With AI, fraud is becoming more diverse, tools are becoming more widely available… and the fear of being detected is diminishing
A fraud market that’s becoming more professional, general-purpose generative AIs that promise significant productivity gains… Fakes have never been so close to the real thing. Finovox helps its clients avoid the pitfalls.
Question: Would your company be willing to entrust its customers’ sensitive data to general-purpose AI systems operated by foreign entities?
Today, in fact, systems like ChatGPT, Claude, and Gemini are capable of automatically analyzing sensitive documents to facilitate the creation of compliance files, particularly as part of KYC (Know Your Customer) processes: identification documents, proof of address, contracts, articles of incorporation, financial documents… At the same time, in the hands of fraudsters, these same general-purpose generative AI systems are valuable tools for creating forgeries!
“While these new features promise significant productivity gains for businesses, they raise numerous questions regarding privacy, data governance, and document authenticity,” adds Marc de Beaucorps, co-founder & CEO of Finovox. Indeed, how are these documents processed once they’re transmitted? What safeguards are in place to protect this data? And, above all, how can we ensure that the analyzed documents are authentic and have not been tampered with before being included in a compliance file?
Fraudsters’ risk profile has changed
In a recent study, Finovox showed how generative AI and the digitization of communications stand out as the two factors most strongly associated with the evolution of fraud. “This increase, as Marc de Beaucorps notes, is accompanied by a normalization of the act of committing fraud. Photoshop, Word, Canva, and even online PDF editors are now widely accessible, often for free; they make it easy to produce visually convincing forged documents without any special technical skills…”
The most significant shift concerns fraudsters’ attitude toward risk, which underwent a profound change between 2024 and 2026. In 2024, 69% of them reported being unaware of the legal consequences of their actions. By 2026, 64% were committing fraud knowingly and without fear. And 65% say they are willing to do it again, compared to 3% two years earlier… “We’ve gone from ignorance to open indifference!”
Similarly, fraud schemes are becoming more diverse, the tools are becoming more widely available, and the fear of being detected is diminishing. The French company Finovox —which is based in Luxembourg within the incubator Le Village by CA—is countering general-purpose AI with its own AI, specialized in verifying document authenticity and detecting document fraud, which is now operational in about 100 companies across 70 clients.
Fraud Is Becoming More Convincing
“The rise of AI in document-related processes is undeniable. But it must be accompanied by careful consideration of data protection and the reliability of the analyzed documents,” insists Marc de Beaucorps. “We go a step further by focusing on document consistency—by analyzing not only the content but also the structure.”
In fact, fraud schemes are becoming more convincing—undetectable to the naked eye—with the use of AI-powered image and document generation tools. This is also why Finovox has integrated the analysis of documents’ digital fingerprints into its solution to detect repeat fraud.
To combat forged documents, supporting documents, and certificates, Finovox relies on technology based on models trained using anonymized document data and historical fraud patterns, from which it continuously learns to improve.
Finovox’s strength also lies in data sharing, while complying with GDPR regulations. “The more we’re integrated into the workflows of banks, insurance companies, and public agencies, the more we learn—and the more robust the system becomes!”
Fraud Is Becoming More Sophisticated
The solution is also designed to integrate directly with clients’ tools via APIs to facilitate fraud detection. Clients are estimated to have prevented over 100 million EUR in losses as a result.
During the onboarding and KYC phases, Finovox verifies each document in real time (passport, ID card, driver’s license) to eliminate forgeries from the very start of the relationship. When it comes to claims processing, it automatically blocks falsified invoices, estimates, and supporting documents to prevent improper payments, while speeding up claims processing. Finally, in the area of financing, it secures loan disbursements by verifying the validity of balance sheets, pay stubs, or bank account information before any payment is made.
Development is ongoing, accounting for a good half of the company’s 45 employees—which closed an 8.2 million EUR Series A funding round led by TX Venture last June. “Fraud is becoming more professional,” notes Marc de Beaucorps, “whereas in the past, it was the domain of opportunists. I see a similarity with cybersecurity. Actors are selling their methods and tools. Some even go so far as to provide training. And all of them, of course, are leveraging the high performance of AI! That’s why we’re honing our expertise in technology and, even more so, in our clients’ business lines.”
