Trust, then re-verify: Securing the customer lifecycle in AI-driven insurance

As insurers automate more of the customer lifecycle, a verification check completed at onboarding may not be enough. Re-verifying identity and account details at onboarding, when important customer or account details change, and before claims payouts can help insurers identify inconsistencies before funds are released.

• Trust established at onboarding should not be treated as permanent. As AI and automation reduce human intervention across insurance workflows, insurers should re-verify identity and account details at critical points throughout the customer lifecycle.
• Three moments deserve particular attention: onboarding, servicing and account changes, and claims payouts. Each creates a different opportunity for synthetic identity fraud, account takeover or manipulated payment details to go undetected.
• Re-verification can help insurers add security without sacrificing automation. Rechecking identity and account information at key moments can surface inconsistencies before changes are accepted or funds are released.

Meta description: Artificial intelligence is a powerful tool for insurers – and fraudsters. Discover three points in the customer lifecycle where the risks are greatest.

Automation now runs much of the customer lifecycle. Artificial intelligence (AI) pulls data from submissions, triages cases and drafts routine documentation. Instant payment rails move the money. Fewer people review each step.

As insurers automate more of the customer lifecycle, a verification check completed at onboarding may not be enough. Re-verifying identity and account details at onboarding, when important customer or account details change, and before claims payouts can help insurers identify inconsistencies before funds are released.

Account takeovers jumped 297.3% from 2023 to 2024. Fraudsters requested US$228 million from policyholder accounts in 2025, according to LIMRA.[note1] Nearly three in four of those attempts came through the customer portal.

Fraudsters access the portal in one of two ways: passing the initial verification check or infiltrating a genuine account. In both cases, they rely on the trust established at onboarding to move through insurance workflows.

Insurance runs on a chain of trust. An insurer establishes trust at onboarding and then relies on it throughout servicing, claims and payouts. Each automated stage gives fraudsters a different way to exploit that trust. Here’s where to look.

For most insurers, verification happens once: document checks and know-your-customer (KYC) review when the policy is written.

Synthetic identities combine real and fabricated details to pass traditional checks. They present new risks at onboarding, especially in automated workflows. In a manual process, an underwriter reads the file and may notice the inconsistency. An automated system can miss the flaw, letting the fraudster move into servicing and claims unchecked.

Nearly a quarter of identity-theft claims referred to the National Insurance Crime Bureau (NICB) involved a synthetic identity.[note2] They account for up to 85% of identity fraud in life insurance, according to the Business Information Industry Association (BIIA).[note3]

After onboarding verification, a new address, beneficiary or bank account looks like a routine change. Fraudsters who pass that initial check then change the account details. Those who fail it take over a genuine customer’s account instead.

Automation leaves fewer people to catch them.

The scale is already visible. Fraudsters made 9,567 account takeover attempts in 2025, according to LIMRA. When they came through the customer portal, they accessed the account 79% of the time. The average disbursement requested was US$88,332, and insurers took an average of 18.9 days to notice.[note4]

Fraudsters pass the onboarding check, change the account details and file the claim. Now comes their costliest move for insurers: collecting the payout.

Even if insurers catch them, instant rails leave less time to recover the money. In one US emergency relief programme, tornado victims received instant payments at midnight on a Saturday.

A fraudster’s payout moves just as fast. And the sums can be large. The average account takeover request across channels was US$110,213 in 2025.1

Some requests get paid. A fraudster with a couple’s personal details passed the insurer’s contact-centre verification and filed a death claim on the wife.4 The insurer paid it by Automated Clearing House (ACH) transfer. The fraudster then filed a claim on the husband. While it was processing, he opened a benefits letter and called the insurer. Only then did an audit find the red flags on the fake death certificate.