AI is helping scammers produce convincing fraud attempts faster and at lower cost. For finance teams, the challenge isn't just spotting suspicious messages. It's verifying that payment instructions are legitimate.
Global scam losses reached an estimated US$442 billion in 2025, according to figures cited in a recent PYMNTS report.
The report also highlights an analysis by cybersecurity company ZeroFox, prepared for Bloomberg, which found that phishing pages hosted on the free tiers of nine developer platforms and AI app builders had increased more than fourfold over two years, reaching 10,600.
These findings point to a shift in the economics of fraud. AI is making it easier and cheaper to create convincing scams, allowing criminal networks to launch more attempts with fewer resources.
For businesses, that raises a practical concern: fraudulent payment instructions are becoming harder to distinguish from legitimate ones.
AI is lowering the cost of deception
Scammers can now use AI tools to generate phishing websites, write convincing messages and impersonate legitimate businesses without extensive technical expertise.
Some of these capabilities are already widely accessible. As Eftsure recently explored, cloning a person's voice can now take just three seconds of audio and little to no financial investment.
The financial incentive is significant, too.
INTERPOL's 2026 Global Financial Fraud Threat Assessment reports that AI-enhanced fraud is 4.5 times more profitable than traditional methods.
This is particularly concerning given the adaptability of organized fraud networks. As covered in Eftsure's earlier analysis of the scam economy's evolving business model, criminal operations are adapting their methods to sustain profitability.
AI gives those operations another way to reduce costs and increase the volume of attempted fraud.
When a convincing message becomes a fraudulent payment
For finance teams, one of the most relevant threats is payment redirection fraud.
A criminal impersonating a legitimate vendor might request updated bank account details or submit a fraudulent invoice. The communication may reference a genuine business relationship, use a familiar writing style and contain no obvious errors.
AI makes this kind of impersonation easier to produce.
The problem is that traditional warning signs are becoming less reliable. A professional-looking email, accurate company details or even a familiar voice doesn't establish that a payment request is authentic.
And while an accounts payable employee may follow the organization's established approval process, that process doesn't necessarily verify who controls the receiving bank account. An approved payment isn't necessarily a verified payment.
Payment controls need to work even when scams look legitimate
Expecting finance employees to identify every AI-generated scam isn't a sustainable defense.
Instead, organizations need controls that don't depend on recognizing fraudulent communications. That means strengthening verification at the points where fraudulent instructions can become actual payments.
Finance leaders should consider whether their organization can:
- Independently verify vendor bank details: Confirm account information through a trusted verification process rather than relying solely on details supplied by email
- Control changes to payment information: Verify bank account changes before updating an existing vendor's payment details
- Identify payment risks before funds leave: Check payment details against verified vendor information before releasing payments
- Maintain an audit trail: Record how vendor information was verified and who authorized changes
These measures won't eliminate every fraud risk. But they can reduce the likelihood that a convincing impersonation attempt results in money being sent to a criminal-controlled account.
The case for a unified trust layer
As AI makes fraudulent communications harder to distinguish from legitimate ones, businesses need a more reliable basis for deciding who to trust. For finance leaders, that means moving beyond isolated checks and manual approvals toward verification that connects vendor identity, bank account information and payment decisions.
It's the thinking behind Eftsure's vision for a unified trust layer and a new standard of payment integrity.
Because when scammers can make almost anything look authentic, businesses need more than convincing information: they need independent verification before money moves.
