CYBERINSURANCE · IN DEVELOPMENT
Your cyber policy, read against AI risk.
A tool that reads your own cyber and tech E&O policies, runs AI-related loss scenarios against the wording, and shows which clauses each answer turns on, with the page reference so you can check it yourself. It never gives a verdict. It gives you the questions to put to your broker.
WHAT YOU GET
One finding, as the tool produces it
This sample comes from a synthetic policy written for testing. No real insurer’s wording appears on this page.
Our own coding agent misconfigures or deletes production
Loss type Business interruption
Agreement Synthetic Agreement A
Interpretation unresolved
The tool found the clause this answer turns on and does not decide what it means.
What the policy shows
A synthetic declarations amendment identifies Agreement A for review.
AGREEMENT A REMAINS SELECTED SUBJECT TO THE TABLE BELOW.
ABC-ZZZform page 1 of 2PDF page 4
What is unresolved
The Automated Instruction definition may require broker interpretation for a software-produced command.
Automated Instruction means a command produced by software without a contemporaneous human keystroke.
ABC-ZZZform page 1 of 2PDF page 4
What to do next
Ask the broker how the Automated Instruction wording applies to the scenario.
Automated Instruction means a command produced by software without a contemporaneous human keystroke.
ABC-ZZZform page 1 of 2PDF page 4
What this turns on
The clauses
selection
AGREEMENT A REMAINS SELECTED SUBJECT TO THE TABLE BELOW.
ABC-ZZZform page 1 of 2PDF page 4
automated-instruction-definition
Automated Instruction means a command produced by software without a contemporaneous human keystroke.
ABC-ZZZform page 1 of 2PDF page 4
The reasoning
- The amendment identifies the agreement before the definition is considered.
- The scenario turns on whether the software-produced command meets the cited definition.
What would change the answer
Different wording that required a contemporaneous human keystroke would change the unresolved question.
WHERE YOUR POLICY GOES
It stays on your machine.
The tool is built to read your policy from your own computer and write its results there. It makes no network call unless you turn one on: an export, or an optional AI step that sends definition text to your own AI provider, under your own key.
Your insurance policy is a confidential document. Nothing about this tool should make that harder to keep.
WHAT IT DOESN’T DO
It reads. It doesn’t rule.
- It never gives a verdict on a loss. Every finding is a question with its evidence attached.
- It doesn’t say what insurers usually do. Every finding comes from your policy’s own wording.
- It isn’t legal advice, and it doesn’t replace your broker or coverage counsel.
STATUS
In development.
Before anyone sees output from a real policy, an outside reader checks that every finding follows from the clauses it cites. We built that check in because the people who wrote the tool share one reading of the wording, and a mistake we all make would pass all of us.
