Fraud’s ‘Detection Gap’: What are Fintechs Doing About It?

Financial fraud is impacting the financial services sector at an alarming rate.

In particular, recent data from Experian indicates that identity driven attacks accounted for 71% of all confirmed fraud cases in 2025. 

As the conversation grows more around what fintechs can do to prevent financial fraud, focus has turned to how AI, the popular technology aiding in the efficiency of financial services, can provide assistance or hindrance to assess and deflect upcoming threats. 

Paul Weathersby, Chief Product Officer, Identity, Fraud & Financial Crime Compliance at Experian, explains how to best understand the current opportunities for financial services to prevent fraud.

What are AI-powered fraudsters? How do they affect financial institutions? 

I’ve been working in the industry for more than 25 years and throughout that time have seen the fraud environment shift and evolve hugely – most recently through newer technologies such as AI.

Data and technology are central to everything we do at Experian and understanding new threat vectors and articulating the best course of action for our clients is core to my everyday responsibilities. 

In recent years criminals have started to use AI to launch attacks at a scale and speed that traditional fraud systems simply cannot match.

The increasing accessibility of AI means that well-coordinated and orchestrated attacks are now occurring regularly.

As a result, all businesses are having to step up their tactics to ensure they are protecting themselves and their customers.

In financial services, we’re seeing this drive an increase in the volume of fraud attacks, which makes it harder to manage and increases operational costs.

Another challenge is that this AI-fuelled shift has created a ‘detection gap’, where attacks are evading traditional security. 

When organisations try to tighten their systems in response, they often see a rise in false positives, creating friction for genuine customers and driving up costs further as teams are forced to manually review harmless transactions. 

Are financial services more prone to a certain kind of fraud? If so, why is this?

Fraud is always closely linked to the environment and the asset being targeted, with customer funds being the ultimate goal.

Organisations that hold customer money, such as financial services (FS) companies, are therefore an obvious target for fraudsters because successful attacks result in a direct financial gain. 

There are indirect fraud schemes including premium-rate phone scams, gift voucher fraud, advertising fraud and SMS scams that FS companies are more prone to seeing too because the path to cash is immediate.

However, FS firms have become increasingly resilient, with controls such as multi‑factor authentication now standard, making direct account compromise more difficult.

As a result, two key fraud trends have emerged: Authorised Push Payment (APP) fraud, where criminals manipulate account holders into willingly transferring money themselves and SIM‑swap fraud, which has increased by more than 1,000% and is used to intercept one‑time passcodes sent via SMS, enabling attackers to circumvent authentication controls.

What can financial services companies do to assist in the innovation race against fraud?

Data sharing is critical. Sharing fraud outcomes with other organisations so that it becomes an industry vs fraudster battle – as opposed to just an individual company versus fraudster – enables businesses to approach the fight on much firmer footing.

Businesses must also adopt a multi-layered approach to defence. A mix of behavioural and biometric technology, AI solutions, as well as traditional prevention systems are all crucial in the fight against fraud.

Businesses must be sure of a customer’s identity, something which sounds simple enough but is increasingly difficult as attacks become more sophisticated.

Selfie scans and document verification are important and continual monitoring of applications is now essential. Behavioural analytics helps to confirm the user is a person, not a bot. 

For instance, is the phone’s data consistent with how a person would naturally hold a device, as opposed to it lying face down on a table?

The fundamental challenge is ensuring it’s a genuine person performing a genuine action.

Combining robust identity verification with continuous, intelligent monitoring is the key to staying ahead.

Are fintechs using AI against fraud enough? 

Using AI tools to fight against AI threats will be on the radar for most fintechs and is becoming increasingly common as the landscape constantly changes.  

An area many struggle with is the difficulty of joining up siloed data pools which can prevent the company from getting a comprehensive, holistic view of customer behaviour, which means potentially criminal activity can go unnoticed. 

The application of AI really is opening-up new avenues in fraud prevention, avenues which would have been unthinkable just a few years ago.

As such, we’re always looking at ways to help businesses overcome these challenges. 

For example, Experian and Resistant AI recently launched a new solution, Transaction Forensics.

Combining advanced behavioural and transaction analytics and Experian’s leading proprietary consumer and commercial data assets, Transaction Forensics uses more than 80 AI models to deliver an all-encompassing, granular view of fraud risk across bank-to-bank payments.

What should financial institutions be thinking about when using AI in compliance and onboarding?   

All financial companies must ensure they have a robust AI data governance framework in place in order to give the systems – and the outcomes it generates – the best chance of success.

Confidence in your proprietary data, the data used to feed AI systems is fundamental.

Recent research from Experian illustrates the issue. A recent report notes that 89% of UK businesses said AI had a positive impact on performance but turning AI from principle into practice remains a major challenge. 

Three quarters (76%) of business leaders cite this as one of their biggest hurdles, driven by gaps in technical expertise and tensions between speed, innovation and governance.

Although 90% agree that high-quality data is essential, only 43% are confident their data is fit for purpose and just 48% believe their teams are adequately prepared.

The findings highlight the need for stronger data governance, practical training and clear guidance.

Is there a collective element to fighting fraud in the fintech space? 

A collaborative, collective approach has been a long-standing way of identifying fraud in financial services and sharing data is hugely important.

A company working in isolation is unable to see the whole picture and there are data sharing consortiums which allows businesses to share information, including Experian’s Hunter service.

If you had one piece of advice to CEOs of financial institutions for using AI in the fight against fraud, what would it be? 

Start with one narrow, high-impact fraud use case and prove it works before scaling.

The biggest mistake is trying to transform everything at once.

In fraud, AI is most effective when it is introduced in small, manageable steps that are easy to understand, monitor and measure. 

That allows you to test performance properly, optimise quickly and build confidence in both the model and the operating process around it.

Once you can clearly see the value and control the outcomes, expand incrementally into adjacent use cases. In this space, disciplined progress beats ambitious overreach every time.

It’s also important to remember that technology alone will never be enough.

It requires building a team that is not only equipped with the right tools but is also empowered and expected to switch up its approach as the threat landscape evolves.

In a sector where you are constantly dealing with highly sensitive information, keeping this in mind will help embed a more holistic approach to AI fraud prevention.