The purchase looked wrong before fraud software had a chance to say so. A passenger was flying from New York to Accra, Ghana; the credit card carried a California billing address. In 2001, Mark Michelon was an accountant at Orbitz who had just inherited a stack of chargeback memos and the title of fraud analyst, more or less overnight. One odd itinerary became a clue. The stack became a systems problem.
Six years later, Mike Long and a group of travel-industry executives founded Accertify, and Michelon joined them. They built a way to pull scattered signals into a coherent picture of a transaction. The first fraud the young company says it caught was almost theatrically appropriate: tickets to a Los Angeles performance of Wicked, bought through Ticketmaster.
That anecdote still explains the business. Digital fraud is rarely one conspicuous fact. It is a collection of details that become suspicious together - device, location, behavior, identity, payment history, account changes and what happened after the sale. Accertify sells large companies the software and expertise to assemble that picture fast enough to make a decision without making a legitimate customer feel like a suspect.
A decision layer, not a digital bouncer
The old mental model of fraud prevention was a guard posted at checkout: score the order, approve or decline, move on. Accertify now describes a much longer beat. Its platform watches for fraudulent account creation and account takeover before checkout; evaluates card-not-present risk at payment; looks for refund, return and policy abuse afterward; optimizes strong customer authentication; and automates the dispute that can arrive weeks later.
The portfolio reflects that journey. Account Protection uses device, connection, behavioral and reputation signals to look beyond a correct password. Fraud Decisioning scores transactions in real time. CARE, its first-party fraud product, looks for customers manipulating returns or refunds. Payment Optimization manages authentication rules and exemptions. Chargeback Management collects evidence, routes responses and tracks results across networks and processors. Strategic Risk Services puts analysts and data scientists beside the client team to tune the machinery.
One customer · four risk moments
The connective tissue matters more than the product names. A chargeback is not merely an administrative chore; it tells the system which earlier decision ended badly. A login anomaly can inform the payment decision. A pattern of refunds can change what happens on a future order. Accertify’s thesis is that fraudsters move across departmental boundaries, so the defense needs a shared memory.
Built for the places where a false no gets expensive
Accertify is not primarily aimed at a neighborhood shop installing its first checkout plug-in. It sells to enterprises with enough volume, complexity and international reach to make configurable controls worthwhile: airlines, hotel and travel businesses, large retailers, ticketing companies, marketplaces, financial institutions, and gaming and betting operators. Publicly named clients include United Airlines, Marks & Spencer, Sportsbet, Target Australia, Fnac Darty, Ubisoft, Viator and easyJet. Accel-KKR said, when it acquired the company, that Accertify served 40 percent of the top 100 retailers.
Their shared problem is not simply theft. It is the seesaw between loss and acceptance. Make controls loose and fraud rises. Make them blunt and legitimate orders fall into manual review or disappear at an authentication challenge. The fraud dashboard may look tidy while the revenue report absorbs the damage. Airlines feel that tension sharply: inventory is perishable, transactions are international, loyalty accounts hold value, and a good customer may quite reasonably buy a ticket far from home at midnight.
Accertify’s reported outcomes make more sense viewed as balancing acts. United Airlines increased its chargeback win rate by more than 40 percent. Sportsbet reported fraud losses down more than 80 percent year over year and chargebacks down more than 82 percent. Target Australia reported a 71 percent reduction in chargebacks while keeping 99 percent of customer purchases frictionless. Marks & Spencer said 99.98 percent of customers received a frictionless journey after strong-customer-authentication rules arrived.
Those figures come from individual company case studies, not controlled comparisons, and every merchant has a different baseline. Still, they illuminate what buyers pay for: fewer losses, higher approval, less manual review, better dispute recovery and fewer needless hurdles for recognizable customers.
Scale is useful. Operational memory may be better.
The fraud market does not lack machine learning. Forter, Riskified, Signifyd, Sift, Kount, Cybersource, Stripe Radar, Feedzai and others compete for overlapping budgets. Some rivals emphasize a financial guarantee on approved orders. Network-owned providers can draw on vast payment flows. Smaller tools compete on quick integrations and transparent digital signals.
Accertify’s clearest distinction is the combination of breadth and control. The system spans account activity through chargebacks, but it also lets enterprise fraud teams configure policies instead of surrendering every decision to a black box. Its community data brings patterns seen across recognizable global brands; adaptive models change with new behavior; device and behavioral tools provide identity context; and experienced analysts help clients translate all of it into operating rules.
That last piece is difficult to reduce to a feature checklist. Accertify grew out of practitioners doing the job, not a general-purpose model hunting for a market. Its careers material still makes a point of pairing technology with people. Strategic Risk Services can be fully involved, while customers who want direct control can run the system themselves. The product is therefore both software and accumulated operating judgment.
The data flywheel is equally important. Accertify says it processed more than 10 billion transactions worth over $1.2 trillion in 2025, protected more than 1.63 billion account events and helped stop 30.7 million fraud and abuse attempts worth nearly $5.8 billion. More volume can expose patterns that a single merchant cannot see. The merchant’s own outcomes then sharpen its strategy. The result is not omniscience; it is a bigger set of clues arriving before the decision window closes.
A 14-year detour through American Express
American Express bought Accertify in 2010 for roughly $150 million, three years after its founding. The deal supplied global distribution and payment-industry reach. It also gave Accertify time to extend beyond transaction screening: chargeback management arrived in 2011, the footprint expanded across Europe and Asia, machine-learning tools followed, then account protection and refund-abuse products.
In May 2024, technology investor Accel-KKR completed a carve-out of the business. The price was not disclosed. Accertify became independent again, this time as a mature enterprise platform rather than an eight-person startup. The structure now looks like familiar private-equity-backed enterprise software: direct, demo-led sales; undisclosed contract pricing; integrations with commerce and payment systems; and a mixture of recurring platform access and expert service. Customers can choose self-managed controls or a closer managed relationship.
Independence has brought visible changes. The company opened a Gurgaon office in 2025 for data science and Strategic Risk Services. It partnered with RegTech Solutions to take account protection into Australian superannuation and wealth platforms. In January 2026, six-time technology CEO Greg Dukat joined to lead commercial growth while Michelon remained president, focused on clients and product innovation.
A few months later came the “More Yes” positioning and the Predictive Yes Platform. The phrase is marketing, but it identifies a real shift in category. Fraud prevention has traditionally justified itself by avoided loss. Commerce enablement asks risk software to prove what it allowed: more approvals, fewer challenges, faster operations and recovered revenue. That gives Accertify a larger story to sell - and a larger set of numbers by which customers can judge it.
Can one platform keep its promise of one picture?
Unified platforms are appealing because fragmented tools create gaps and duplicate friction. They are also hard to execute. Account security, card fraud, authentication, return abuse and disputes have different owners, data and clocks. A model that is excellent at checkout is not automatically excellent at recognizing a synthetic account. An airline and a gaming operator may share fraud patterns but not tolerance for the same intervention.
Accertify’s market position rests on making common data useful without flattening those differences. Its configurable controls, vertical experience and partner connections - including Shopify, Salesforce Commerce Cloud, Mastercard, FreedomPay, Computop, Outpayce and VTEX - give it the plumbing. Its analysts give clients a way to adjust the machinery. The open question is whether the newly independent company can package that depth simply enough to expand beyond the very large merchants that built its reputation.
There is a pleasing loop in the company’s history. Michelon’s chargeback pile showed what happens when decisions are separated from outcomes. Accertify gradually connected the two, then moved outward to the account and inward to the payment experience. Its platform now tries to remember the whole encounter: who arrived, how they behaved, what they bought, whether they complained and what the merchant learned.
The objective is not to make fraud disappear, which no credible risk company can promise. It is to make uncertainty manageable at enormous speed. Somewhere inside those 10 billion yearly transactions is another improbable itinerary, another compromised account and another perfectly innocent customer doing something unusual. Accertify’s business depends on telling them apart - and knowing that the valuable answer is often yes.