Your Cyber Insurance Policy Says One Thing. Can Your Business Prove It?

A practical look at how comparing your own cyber insurance application, policy, procedures, and records helps your business find mismatches before a loss.

The scenario below is invented. No client, no matter, no real company.

Imagine a 180-person logistics software business. On a Tuesday the controller gets an email that looks like it came from a vendor the company has paid for four years, asking that this month's payment go to a new account. The thread carries the right invoice number and the right contract date. She routes it to a second person for approval, exactly the way the written procedure says to. That person compares the wire request against the paperwork already sitting in the file, sees that the two match, and releases $412,000.

Three weeks later the vendor calls about a missing payment. The company files a claim on its cyber insurance, the policy meant to pay for losses from things like hacking, ransomware, data exposure, and fraudulent payment instructions. The CFO reads that policy closely for the first time. He finds a sentence that makes coverage turn on whether the authenticity of the transfer request was verified in line with the company's own internal procedures.

So he goes looking for the procedure. Then he goes looking for proof that somebody followed it that Tuesday.

There are two separate problems in that story, and they have different fixes and different owners. The first is whether the control did any real work. Comparing an email against a filed copy of that same email may never have confirmed who sent it. The second is whether the company can show what it did at all. A business can fail on either one, or on both.

The four documents

Here is the part that catches people. The answer to "are we covered" is not in one place. It is spread across four documents, written by four different people at four different times.

  1. What you told the insurer. The answers on your application or renewal form, as of the day somebody signed it.
  2. What you actually bought. The policy: the pages that promise to pay, the conditions attached to that promise, and the list of things it will not cover.
  3. What your own written procedures say your employees do. The wire procedure, the access rules, the patching standard, the incident plan.
  4. What your records show actually happened. Logs, tickets, call notes, approvals, screenshots, and the account of the people who were in it.

In the story above, the company has all four of those. Nobody has ever read them in the same sitting. That is the entire exercise. Not whether the documents exist. Whether they say the same thing about the same control on the same date.

The form you fill out, and the clause that points back at it

Before a policy exists, somebody at your company fills out an insurance application and signs it. Those answers can matter later, because a policy may point back at them when a claim comes in. So it helps to see what one of these forms actually asks, and what a contract does with the answers.

One published example is the cyber insurance application published by Beazley, form F00863, 042023 edition, written for applicants with revenues under $250 million. It tells the applicant to make the answers accurate as of the date the application is signed. The security questions sit as line items inside a longer form. Two of them ask whether a second login step beyond a password is required for remote access to the network, and whether it is required for web-based email. The form asks whether you require that step, not how well it works.

Now the contract side, from a different insurer in a different year: the specimen policy published by Travelers, form CYB-16001, edition 01-19. It includes a condition saying the insurer issued the coverage relying on the accuracy and completeness of the statements the insured made to it.

That clause is narrower than it first sounds. It does not say a wrong answer ends the policy. In that form, an untrue statement takes away coverage for a loss only if the untruth mattered to the insurer's decision to take on the risk, and only if it also matters to the loss being claimed. There is a third limit on top of those two. The person who knew the statement was untrue when the policy started has to be the individual seeking coverage, or, for a company, one of the senior officers the form describes: the chief executive officer, the chief financial officer, the chief information security officer, the risk manager, or in-house general counsel, plus anyone whose role is the functional equivalent of one of those.

Keep one thing straight about both documents. They are two dated forms from two different insurers, and neither one is your policy, a current market standard, or a general rule. Your own contract may read differently, or carry no comparable condition at all.

So come back to the Tuesday wire. If that company had answered a question like those two, and the answer stopped being true months before the loss, which sentence in its own policy would decide whether that mattered? That is a question you can go read the answer to today.

What none of this proves

Nothing above creates a rule that a lapse in multi-factor authentication, patching, or backups defeats a cyber claim. Nothing above creates a rule that a weak answer on an application produces automatic denial or unwinds the policy. Coverage turns on the wording you bought, your facts, and the governing law, and an imperfect record is not the same thing as no coverage. The narrower and more useful point is that you can know today which sentence in your own policy would decide your most likely claim.

One case where that kind of clause decided the claim

A Florida settlement agent wired roughly $180,900 of mortgage payoff funds to someone impersonating the lender. Its policy covered that kind of theft by deception only if four requirements were met, and the fourth required that the authenticity of the transfer request be verified in accordance with the insured's own internal procedures.

On September 1, 2021, the trial court ruled for the insurer without a trial on two grounds, and said either one alone was enough. The first was that the lender was not an employee, customer, client, or vendor of the insured, the categories the clause named. The second, at pages 10 and 11, was that the insured made no attempt to verify authenticity. It was undisputed that the company had told the insurer, while the policy was being underwritten, that it verbally confirmed wire instructions, and undisputed that it did not do so in this transaction. The court described the company's two-person review as a comparison of the final instruction to the bank against instructions nobody had ever verified.

The limits on that second ground are the point. It is trial-level reasoning, not a rule approved on appeal. The Eleventh Circuit affirmed on September 6, 2022 on the first ground only, and at page 9 note 3 it expressly declined to decide whether the company had failed to verify under its procedures.

That opinion is unpublished. It applied Florida law, under which the court read the policy by its plain meaning and would have read a genuine ambiguity in favor of coverage.

It sets no rule about multi-factor authentication, patching, backups, or ongoing security promises, and no rule that an operational lapse defeats a claim. It is one illustration, not a prediction about your policy. What it does show is that a clause pointing back at your own internal procedures can be the clause the fight is about.

Who owns the comparison

So here is the diagnostic question worth asking inside your own company: who owns the comparison of all four records? If you can name one person who could put the control, the application answers, the contract, and the records side by side, then it is someone's job and it can be scheduled. If the four sit with different people and nobody is responsible for reading them against each other, the comparison may not belong to anyone yet. That is worth finding out before a loss, not after.

The worksheet

Five entries. Run them on one loss scenario and one policy, not on everything you own.

1. The clause your most likely loss runs through. Name the single loss most plausible for your business: a fraudulent payment instruction, ransomware, an outage, a data exposure. In your own words, summarize the pages that would pay and the conditions attached to them, with the form number and edition date printed on the page. Name the individual, not the department, who can produce that page on request, and the date it was last checked against the current policy. Policies get amended mid-term.

2. Whether your policy ties coverage to what you told the insurer. Find out whether your contract has a condition about the accuracy of your statements, and summarize its actual terms with a page reference. If it measures knowledge at a senior-officer level, list by name the people who fit that clause's own definition. Then record what each of them was shown before the last form went in, and who showed them.

3. The answers you actually gave. Pull the last signed application or renewal form from the broker file. In one column, write the answer given and the signature date. In the next, write the present state of that control, confirmed in writing by the person who runs it. Mark every row where the two columns differ and note when the difference started. Take the marked rows to your broker and counsel now to work out what your own policy requires of you and on what timing. Do not assume a mismatch can wait until renewal, and do not assume anyone has agreed to overlook it.

4. What you owe the insurer after a loss. Write down your policy's own trigger for notifying the insurer and its deadline language, with page references. Name the person who would send that notice at 6 a.m. on a holiday, and the backup. Then walk a practice drill through those specific clauses rather than a generic incident script, and write down both the date and what you could not produce.

5. Any security-maintenance exclusion. Check whether your policy has one, without assuming either way. If it does, write down the standard it points to, who publishes that standard, which version, and who inside the company tracks changes to it. If it does not, record that, with the form edition and the date you checked, so the next comparison starts from a known baseline.

Four questions for the leadership team

  1. Which single policy condition decides our most likely loss?
  2. Who signed our last application, and who has checked those answers since?
  3. Which named person produces the evidence inside our notice deadline?
  4. What has changed in our environment since the policy started that the insurer has not heard?

What this means for each seat at the table

CEOs and founders. You are carrying an asset you have probably only read in summary. Ask for the four-document comparison by name, ask who owns it, and ask when it was last refreshed. If the owner turns out to be nobody, that is the finding.

CFOs and finance leaders. If the renewal form lands on your desk, the security answers on it belong to the people who run those controls. Before it goes back in, have each answer confirmed by the person who runs that control. If the honest answer changed in March and last year's answers get copied forward in August, you have created a document somebody may ask you about later.

CISOs and security leaders. Your controls may be stronger than your record of them. The ask is retrievable proof rather than more control: who approved, when, out of which system, kept for how long, and findable by someone who is not you.

Boards and investors. The useful question is not whether the company has cyber insurance. It is whether anyone has compared the policy conditions, the application answers, the written procedures, and the records, and what turned up the last time somebody did.

Practical takeaways

  1. Pull your own application or renewal form from the broker file and copy the control answers, with the signature date, into one document. You cannot audit an answer nobody has read.
  2. If your policy conditions coverage on the accuracy of your statements, find out which individuals fit that clause's own definition of a senior officer, because some clauses measure knowledge only in those seats and only as of the day coverage started.
  3. Before the next form goes in, have each control answer confirmed in writing by the person who operates that control, and keep the confirmation with the form.
  4. Where a coverage clause points at your own internal procedures, read the procedure and ask what evidence would show somebody followed it on one specific day. If the procedure never says to confirm who a counterparty is, that silence is the finding.
  5. Keep the four records distinct in whatever you write down. An application answer, a policy condition, a written procedure, and a log entry prove four different things, and blending them is how a mismatch stays hidden.

Common questions

What does this comparison actually cover?

Four records about one control: the latest signed application or renewal form, your policy, your written procedure, and your records of what happened. The point is whether they say the same thing.

When should a company run it?

Before a loss, while there is still time to read your own contract and current facts calmly. If you find a mismatch, take it to your broker and counsel rather than deciding on your own what it means.

What is a reasonable first step?

Pull the latest signed application or renewal form from the broker file, and ask the person who runs each control to confirm the answer in writing. Then compare it against what your own policy says.

Closing perspective

Nobody can tell you what an insurer or a court would do with your facts. That is the honest answer, and it is also why the comparison is worth running. What you control is whether your own four records agree before anyone has a reason to test them against each other.

An insurance contract is a promise that gets tested through documents. The only real question is whether you learn which document mattered on your own schedule or on somebody else's.


This article is for informational purposes only and does not constitute legal advice. Every company's situation is different, and you should consult with qualified legal counsel before making compliance decisions based on the developments discussed here.

Share this article and about the author

Meetesh Patel, Esq., founder of Consilium Law LLC

Meetesh Patel

Founder and Managing Attorney

I write SparkPoint myself. I built and sold a law firm, ran a clean energy company as CEO, and spent a decade advising founders before building this practice.

More about the firm
Contact

If this touches the work in front of you, start a conversation.

Send a short note about what changed, what you are building, and where legal judgment needs to sit closer to the work.

Disclaimer. This article is provided for informational purposes only and does not constitute legal advice. Readers should consult independent counsel before acting on any analysis. The views expressed are solely those of the author and do not necessarily reflect the views of Consilium Law LLC.