How P&C insurers can detect fraud at quote, policy and claim
When people talk about insurance fraud, the conversation usually jumps straight to claims.
But a lot of the damage is done much earlier.
Customers quietly “tune” their answers to shave a bit off the premium. Young drivers buy dodgy policies from social media accounts. Fraudsters build synthetic identities that pass standard checks and sit on your book for months before turning into losses.
By the time those risks land with the claims team, you’re already on the back foot.
This guide walks through:
- What policy and application fraud looks like now
- Why it’s getting harder to spot
- The key signals to watch across the lifecycle
- How KBS software can help you join the dots
1. What we mean by policy & application fraud
Policy and application fraud is any false, incomplete or manipulated information given at quote, application or mid-term change, whether that’s deliberate or “just a little white lie”.
Typical patterns in motor and home include:
- Non-disclosure and misrepresentation
- Understating annual mileage
- Saying the car sleeps on a driveway when it’s on-street
- Leaving out previous claims or convictions
Surveys consistently show that more than a quarter of drivers admit to bending the truth to reduce their premium. [1]
- Fronting and first-party fraud
- Business misstates how an asset is used, e.g., a “back-yard” rental car fleet insured as a standard low-risk vehicle.
- A household “shares” a policy in a way that clearly doesn’t match real usage
- Ghost broking and fake policies
- Fraudsters posing as brokers on social media, selling fake or heavily altered motor policies to price-sensitive drivers, often leaving them completely uninsured. [6]
- Synthetic and third-party identity fraud
- Fraudsters mix genuine credentials (ID numbers, addresses, dates of birth) with fabricated details to create a “new” identity that passes basic checks and then slowly build a history before hitting you with losses. [5]
None of this is brand new, but the volume, the sophistication and the digital reach are all growing.
Insurers are blocking more fraudulent applications at quote and bind, and some report application and identity fraud growing faster than traditional claims fraud. [2][3]
At the same time, first-party fraud (where the applicant themselves is complicit) has become the single biggest category of fraud attacks globally, accounting for around a third of cases in the latest large-scale digital fraud data. [4]
2. Why policy fraud is getting harder to spot
2.1 Digital journeys means less natural friction
More business run through direct digital channels and self-service portals.
That’s good for growth and customer experience, but it also means:
- Less time to spot “off” behaviour
- Fewer conversations where a broker or agent says, “This looks odd, talk me through it”
- More opportunities for fraudsters to test different answers and see what gets them the best price
Fraud teams end up reading the final version of the application, not the journey that got there.
2.2 Identity and synthetic ID fraud are spiking
Identity-driven fraud has jumped to the top of the worry list. Identity and synthetic ID fraud are now among the hardest fraud types to detect, well ahead of many claims typologies. [3]
At the same time, synthetic identity fraud is gaining ground across financial services, with analysts classing it as one of the fastest-growing forms of financial crime. [5]
The challenge is that synthetic customers:
- Often pass KYC and credit checks
- Behave “well” for a while
- Don’t have a real victim who will raise a complaint
So they sail through onboarding and rating, look fine on paper, and only show their hand later.
2.3 Data is still fragmented across the lifecycle
Many insurers still have:
- Quote and application data in one platform
- Policy and billing in another
- Claims, devices, and external checks elsewhere
Fraud surveys repeatedly highlight this lack of a single customer or entity view as one of the biggest blockers to better detection and prevention. [3]
If you can’t easily see the full picture – quote history, prior policies, claims, contact points, devices – risky applications will keep slipping through as apparently “clean” new business.
3. Signals to watch across the lifecycle
Rather than thinking of policy fraud as a one-off event at inception, it helps to look for patterns over time.
3.1 At quote and application
Useful signals here include:
- Risk details that don’t quite hang together
- Very low mileage but a long commute
- High-performance vehicles being priced as if they were low-risk run-arounds
- Suspicious quote behaviour
- Many quotes from the same device or IP with different personal details
- Rapid tweaks of key rating factors (address, occupation, drivers on the policy) to chase a lower price
- Channel and context clues
- Policies arranged via informal channels (e.g. messaging apps, social platforms) instead of regulated brokers and direct channels, a common pattern in ghost broking.
3.2 Mid-term and renewal
Fraud doesn’t stop at day one:
- Frequent mid-term changes that move the risk profile up and down
- Churn between different addresses or keepers that don’t match genuine life events
- Multiple policies linked to the same device, payment method or contact details under different names
These are often early signs of policy farming or identity abuse.
3.3 At claim
By the time a claim appears, you’re often seeing the consequences of policy or identity fraud:
- Accident circumstances that contradict declared use, mileage or overnight parking
- Claimants, third parties or repairers who appear in other suspicious networks
- Identity information at claim that doesn’t line up cleanly with what you saw at application
The goal is to capture enough of these signals at quote, bind and mid-term to avoid only spotting them when there’s money on the table.
4. Practical moves for fraud leader
You don’t have to fix everything at once. A few focused shifts can make a big difference.
4.1 Make policy fraud a shared problem
Application fraud sits at the intersection of fraud, underwriting, pricing and distribution. Treat it that way:
- Agree simple, shared measures such as fraud blocked at quote, confirmed policy-fraud loss, ghost-broker cases identified
- Include policy and identity fraud explicitly in your fraud strategy and risk appetite, not just claims fraud
That way, nobody is “owning” only one slice of the puzzle.
4.2 Build an entity and lifecycle view
Even if you start small, aim for an entity-centric view that links:
- Quote and application history
- Policies and claims
- Devices, email/phone, addresses and payment methods
- Relevant external / contributory data where you’re part of a consortium
Industry work on fraud trends keeps coming back to the same conclusion: that broader, better-linked data and collaboration are some of the biggest levers you have. [3]
4.3 Treat synthetic ID as its own beast
Synthetic identity won’t be solved by “more KYC” alone.
- Combine identity data with behavioural and device signals – how and from where the customer interacts – and look for patterns over time, not just at onboarding [5][7]
- Share synthetic ID patterns and cases across fraud, credit control and claims so you recognise the same synthetic persona wherever it turns up
The aim is to spot the “fabricated but consistent” identities that fall between the cracks of traditional controls.
5. How kbs software can help you
kbs software is built to help fraud and underwriting teams tackle policy and application fraud across the lifecycle, not just once a claim arrives.
In practice, that means you can:
- Connect the dots across data
Bring quote, policy, claim and external data into a single view so you can spot entities, links and networks, including policy farming, ghost broking patterns and repeat devices that span multiple identities. - Score applications in real time
Use advanced analytics to highlight high-risk applications at quote and bind, with clear reasons that underwriters, fraud analysts and brokers can understand and act on, not just a black-box score. - Run smarter workflows with proper governance
Automatically route risky applications for review, keep a clean audit trail of what was checked and why, and support the governance expectations regulators are setting around fraud, identity and AI-assisted decisions. [7]
The point isn’t to turn every quote into an investigation. It’s to quietly filter out bad and synthetic risks so your teams can focus on genuine customers and genuinely complex cases.
References
- [1] Which? “5 car insurance lies that could invalidate your policy”, 14 May 2024.
- [2] Aviva 2024 fraud results releases (claims and broker updates, April–May 2025).
- [3] Insurance Post – Insurance Fraud Survey 2024 and related articles.
- [4] LexisNexis Risk Solutions Cybercrime Report 2024 (“The Calm Before the Storm?”)
- [5] LexisNexis “Understanding synthetic ID fraud in the UK insurance industry”
- [6] LexisNexis “Tackling insurance fraud: rising challenges and effective strategies” and related fraud-trend content.
