiming to be in Austin whose GPS signature is consistent with an emulator running on a server farm. A new account whose "home" location doesn't match any real-world location behavior. These aren't things a password check catches. They're not things a CAPTCHA catches. Incognia catches them.

The product line covers the full fraud lifecycle: from account creation through every transaction. No single point of failure, no single layer a fraudster can bypass alone.

What the data actually shows.

The platforms that decided the old fraud tools weren't good enough anymore.

Grubhub adopted Incognia to protect its delivery platform - the kind of reference customer that says something about the quality of the underlying technology. Delivery Hero, one of the largest food delivery companies in the world, rolled Incognia out across six of its brands and reported a 7x return on investment. Upwork, the freelance marketplace with chronic trust and safety challenges, brought Incognia in to improve marketplace integrity.

These aren't pilot customers. These are scaled deployments across platforms that collectively touch hundreds of millions of users. When over 50% of the world's top delivery companies are running your fraud stack, the product has moved past the "interesting idea" stage.

$46.5 million to make fraud prevention actually work.

200% net revenue retention and a 100% trial-to-customer conversion rate are the kinds of numbers that make Series B terms easier to negotiate. The company tripled revenue in the year preceding the Bessemer round - not a coincidence that those two facts appeared in the same press release.

Privacy-first identity is a regulatory inevitability. Incognia built it before it became mandatory.

Incognia's pitch is not just "we're better at catching fraud." It's that the industry was heading toward a wall: increasing regulatory scrutiny on biometric data, growing user resistance to friction-heavy verification flows, and a fraud ecosystem that had adapted to every rules-based countermeasure in the market. The traditional fraud prevention toolkit was producing diminishing returns.

The location-behavioral approach sidesteps the biometric data problem entirely. Incognia doesn't store your face. It doesn't know your name. It processes signals from your device - signals that collectively form a pattern as unique as a fingerprint, but that don't require you to hand over anything personally identifiable. The system works better for users and creates fewer regulatory headaches for the platforms deploying it. In a GDPR world, that combination is not a small advantage.

Add generative AI to the fraud side of the equation - and the 2025 threat landscape includes AI-generated fake accounts, synthetic identities, and automated fraud farms that traditional detection can barely keep up with - and Incognia's timing starts to look less like luck and more like a calculated bet that paid out.

Back to that delivery driver in São Paulo.

The fraudster with the cloned account and the spoofed GPS never made it past the login check. The platform flagged the device anomaly, suppressed the account activation, and moved on. The whole thing took milliseconds. No friction for the tens of thousands of legitimate drivers who logged in that same day. No manual review queue. No customer support ticket from an angry real driver who got blocked by accident.

That's the business Incognia is building: fraud prevention that's invisible when it works, and decisive when it needs to be. In a world where the fraud tools came before the fraud tactics, that used to be impossible. Incognia made it routine.

Nearly a billion devices. $46.5 million raised. Over 50% of the world's top delivery platforms. The company is four years old.