Fraudio is a Fintech startup based in Amsterdam that focuses on helping companies in the payment ecosystem to fight payment fraud and financial crime utilising artificial intelligence, machine learning, and multi-dataset network effects. Our mission is to connect merchants, payment service providers, merchant acquirers and card issuers of all sizes to a powerful centralized AI / smart brain that prevents, detects, and fights fraud in real-time, creating unrivalled value.

Fraudio boasts a proprietary plug & protect centralised artificial intelligence brain. This brain does not require costly configuration, facilitates easy integration and continually learns from all transactions. This makes Fraudio a Generation 3 provider - implementing a disruptive leap from rules-based and machine learning-based solutions trained on individual customer’s data.
Fraudio sets no barrier to entry: our accessible, democratic solution is on a pay-per-use basis only.

Thanks to the patent-pending technology behind our smart brain, we can disrupt the payments fraud industry by moving away from a professional service approach and into a modern SaaS solution - to the benefit of consumers’ overall payment experience and our customers' bottom-line results.

Fraudio gives access to a top-performing, top quality, yet deceptively simple fraud detection and prevention API that, within real-time, returns advanced fraud-related AI insights about customers transactions. This allows every one of our customers to maintain conversion rates while reducing the direct and indirect cost of fraud - maximising revenue.

Fraud is rising, and fines for poor risk management are consequently growing. Financial institutions, especially in Europe, are under increasing scrutiny.

This problem impacts all institutions touching payments, with more than 80% of organizations affected by payment fraud. Merchants lose up to 1.8% of revenue to fraud and spend up to 23% of operational budgets on managing it. Payment Service Providers (PSPs) are in a race to zero on transaction processing fees and subsequently need to offer customers ancillary services. In an increasingly competitive marketplace, and rapidly developing ecommerce marketplaces, all contribute to the rising tide of fraud. Licenses to operate can be lost if fraud ratios become too high. Acquiring Banks can be fined by the Card Schemes and regulators if they don’t control and actively mitigate the prevalence of fraud.

Additionally, money laundering by micro-merchants and instant onboarding results in $350bn laundered per year, with easy digital onboarding increasing merchant fraud and scams at the expense of acquirers. Issuing banks and lender customers face increasingly sophisticated fraud and scam attacks. Money Laundering is crucial for EU regulators as banks have extensive and expensive teams working with outdated techniques and technology.

Existing solution providers in the transaction fraud and money laundering spaces have some degree of efficacy but are hugely inefficient: The implementation process is long and unnecessarily expensive, resulting in an under and poorly-served market. The existing professional service-based business model means vendor’s businesses scale poorly - these current business models require replication of technical teams and are consequently unaffordable to small/medium players. They take a long time to produce models and results for their customers, with proof-of-concept (PoCs) that take six months and cost around six figures - the industry norm.

Usually, the steps of professional services based integration of detection systems include at least:

• Historical data transfer; customer collects and transfers historical data to the vendor.
• Data science teams assigned to work with that customer’s data.
• Custom product development, training and optimizing, customized machine learning models, and often also rule-setting to further try to improve protective measures.
• Localised product testing, problem-solving, and refinement.
• Deployment and hosting of the customers, non-transferable, tailored solution.

Fraud analysts were initially performing manual fraud analysis and using rule-based systems. With improving and maturing of technologies, top-tier vendors began to offer fraud detection systems that use artificial intelligence. These systems were created to facilitate analysis and to automate decision making. The application and use of artificial intelligence effectively improved the accuracy of fraud detection and systems - now detecting larger fractions of fraudulent transactions than ever before. Fraud detection sophistication increasingly determined a company’s value proposition and therefore became proprietary technology isolated within banks, credit card networks or schemes, and payment service providers.

These professional services based on proprietary fraud detection systems provided a negligible network effect. Credit and debit card transaction records, though collected by everyone, have no unifying standard, making this data highly diverse. This diversity reinforces already secluded data silos within each company and prevents aggregation, thereby failing to leverage the totality of this data.

Take, for instance, a card issuing bank. A vendor processing their transactions will have access to every payment that this bank’s cards make. The typical vendor will produce machine learning models that will learn from behaviours and patterns specific to this bank’s customers. Suppose the same vendor also processes transactions from another bank, in the same country. In that case, it will isolate the data sets from these two customers and produce individual machine learning models - the opportunity to produce stronger insights by aggregating data from the same segments is lost. Often this forced segregation is not able to capture underlying patterns and behaviours that are customer independent and that occur homogeneously across one specific segment.

Incumbent vendors, using this siloed data approach, therefore need to start from scratch for each new project. They need to go through lengthy data exploration processes and adapt their complex machine learning pipelines for each new customer. Because of this cumbersome and highly inefficient business model, their customers are forced to sign multi-year contracts with exorbitant setup costs and annual fees (up to €2m).

Fraudio's Next Generation Solution:

Fraudio has applied for a patent on a state-of-the-art method for applying a centralised artificial intelligence system for fraud detection. The AI brain is trained in batch using a centralised data set, with a known and very well understood data schema, to score events from multiple streams of data in real real-time (<100ms on average), i.e., to streams of card payment transactions from diverse customers.

Data sets are aggregated into one single extensive data set using advanced data transformation techniques to map these new data streams into our central data schema. In a nutshell, each inbound stream of card payment transactions has its data language as spoken by that specific customer’s systems. Upon integrating with Fraudio’s API, each of those streams are instantly translated into the same data language spoken by Fraudio’s Artificial Intelligence brain.

This is a novel and disruptive approach to the problem of fraud detection in payment transactions, as it is opposed to having different artificial intelligence models specifically - and laboriously trained for each stream of payment transactions.

But what value does this bring for our customers?

Fraud detection SaaS with simple and easy API integration & pay-for-use pricing

Because of this technological leap forward, we can significantly lower the barrier to entry into the realm of top tier fraud detection through accessing Fraudio’s high-end fraud-fighting power via a developer-friendly API - fully hosted in the cloud.

We have three products which can be connected to in days, not months:

• Credit/debit card transaction fraud scoring.
• Merchant fraud/scam monitoring.
• Money Laundering detection.

Our USPs

• No integration/setup costs vs competitors six-figure demands.
• Fraud scores from day zero, with responses below 100ms on average.
• Reduce false positives and maximise conversion while cutting operational costs of fraud management.
• Machine learning capabilities continue to drive improvements for all customers, allowing them to benefit from our strong network effect.

Proof points:

We're a team of experts from the payment industry, backed by ING bank, Payvision and Viva Wallet. Our products are powered by ever-improving centralised models trained on a proprietary dataset of over 1B transactions.

Together with the largest card issuer in Portugal, we have proven to make 15 times fewer false positives than the fraud detection system from one of the largest card schemes. We make 30 times fewer false positives than custom made ML models for a leading German PSP and deliver 40% fewer false positives than one of the top 3 market leaders who serves a Dutch merchant acquirer.

‍