AI will define the next generation of fraud prevention — on both sides, thinks Laurent Sarrat, CEO and co-founder of Sis ID, a French fintech that provides collaborative payment-fraud prevention solutions.
When AI tools are both advanced protection and a growing threat to payment companies, major challenges arise. PaySpace Magazine Global had an exciting conversation with Laurent Sarrat about novel fraud techniques, AI-driven fraud prevention solutions, the role of new EU Instant Payments Regulation and Verification of Payee (VoP) service, and much more.
Laurent Sarrat is the CEO and co-founder of Sis ID, the company behind ‘My Sis ID’ platform, which enables businesses to verify supplier and beneficiary bank details, preventing fraudulent transfers and ensuring data integrity. Built by finance professionals for corporate users, the company helps secure the purchase-to-pay process and reduce payment fraud risks worldwide.
Fraud prevention is evolving with the advent of AI technologies. However, so do fraudulent techniques. Can you tell us more about AI use in unauthorised fraud prevention, and how often do you spot AI use by criminals?
64% — that’s the number to remember. Today, nearly two-thirds of fraud attempts involve some form of AI. It’s no longer something you can “spot” — AI is now embedded throughout the fraud lifecycle. In some cases, the entire scam is AI-generated, from deepfakes to fake CEO voice calls.
When we detect a fraud attempt, we assume AI played a role. The key questions are when and how — and how quickly it’s learning. AI targets the weakest link in any security system: human interaction. As it evolves, so must our defences.
Today, agentic AI is emerging as a facilitator of new e-commerce and payment models. What are the associated risks from a cybersecurity perspective? How are security tech professionals preparing to meet new challenges linked to payment protocols for agentic commerce?
Gartner estimates fewer than 1% of companies currently use agentic AI — but adoption is rising fast. Its goal is to understand customer behaviour and automate strategic actions to drive growth.
From a cybersecurity standpoint, the biggest risks are compliance and decision accuracy. AI agents can initiate payments independently, so they must operate within strict frameworks such as GDPR. When implemented correctly, agentic AI can enhance security — provided payment service providers and banks fully align with standards like Confirmation of Payee (CoP) and Europe’s Verification of Payee (VoP). These safeguards are essential at the final point of transaction.
Image: Sis ID
What type of fraud remains the toughest challenge to tackle, despite all the technology advancements, and why is it so difficult to eliminate?
Human-focused attacks remain the toughest to stop. Even the strongest system can be undone by a single moment of misplaced trust. Fraudsters win by sounding convincing — through realistic emails, voice clones, or urgent payment requests.
The answer is constant awareness, backed by technology that minimises human error. At Sis ID, we integrate directly into our clients’ systems so payment data stays consistent, verified, and consolidated. Every transaction is scored in real time, helping us spot mismatches before money moves.
Can authorised push-payment fraud levels significantly decrease with the help of security tech, or is public awareness/education the only way out of being manipulated?
Both education and technology are essential. People need to recognise when something feels off — but tech must step in before a mistake becomes a loss.
CoP in the UK and VoP in Europe add that crucial last check: an automated name verification before funds leave an account. It compensates for human vulnerability and ensures the recipient truly matches the intended payee.
Sis ID has received a fair share of negative public feedback from customers regarding some slow and manual account verification procedures, which seem outdated in the age of extensive automation. How are you addressing these dissatisfactory use cases? Are you implementing any tech innovations into the verification of the payee procedure?
We process around three million checks every month, protecting customers in more than 200 countries. Accuracy and integrity come first.
Today, nearly 90% of checks are fully automated. When manual review is needed — around 3,000 cases per month — our accredited team steps in to ensure absolute reliability.
We have about 60 public reviews, the vast majority positive. The small share of negative comments mainly relate to timing — and we’re addressing that through automation and continuous improvement.
Since VoP went live in Europe last October, we’ve partnered with multiple PSPs — including BNP Paribas — to launch a fully compliant solution, now recognised by the European Payments Council as a Registered Verification Mechanism (RVM).
There is often a fragile balance between enhanced security requirements and customer experience. How to make sure that your security prevention methods are not being overzealous and not leading to false fraud alerts?
This was an early challenge for the UK under CoP, and Europe can learn from that. The key is alignment. When PSPs and banks follow consistent standards — especially on “no-match” outcomes and name-matching rules — customers get strong protection without added friction.
Fraud prevention must be robust, but never at the expense of trust or usability.
In your opinion, what does the next generation of fraud prevention technology look like? Which innovations in the segment seem most promising?
AI will define the next generation of fraud prevention — on both sides. What matters most is how responsibly and rapidly the financial ecosystem adapts. It’s an exciting time, and we expect major advances in the years ahead.
