AI Made B2B Fraud Cheap And Trust Expensive

TL;DR

Prime Big Deal Days · Oct 6–7Offer from Amazon

Get smart everyday buys delivered free — and shop member deals

• Fast, free delivery on millions of items
• Access to Prime Big Deal Days deals on October 6–7
• Prime Video, Amazon Music and more included

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

Bottomline senior risk and fraud officer Katie Elliott says AI lets fraudsters make B2B payment scams broader and faster, while manual verification remains costly for finance teams. Her comments point to a shift toward layered, ongoing supplier checks, with automated screening for routine payments and human review of anomalies.

AI is enabling fraudsters to send B2B payment scams at greater scale, while finance teams still face the expense of checking whether suppliers and payment instructions are genuine, Bottomline senior risk and fraud officer Katie Elliott said in an interview with PYMNTS. Her account points to a growing challenge for corporate payments: verifying counterparties before funds move, without slowing legitimate payments or relying on manual checks for every transaction.

Elliott said attackers use available AI and other tools to make their attempts “bigger, broader, faster.” She described a shift from more targeted activity to mass phishing and spamming, which can raise the number of suspicious requests finance departments must assess. The interview did not quantify how many businesses have been affected or how much fraud losses have changed.

The timing of checks matters because authorized payments can move quickly. Elliott said funds may be taken out as soon as a recipient gains access to them. That makes prevention before authorization more important than relying on recovering money after a mistaken or fraudulent transfer.

She recommended evaluating multiple signals rather than trusting one data point, citing digital identity, phone details, email-domain history and payment information. Elliott also described a model in which AI processes normal activity and sends exceptions—such as changed payment instructions, an unusual rise in payment volume or a supplier requesting an abnormal amount—to a person for review.

Why Supplier Checks Must Scale

The issue is a mismatch in operating costs: attackers can make many attempts and need only some to succeed, while businesses must screen payments without creating delays or staffing large manual-review teams. If AI increases the volume and plausibility of fraudulent requests, checking every payment by hand may not scale.

That pressure could make shared verification services more valuable. Elliott said relationships with third-party data providers and verification tools can be expensive, especially for smaller businesses. Payment networks may be able to spread some of those costs across more transactions, though the interview did not establish how widely such services are available or whether they reduce fraud in practice.

For finance leaders, the operational question is how to combine automated checks with human review of meaningful exceptions. Faster payments leave less time to identify a problem after authorization, increasing the importance of checking supplier identity and payment instructions before funds are sent.

From Onboarding to Ongoing Verification

Traditional accounts-payable controls often include checking a supplier when it is added and reviewing payment details when they change. Elliott’s account suggests that a one-time onboarding check may be less dependable if convincing identities can be generated or impersonated more easily. She said the ability to use AI to create “a whole identity” is a particular concern.

The interview, titled “Stopping B2B Fraud Before AI Moves the Money,” was part of coverage for the PYMNTS B2B Payments Event 2026. Elliott’s comments are an industry practitioner’s assessment of the threat and possible controls, not an independently measured study of fraud trends. The source provides no comparative loss figures, prevalence estimates or evaluation of a specific verification product.

“They are using the technology that’s out there, the AI, every tool available to them in order to make their attempts bigger, broader, faster.”

— Katie Elliott, senior risk and fraud officer at Bottomline, speaking to PYMNTS

The Scale of the Threat Is Unquantified

The interview does not provide loss totals, incident rates or a comparison over time, so it cannot establish how much AI has increased B2B fraud or which sectors face the greatest exposure. Elliott’s statements describe her assessment of the threat, rather than independently verified measurements presented in the report.

It is also unclear how effective shared network-based verification would be, what it would cost businesses, or how systems should balance fraud screening with payment speed and privacy. The source does not identify a specific product, implementation timetable or formal policy change.

Controls Focus on Payment Exceptions

The practical approach outlined by Elliott is to use automated checks for routine payments and route changes, unusual payment amounts and sudden activity spikes to human reviewers. Businesses considering that model would need to decide which signals to combine and what level of risk should trigger a hold or additional confirmation.

PYMNTS said its full interview with Elliott contains further discussion of verification, shared payment infrastructure and human review. The source material does not announce a new program or a scheduled next milestone, so whether companies adopt these measures—and how well they perform—remains to be seen.

Key Questions

What did Katie Elliott say AI is changing about B2B fraud?

Elliott said AI and other available tools let fraudsters make attempts “bigger, broader, faster”, including through mass phishing and spamming. The interview does not quantify the resulting change in fraud losses or incident rates.

Why does payment speed affect fraud prevention?

Elliott said an authorized payment can move quickly and funds may be withdrawn soon after receipt. That leaves businesses less time to act after a fraudulent payment and puts more emphasis on checking instructions before authorization.

What information should businesses use to verify a supplier?

Elliott recommended using multiple signals, including digital identity, phone information, email-domain history and payment details, rather than relying on one piece of data. She did not specify a single required verification standard.

Does Elliott recommend replacing human review with AI?

No. Her suggested model uses automation for normal activity while directing anomalies—such as changed payment details or an unusual request amount—to human approval.

Are shared payment-network checks proven to reduce fraud?

The interview presents shared verification as a possible way to spread the cost of data and tools across more transactions. It does not provide evidence that a particular network service has reduced fraud or details on its cost and availability.

Source: rss