Senior insurance fraud professionals give a candid assessment of the threats facing the sector and where the gaps in detection and intelligence sharing still lie. Jamie Melling reflects on the insurance fraud panel discussion at the Smartnumbers Consortium Conference in June.
Moderated by insurance fraud consultant Matt Gilham and featuring Anna Phelps, Fraud Risk Manager at Marshmallow, and Andy Pickard, Policy Fraud Manager at Hastings Direct, the conversation covered evolving threats, the role of the contact centre and the case for deeper data sharing across sectors.
Fraud is an evolving threat
Both panellists confirmed that fraud, including ghost-broking and policy applications using stolen identities, remains a significant and persistent threat for insurers.
Andy Pickard noted that the threat is adapting as different controls are put in place. Investment in fraud prevention can redirect fraudsters rather than stop them entirely, he said. Those blocked at the digital front end will shift to the contact centre, where controls are less mature and friction is lower, to find ways to get through.
It is a useful reminder that fraud prevention should not be customer channel specific. Because strengthening one access route without monitoring the others creates an opportunity that organised fraudsters will find.
Synthetic identities are more than just a banking problem
We heard why synthetic identity fraud, long associated with the banking sector, is now a material concern for insurers. The mechanism is straightforward. Insurance, particularly products such as travel or home cover that involve less stringent KYC (know your customer) checks, offers a relatively accessible way for fraudsters to build up the profile of a fictitious identity. A synthetic customer who takes out a cheap policy, pays on time and behaves well begins to accumulate a credible financial history. That identity can then be used to pass checks at banks and other financial institutions.
Anna Phelps explained, “Insurance might be a stepping stone to build legitimacy for synthetic identities.” The insurer may never experience a direct financial loss. But they have contributed, unknowingly, to the creation of a fraudulent persona that causes harm elsewhere in the financial system. It is a collective responsibility issue that argues strongly for cross-sector intelligence sharing.
The contact centre remains a significant blind spot
Andy Pickard raised the issue of withheld number calls in the contact centre, a tactic fraudsters often use when calling the contact centre.
Fraudsters calling the contact centre with a withheld number can set up multiple policies with different quote data, safe in the knowledge that nothing connects those calls to the same person. Without the ability to identify the number behind each call, what is effectively one fraudster setting up multiple policies appears to be entirely separate, unrelated activity.
As Andy put it, the same number could be calling fifteen times a day, but without being able to identify it, those interactions remain invisible to investigators. “That’s the real blind spot for us [we need to know] who’s behind that [call], and what else does that link to?”
Andy noted that the response a call from a suspicious number doesn’t have to be an outright block because that simply pushes fraudsters to change numbers and try again. The smarter approach is to route these calls to a team of skilled agents equipped to handle them appropriately, while giving no indication to the caller that they have been identified. This can be a more sophisticated and sustainable response than blocking alone.
Anna Phelps added a further dimension from Marshmallow’s experience: a fraud case where the same voice was used across ten separate FNOL calls, with no other connecting data points. No device intelligence, no address match, no banking signals. Each call, viewed in isolation, appeared entirely legitimate.
Taken together, these examples make the same point. Just as device fingerprinting has become a standard tool for linking customer and claimant data in digital channels, voice and caller number can link cases in the contact centre.
Data sharing is insurance’s strength but there is more to do
The insurance sector has a genuine track record in fraud intelligence sharing. The Insurance Fraud Bureau (IFB) reaches its twentieth year in 2026, and the Insurance Fraud Investigators Group (IFIG) has been operating for over 25 years. Andy and Anna both acknowledged the collaborative culture that exists between insurance fraud teams and the recognition that there is no competitive advantage in allowing fraud to continue.
But both were clear that more is needed, particularly across sectors. Anna identified a structural gap that many in the industry will recognise. When a claims incident raises concerns about money laundering, the typical route to the banking sector is via CIFAS and an ad hoc request that may or may not yield useful intelligence. “So much fraud permeates across different sectors,” she said, “but how are we linking all that data together?” For smaller insurers in particular, the intelligence banks hold can be the missing piece of a much larger puzzle — significant to them, even if it appears minor from the outside.
Andy Pickard was equally direct about where the current data-sharing relationships fall short. CIFAS, he noted, used to allow more open dialogue, but that has been scaled back. What insurance fraud investigators really want to know is where money is coming from and where it is going.
A bank account number appearing across multiple fraudulent policies may be the only lead available but without knowing who owns that account, making a referral to the Insurance Fraud Enforcement Department becomes significantly harder. Account ownership, addresses, unusual payment patterns and other similar data, shared systematically rather than on an ad hoc basis, could make a material difference to enforcement outcomes.
Anna Phelps summarised the broader argument plainly: “As much [intelligence] as you give, you take way more.” It is a principle the insurance sector understands well within its own ranks. Extending it fully across sectors, and acting on it, remains the work still to be done.
