A finance team can approve the right invoice, follow the right process and still send the money to the wrong bank account.
That process is behind Basware’s acquisition of Trustpair, the Paris-based payment fraud prevention company. Announced in August, the proposed deal would extend Basware’s invoice controls through the payment stage, where altered supplier details or a compromised email can redirect money after the invoice itself has already passed review.
“A finance team can do everything right on the invoice and still send the money to the wrong bank account,” Basware CEO Jason Kurtz said when the acquisition was announced. “Basware has spent more than 40 years protecting the invoice. Bringing Basware and Trustpair together extends that protection through the payment itself.”
The strategic logic is fairly direct. Basware focuses on whether an invoice is legitimate, correctly processed and approved. Trustpair checks whether the bank account receiving the payment actually belongs to the intended supplier.
Those two controls used to look like different problems. AI-enabled fraud is making the space between them harder to ignore.
Fraud Is Getting Better at Looking Ordinary
Supplier payment fraud rarely needs an obviously fake invoice. A criminal can compromise a real email account or alter payment instructions around a legitimate transaction, which means the invoice may look exactly as expected while the bank details do not.
Generative AI lowers the cost of making those attacks more convincing. According to the Association of Certified Fraud Examiners, 75% of anti-fraud professionals reported an increase in generative AI document fraud and forgery over the previous two years. Just 7% said their organizations were more than moderately prepared to detect or prevent AI-powered fraud.
Kurtz sees invoice fraud as one of the areas where finance teams will have little choice but to use AI defensively.
“Fraudsters are using AI tools to help breach organizations,” he said in a recent Q&A. “We need AI to combat these bad actors. Invoice fraud is a perfect place to leverage AI as it can learn from trends being seen across thousands of companies.”
Trustpair approaches that problem through bank-account validation. This approach has become increasingly relevant after Nacha tightened its fraud-monitoring requirements for ACH participants in 2026.
Nacha's expanded 2026 ACH fraud-monitoring rules require applicable financial institutions, non-consumer payment originators and certain third-party participants to implement risk-based procedures intended to identify potentially fraudulent transactions, including payments induced by false pretenses. The rules do not prescribe a single detection method or require every transaction to be screened before processing.
That puts more pressure on finance teams to catch a bad payment before it leaves rather than relying on recovery afterward. Faster payment rails make that window smaller.
The Control Point Is Moving Closer to Payment
The acquisition also says something about where financial controls are heading. Invoice automation has spent years removing manual work from capture, matching and approval. Fraud prevention now has to follow the transaction further.
Basware and Trustpair together bring intelligence drawn from 2.5 billion invoices, 20 million suppliers, and more than 10 million verified supplier-to-bank-account pairs. The value of that scale is less about the raw number than the patterns buried inside it. A system watching supplier behavior across a large transaction base can potentially spot changes that look unusual before an individual finance team would know to question them.
That still does not eliminate the need for human judgment.
“AI should surface the signals that make a payment worth investigating,” said Baptiste Collot, Trustpair’s cofounder and CEO. “A human should make the call before you block a payment or accuse a supplier, because false positives carry real relationship and legal costs.”
That division is especially important in fraud prevention, where stopping every unusual payment would create its own operational mess. The useful role for AI is to narrow attention: identify the bank-account change or transaction pattern that deserves another look, then give a finance professional enough context to decide what happens next.
Basware’s acquisition of Trustpair pushes that idea beyond the invoice itself. The company is betting that financial integrity increasingly has to cover the entire path money takes, right up to the account that receives it.
For finance teams, the distinction is becoming harder to ignore. Approving the correct invoice is only half the job if the payment can still be diverted after everything else checks out.
