// Questions
Answers, before the call.
Everything a first conversation usually spends thirty minutes on. If your question is not here, it is worth an email — and it will probably end up on this page.
// The firm
What does Dearborn Labs do?
Dearborn Labs is an AI-native software development firm built specifically for insurance. We rebuild the workflows insurance runs on, and we build the AI systems that run them. You hire us; you don't buy software from us. What we build, you own.
Do you work outside insurance?
No. Insurance only. That is the whole credibility axis and we are not diluting it.
Are you a product company or a services firm?
A services firm. You hire us to build something, and when it is running your team owns it. The build fee is fixed before work starts, and continuing to run and improve what we built is a separate agreement and your call.
Is partnership just a nicer word for a contract?
It is a description of how the work runs, not a softening of the commercial relationship. You hire Dearborn Labs, and what we build is yours. Partner here describes how the engineering runs, not a shared-risk arrangement of the kind your reinsurance panel would mean.
How big is the team?
Small and senior, deliberately. The economics of a delivery pyramid require juniors to bill, and that is the model we exist as an alternative to. It also means we run a handful of engagements at once, which is why what we build sits on your infrastructure from the first commit rather than ours.
What do you mean when you say partner?
Four specific things, and they are checkable before you sign. The people who sold the work build it. Your engineers are on the team rather than on the other side of a handover. You own the asset from the first commit. And we are measured on your team running it without us.
Why hire Dearborn Labs instead of a large integrator?
Because the person in the room is the person writing the code, and because insurance is the only industry we work in. What is checkable rather than argued sits on the diligence page: surfaces, ownership, governance, transfer, security and cost.
How senior are the people who show up?
The engineers on your account are named in the proposal, with the systems they've built. You meet them before you sign, and there is no delivery team behind them.
// The work
Do you replace Guidewire or Duck Creek?
No. Dearborn Labs works at the workflow level, on top of whatever core systems you run, through sanctioned integration surfaces — Guidewire Cloud API, Application Events, Integration Gateway and Cloud Data Access; Duck Creek's gateway; ACORD AL3. A build needing an unsanctioned surface is declared in the proposal, not at implementation.
We already have an accelerator program with our core vendor. Does that rule you out?
No, and it usually helps. We build the workflow on top of what the accelerator delivers, and part of Discovery is telling you which parts of your list are arriving in your vendor's release notes anyway.
What if our historical data is on a mainframe?
If the data behind the workflow isn't fit to use, fixing it is part of the build, scoped and priced with it. We don't ask you to fund a data program first. Any proposal that skips that step is scoped on the new system only, or it is wrong.
Has this run in production, or only in pilots?
In production, in a regulated carrier, for years. Our founders built and operated Clearcover, with AI running live regulated insurance operations since 2017. One of those builds is published in full, with its definitions and its failure modes.
What won't you build?
Automated adverse decisions, ultimate-severity models for commercial auto, materiality conclusions on manuscript-form diffs, dashboards over unfixed ingest, and anything without a named human decision-maker. The full list, with what we build instead, is its own page.
Do you do strategy work?
Only the four weeks of it that produce a buildable spec. We are not a firm you hire for a roadmap you then take somewhere else to implement — though if that is what happens, you still own the map.
// Engagement and commercials
How does an engagement start?
Two doors. If you already know which workflow is broken, bring us the scope and we build it. If you want it mapped first, Discovery is four weeks, fixed scope. Either way, week one is instrumentation. We don't publish build prices because no two builds are the same, and Discovery is quoted before you start.
What does Discovery actually return?
An instrumented baseline of the target workflow, an integration-surface assessment naming which builds are sanctioned and which are not safe to build, a governance-gap read against your states, and a ranked build list including what we would not build and why.
Are we committed to a build after Discovery?
No. Discovery excludes any build, any production change and any commitment to a next engagement. If the finding is that the money belongs somewhere else, that is in the report.
Who owns the code?
You do. Code, data model, tests, runbook and documentation live in your repository, on your infrastructure. We keep the tools and methods we arrived with, the way any firm does, and the contract says so.
How is our data handled?
Role-based access and audit logging are on everything we build, data stays in your tenancy, and the subprocessor list is named. The data-handling terms themselves are in the master agreement, and we send it before you ask.
Can we see your security documentation?
Yes, and you should not have to ask. Subprocessors, data residency and the governance questionnaire are all on the diligence page.
Can you give us references?
We can put you in touch with people who have worked with this team. What we cannot do yet is point at a review-site profile or an analyst placement — we have neither, and analyst relations is a 2027 project.
// Proof
Do you have an ROI number?
We can show you what these systems did inside a carrier we ran. We cannot show you what they will do inside yours, and we will not model it before we have measured it. Before we build anything, we instrument your baseline.
Why won't you just estimate it?
Because an estimate built on somebody else's book is a number you would then have to defend to your own board. The industry is full of vendor-published return figures, and not one of them was measured on your book. We hand you a measurement you can stand behind instead.
What happens if it doesn't work?
You find out in week one rather than month six, because the baseline comes before the build. Discovery is designed to be the cheap place to discover that, and the ranked build list includes the things we would not build.
Still have a question?
Ask it. If it takes a paragraph to answer, it probably belongs on this page.