From fulfilment to finance: How AI is reshaping everyday fraud for businesses

Fraud isn’t standing still

As everyday digital tools become more powerful, they are also making some of the oldest scams faster and harder to spot.

Organisations are now seeing familiar tricks resurface in new forms – particularly in two areas that hit day-to-day operations:

1. the diversion of high-value goods during fulfilment, and
2. the submission of fabricated receipts in expense claims.

These are not new fraud types. What has changed is how accessible, fast and credible the tools have become – backed by data showing broad uptake of AI-enabled fraud techniques.

The numbers are startling

• In September 2025, information security business AppZen reported that around 14% of all fraudulent expense claims were generated using AI-created fake receipts.
• UK Finance has reported that in the first six months of 2025, UK fraud losses reached £629.3m, a year-on-year increase in volume of 17%, across more than 2 million confirmed cases.
• In Q1 2025, 35% of UK businesses reported being targeted by AI-related fraud.
• A finance industry survey found that 74% of UK finance-leadership respondents believed employees could exploit AI tools to manipulate expense processes.

Both patterns carry operational loss, reputational risk and legal exposure. Businesses are responding: in the next year, over two thirds plan to increase fraud prevention budgets, and just over half expect to strengthen their AI-based analytics to detect and deter manipulation.

1. Identity fraud involving high-value goods

Reports from mobile operators and electronics retailers show growing instances of devices ordered using compromised or fabricated identities. AI tools are accelerating this trend by making it faster and cheaper to create convincing identity documents, deepfake images and supporting details that pass standard checks. Fraudsters can now automate bulk applications and generate matching communications at scale, overwhelming review systems.

Cifas reported 26,000 telecom account takeovers in the UK between January and June 2025 – cases where scammers accessed customer accounts to make unauthorised purchases.

Common features include:

• Orders placed just under automated review triggers
• Late changes to delivery instructions or carrier routing
• Rapid resale of goods overseas or on secondary markets

Where instalment or financing vehicles are used, this shifts from mere stock loss into credit fraud, with implications for recoveries, chargebacks, insurer scrutiny and supplier relationships. Even when goods are seized later, the organisation often absorbs the operational cost, chargebacks and reputational impact.

This pattern is increasingly seen in business models with high-value portable products, remote fulfilment and dealer or partner networks – such as luxury retail, specialist equipment supply and B2B subscription hardware.

In these cases the practical leverage lies not after shipment but before release – emphasising the importance of verifying or slowing the transaction if possible.

2. Fabricated receipts in expense claims

In parallel, AI-fuelled expenses fraud is surging.

Finance teams are encountering an uptick in receipts that mimic genuine merchant formatting, complete with correct logos, plausible tax and line-item detail and authentic templates.

Visual checks alone are no longer a reliable safeguard. Because expense systems often rely on trust and speed, they become low-friction channels for otherwise well-disguised misconduct.

Even low-value claims can trigger audit, insurer and governance questions around an organisation’s controls.

Why these patterns are intensifying in larger organisations

Fraud is growing fastest where big businesses are most complex. Six features of large organisations make them especially exposed:

1. Scale and spread

   Large companies sell through many routes – online, in-store, through dealers and logistics partners – each with its own checks. Fraudsters exploit the gaps between these systems, such as when delivery details can be changed or when controls differ by region. Cross-border resale channels make diversion losses harder to trace once goods move overseas.

2. Thresholds and pressure for speed

   Most workflows rely on value limits and delivery or payment deadlines – ship today, approve under £500, reimburse within a week. Offenders learn these patterns and stay just below review thresholds. Exceptions gradually become routine as teams prioritise service levels and speed.

3. Data scattered across systems

   Big businesses often run multiple platforms for orders, payments, and expenses. When those don’t connect, it’s difficult to see that one device, address or account has been flagged elsewhere. Fragmented data weakens visibility and allows repeat offenders to re-enter the system.

4. Convincing fabrications

   AI tools can now generate realistic documents, identities and messages in seconds. Receipts or invoices look authentic enough to pass first review – especially when staff are under time pressure or rely on automated checks. Fraudsters no longer need specialist skills, only accessible tools.

5. Heavy use of third parties

   Couriers, payment processors, and dealer networks hold key information about deliveries, chargebacks, or suspicious accounts. But data quality and response times vary, and contracts don’t always allow quick information-sharing or recall. Many pre-shipment verification steps still depend on automated rules without manual override or ID checks.

6. Cross-border complexity

   Different jurisdictions have different privacy, shipping and consumer-protection rules. That makes it difficult to trace goods, recall stock or share alerts in time to prevent further losses. Recovery becomes slower and more expensive once assets leave the original market.

Closing note

Although emerging technologies are accelerating fraud, all is not lost. Research into how technology enables financial crime is expanding, giving organisations clearer insight into how these risks evolve and how to respond proportionately.

Recent work led by Professor Joe Burton has examined the considerable evidence emerging of a substantial acceleration in AI-enabled crime, and how this understanding can guide prevention and policy responses.

At the same time, AI can also be part of the solution. As Rebecca Rossiter at AlixPartners has highlighted in her work on “AI as your co-detective”, AI tools can strengthen investigations by helping teams find key facts faster, make sense of large volumes of data, and check answers against the original documents.

Both Joe and Rebecca will be speakers at our Compliance Club webinar on 2 December 2025 at 1pm, which focuses on Business Crime in the AI Era.

In the 30-minute session we will explore how legal and compliance teams can navigate these challenges, including:

• The legal landscape – how new laws and enforcement priorities are shaping corporate exposure in the AI era.
• Where the risks are – how emerging technologies can create opportunities for fraud and other misconduct across sectors.
• Turning technology to your advantage – how AI and digital tools can support investigations and strengthen compliance.

Sign up now.

Authored by Reuben Vandercruyssen, Liam Naidoo, Claire Lipworth, and Olga Tocewicz.