Tag Archive for: claim payout protection

A driver is rear-ended at 11 o’clock at night. The bumper is cracked, nobody is hurt, and the policyholder uploads photos before going to bed. A week later the settlement is due.

In that week, the claims portal has to answer four questions:

• Is the person logged in the policyholder?
• Is the payout account still theirs?
• Is the claim genuine?
• Can the repair shop see only what it needs?

Each question maps to a different control and a different failure. This post takes them in turn.

Payout changes need their own check

The settlement is paid into a bank account, which makes the payout account the most valuable field in the portal. A login that is taken over is dangerous mainly because it can change this field.

Consider a policyholder who receives a message that looks like it came from the insurer, asking them to confirm their details. They enter their login on a copy of the portal and update the payout account before they realize the page was fake.

A step-up check before the change is the control that stops this. It asks for a fresh factor even though the session is already signed in. In general practice, the portal also sends a notice to the contact details already on file, not only to the new ones, so the real policyholder hears about the change.

OmniDefend’s customer identity offering describes multi-factor authentication for customer transactions, online or on-premise. Its transaction verification can be called from the insurer’s own applications. The platform does not replace the claims system’s payout rules, which stay with the insurer.

A stolen policy number should not open an account

A policy number is printed on an insurance card and appears on documents that pass to repair shops and rental desks. Treating it as a secret is a mistake.

In general practice, enrollment confirms that the person registering also controls something the policy record already holds. That might be a one-time code sent to the contact on file.

• Policy number alone: enough to start a registration.
• Policy number plus a code sent to the contact on file: the registration proceeds only for the person who controls that contact.

Consider a fraudster who has a policy number and a date of birth from a stolen document. Under the first rule, the registration goes through. Under the second, it stops at the code, because the fraudster does not control the contact on file.

Matching an applicant against a large base of enrolled customers is a different job from checking one account at login. OmniDefend’s large-scale identification describes 1:N matching of biometrics across a database, which is relevant to that second job.

Repair shops see only their own job

A portal with several kinds of users needs boundaries. Each party should see only what its job requires.

• Policyholder: their own claims and payout status.
• Repair shop: the assigned claim, its photos, and its estimate.
• Adjuster: the queue they work from.
• Medical provider: the medical part of one claim, not the full policy.

A body shop account set up for one job can stay active for years after the job ends. Partner access should be scoped to one claim, time-limited where possible, and removed when the job closes. Review the list each quarter, because access tends to outlive the work that justified it.

Credential reuse and session takeover after a good login

A login check runs at one moment. After that, the session is what an attacker uses.

Credential stuffing tries passwords leaked from other sites against an insurer’s login. It works because people reuse passwords. Breached-password screening at sign-up and a second factor reduce that risk.

Take a policyholder who reused a password from a retailer that was breached. The reused password works at the portal, but the second factor stops the attacker, unless the code is captured by a fake page.

Staff accounts are the high-value target

Adjusters and claims managers can see every file, which makes their accounts a target. A single sign-on with multi-factor authentication on the identity provider means one strong sign-in can protect several applications.

The risk is the connections that sit outside the identity provider. In practice, each connected application has to be checked, because some keep their own user lists. A claims tool with a local list can keep a former adjuster’s access alive after the identity provider has removed them.

• Identity provider: where sign-in is enforced.
• Connected applications: each one needs its access tied to the same identity.
• Local user lists: the first place to look when a removed person still has access.

OmniDefend describes workforce features including single sign-on, multi-factor authentication, a universal directory, and full user lifecycle auditing. Its financial industry page describes automating joiner-mover-leaver steps with an HR system.

The regulatory baseline

Insurance regulation in the United States is handled mainly by states. The NAIC Insurance Data Security Model Law, dated 2017, asks licensees to maintain a written information security program based on their own risk assessment. States adopt their own versions, and the NAIC publishes a state-by-state adoption page.

An insurer licensed in several states may find that one state’s version asks for more than another’s. The practical approach is to check the rules where the insurer is licensed and design for the strictest one.

Where the identity layer stops

OmniDefend’s listed capabilities cover identity: enrollment, login factors, transaction verification, large-scale identification, user management APIs, and workforce access.

They do not cover:

• Claim photo forensics
• Damage assessment
• The payout workflow
• Underwriting data

Those stay with the claims platform and the insurer’s own controls. Consider a photo edited before upload. That is a claims-evidence question, and an identity check cannot answer it.

Test one claim from first report to payout

The most useful test is one real claim, followed through every step. Check each step against the question it answers.

1. Enrollment: Does a registration without the contact code fail?
2. Login: Does a login from a new device trigger a fresh factor?
3. Payout change: Does a change from an unknown device get blocked, and does an alert reach the contact on file?
4. Repair shop access: Does the shop see only its assigned claim?
5. Staff review: Does a removed adjuster lose access in every connected application?

Each step should leave a record that shows who did what. Run the test on your own claims flow first.

OmniDefend’s sales team sets up the 30-day free trial, so the month goes to real claims rather than configuration. Start the free trial. If you want to see the identity features before that, request a demo.

Frequently asked questions

Can someone register my account with only my policy number?

They should not be able to. A well-designed enrollment confirms that the person also controls something on the policy record, such as a code sent to the contact on file.

Why does the portal ask for a code when I change my bank details?

Changing where a settlement is paid is a high-risk action. The check stops someone who has taken over your login from redirecting the money. The alert to your existing contact details lets you spot a change you did not make.

Can a repair shop see my full claim history?

It should not. Partner access is normally limited to the claim it was assigned, with the photos and estimates it needs, and it should end when the job is complete.

Is a passkey safer than a password for the portal?

A passkey is bound to the site it was created for, so a lookalike page cannot use it. It does not replace checks on payout changes, which still need their own step.

Do insurers have to use multi-factor authentication?

Many do, under state insurance data security rules based on the NAIC model or under their own state regulations. The requirements differ by state, so check what applies where the insurer is licensed.

Resources

• NAIC, Insurance Data Security Model Law, government affairs brief
• NAIC, Insurance Data Security Model Law (Model 668) state adoption, Spring 2024
• Ampcus Cyber, What is credential stuffing (NIST SP 800-63B-4 password screening, secondary)
• ThreatDown, What is credential stuffing?