The suspicious customer looks almost boring. The name is ordinary. The card clears its first check. The address exists. But the email account appeared yesterday, the phone has no convincing online history, the device is running through an emulator and three other accounts have touched the same IP. None of those facts proves fraud. Together, they start to tell a story. SEON is in the business of assembling that story before money, merchandise or trust leaves the building.
The Austin-headquartered company sells a cloud platform for fraud prevention, identity verification and anti-money-laundering compliance. Its customers feed it events from the moments that matter in a digital business: signup, onboarding, login, account recovery, checkout, payment and investigation. SEON enriches those events with more than 900 live signals, applies customer-written rules and machine-learning models, and returns a decision or a case for human review.
That description sounds clinical because the problem is not. A fraud team lives between two costly errors. Approve the wrong person and the company may absorb a chargeback, a stolen account or a regulatory problem. Block the right person and the company has just escorted a good customer to the door. The useful product is not merely a red light. It is a light whose wiring an analyst can examine.
Born from an expensive annoyance
SEON has the sort of origin story founders usually wish had happened to somebody else. Tamás Kádár and Bence Jendruszák were university friends building a cryptocurrency business when fraud attacks began to overwhelm it. The tools they tried felt slow, expensive or awkward enough to damage the customer experience. They wrote their own system, other crypto and high-risk merchants asked to use it, and in January 2017 the defense became the company.
The founders made digital footprint analysis the early wedge. A fake identity can borrow a name and type a plausible address. Reconstructing years of ordinary online life is more difficult. SEON checks an email, phone number and IP against hundreds of online services, then combines that footprint with device data: browser and hardware traits, proxies, VPNs, emulators, remote-access tools, geolocation mismatches and links to other users. Behavioral signals add the way a person moves a mouse, types, touches a screen or handles a device.
The point is not to make a single clue sound magical. New customers have young emails. Travelers use VPNs. Families share devices. SEON's pitch is that context, velocity and connections are harder to fake in combination. A risk team can set thresholds to match its own appetite, label outcomes and watch models tune themselves on the business's actual results. It can also write a rule in ordinary language, inspect the generated logic and change it without filing a request with engineering.
“We wanted a way to proactively stop fake sign-ups without adding friction for real users.”Olivier Novel, co-founder of Cardmarket
From a score to a command center
SEON no longer sells only the moment of fraud scoring. Its platform now follows risk across six stages. It can pre-screen a person before a costly know-your-customer check, verify documents and faces during onboarding, monitor behavior at login, inspect payments, organize investigations and prepare regulatory filings. Customer and payment screening cover sanctions, politically exposed people, watchlists and adverse media. Transaction monitoring looks for suspicious movement after an account is active.
The expansion became concrete in 2023, when SEON acquired Budapest AML specialist Complytron. Instead of building every compliance database and matching algorithm from zero, it brought in the technology and team. By 2025, SEON was presenting fraud and AML as one operating layer. In 2026 it added identity verification, network detection for coordinated rings, an AI chart builder and a server that lets approved external AI tools query SEON risk data.
This is a meaningful market move. Fraud teams and AML teams have traditionally bought different systems and maintained different queues, even when they are looking at the same customer. A synthetic identity can be both an onboarding problem and part of a mule network. An account takeover can become a suspicious transaction. SEON wants the device clue, identity check and payment trail to remain attached as the case moves between teams.
Who buys it, and how
The customer list follows wherever digital money and incentives attract inventive strangers. Fintechs and lenders use SEON to screen applicants and monitor transactions. Payment providers use it across merchants and methods. Gaming operators look for bonus abuse, multi-accounting and account takeover. Retailers look for bot-driven registrations, chargebacks and refund abuse. Named customers and case studies include Revolut, Bilt, Nubank, Afterpay, Lottoland, Payop, Carbon, Cardmarket and tbi Bank.
The software arrives mainly through APIs and a browser-based administration panel. A public Starter plan lists $699 a month for 2,500 fraud checks, ten users and 50 custom rules. Premium pricing is tailored to volume and can include the AML suite, unlimited rules, implementation help, round-the-clock support and managed risk services. The latter gives a customer a specialist who watches performance and tunes its strategy, a useful admission that sophisticated risk software does not run itself.
SEON also sells through AWS Marketplace, which can turn a new security vendor into an item on an existing cloud bill. A global partner program brings in consultants, integrators and referral partners. Its 2026 work with commerce agency Domaine produced a Shopify-native application, an attempt to put enterprise-style signals closer to merchants that do not have a large risk engineering team.
“SEON gives us system-wide visibility. Even if fraud starts on one merchant, we block it across the platform before it spreads.”Anastasia Semenkova, CEO of Payop
A crowded market, a visible argument
SEON competes in several overlapping neighborhoods. Sift, Forter, Signifyd and Kount are familiar names in commerce fraud. Feedzai and Featurespace have deep positions in financial institutions. Sardine and Unit21 join fraud and compliance workflows. Socure, Sumsub and Persona overlap in identity. BioCatch specializes in behavior; Fingerprint in device intelligence; ComplyAdvantage in financial-crime data. A buyer can choose a broad platform, assemble specialists or use tools already bundled with a payment processor.
SEON's place in that market is the configurable middle: broader than a point-data API, more self-service than a traditional bank stack, and increasingly wider than fraud alone. Its differentiator is not one model. It is the combination of first-party enrichment, visible reasoning, fast integration and rules the operator controls.
Transparency matters because fraud models make consequential mistakes. SEON emphasizes “white-box” decisioning: the analyst can see which signals pushed a score, preserve an audit trail and explain an action to a colleague or regulator. That does not make the system automatically correct. It makes disagreement possible. In a field fond of black boxes, the ability to ask why is a practical feature.
The company reports tangible customer results, though each belongs to a particular deployment rather than a universal promise. Carbon says it automated more than 95 percent of fraud checks. tbi Bank reports a 5 percent increase in loan approvals. Lottoland says it cut bonus abuse by 70 percent. Payop protects more than 500 payment methods. Cardmarket screens more than 2,000 registrations a day. These examples show the range of the product, and also why buyers must tune it to their own economics.
The next contest is over context
SEON raised $12 million in 2021, $94 million in 2022 and $80 million in a 2025 Series C led by Sixth Street Growth. The company says total funding is $187 million. That capital has supported a shift from a Hungarian-founded fraud startup into a global operation with major offices in Austin, London, Budapest and Asia-Pacific, and a public team count above 300. Its latest certifications extend identity verification into regulated European and German use cases.
The geography has shaped the company in less obvious ways. Product engineering grew from Budapest while sales and support spread closer to customers in the United States, Europe and Asia. SEON describes an unusually flat internal culture, with trust, transparency, customer obsession, innovation and ownership as its stated values. Its careers material promises hybrid work, flexible hours, employee equity, language courses and a book allowance. Those details matter in fraud software because the work crosses disciplines: an engineer has to understand the analyst's queue, a compliance specialist has to understand the data model, and a salesperson has to resist promising that any tool can catch everything. The platform's emphasis on inspectable decisions reads like a product version of the same culture. If the reasoning is visible, somebody can challenge it, improve it and own the result.
AI will make both sides faster. Criminal groups can generate identities, automate browsing and vary attacks cheaply. Risk teams can summarize cases, discover networks and draft filings faster. The advantage may not belong to whoever says “AI” most often. It may belong to whoever owns useful context, keeps it current and lets a human understand what the machine did.
That returns SEON to the suspicious customer. The business is not built on knowing that one odd dot is guilty. It is built on seeing where the dot came from, what else it touched and whether the pattern has appeared before, then giving an operator enough evidence to make the next move. Fraud rarely introduces itself. The clues have to do the talking.