Book Miner Agent reads your book export and ranks every account that’s missing coverage you’d expect. Your own rules sit on top of the industry library, and your own commission assumption is behind every figure.
With Gil, our coverage-gap agent · about five minutes · stop and come back any time.
He asks the handful of questions that decide whether a review of your book is worth running at all. Prefer a person from the start? Book 20 minutes with Vasu.
The next year of revenue is already sitting in the book. The manufacturing client with no cyber. The contractor whose umbrella never got written.
Nobody has time to read three hundred accounts line by line looking for them.
The analysis itself isn’t hard. A good producer can do one account in ninety seconds.
The job is three hundred repetitions of that ninety seconds — which is exactly why nobody ever reaches the end of it.
Every row is a review candidate for your licensed team, and never a recommendation to the insured.
Every gap, and every draft email, is for a licensed producer to edit, own and send. The agent never contacts a client.
Your commission assumption is a required answer at setup, precisely because a default there would put a made-up figure in front of a producer. Any number you see is built from yours.
The shared coverage library is generic industry knowledge. The library is never derived from any agency’s data, including yours.
One file, read once. No new system for anyone to learn, and nothing for your producers to install.
Client, business type, policies held, premium and renewal date — straight out of your AMS.
Business type mapped to the coverages you’d expect to see, plus your own rules. The “when the business type mentions X, also expect Y” your agency already knows and has never written down.
Account, missing coverage, why it was flagged, and what it’s worth at your commission assumption — not a number we invented.
Anything filtered by your minimum premium, the lines you never write, or a rule you’ve silenced is counted and disclosed. A shorter list has to explain itself.
The list gives them review candidates to call on. Book Miner Agent contacts nobody.
Worth checking now rather than on the call. Gil asks about all of this anyway.
Client, business type, policies held, premium, renewal date.
The rules map industry to expected coverage. A book with no industry column has nothing to map, and this is the one that stops projects.
A list nobody calls isn’t worth generating.
So any figure is built from your numbers rather than a default we chose.
We’d rather say that here than have you wonder.
We don’t take credit for the sale. Your producer made it, and the report says so.
Someone sits with a producer and works an agreed sample before anything is installed. That count is the baseline, and it stays yours whether or not you continue.
The action rate matters most. A row nobody calls is noise, and too much noise means the rules need tuning — we’d rather know than not.
We never see your policy system, so the closed number is yours to give us or keep. We report it as your figure, never as ours.
We put the before and after in front of you and you decide. If the numbers don’t justify continuing, we’ll say so first.
We won’t fill in the second slider. How many of your accounts are missing something is the exact question a review answers, and guessing it for you would put a made-up figure in front of a producer.
Every row is a review candidate, not a sale. Each one still has to be checked by a licensed producer, called and written. Plenty won’t convert, won’t need the cover, or already have it somewhere else. The figure above is what the list represents at your own commission assumption. That number is not revenue we are promising you.
A synthetic agency, built specifically to try to break it.
A 300-client book with 648 coverage gaps planted on purpose, plus the answer key — so we know exactly what the analysis caught and what it missed. Those numbers are our test set, not a claim about your book. We have no client results, because we have no clients yet: a founding client sees it run on their own accounts, under NDA, before deciding anything.
Gil is our coverage-gap AI agent. He knows what Book Miner Agent needs, and he knows commercial lines. He asks the questions we’d ask on a call: what your AMS exports, whether the rows carry an industry, and who would work the list.
If it isn’t a fit, he says so. And if the real problem turns out to be renewal accuracy rather than unsold coverage, he passes you to Reed rather than selling you this.
What he can do: answer questions about Book Miner Agent, work out whether your book fits, arrange the NDA, and book time with Vasu.
What he can’t do: quote you a price, give coverage advice, or promise a result. Same rules as the product.
The output is a review list for a licensed person, and the agent sends nothing. Nothing leaves your agency unless a producer writes it and sends it.
No. The list tells your producer what to look at. The agent never contacts anyone.
A generic industry rules library, plus your own rules layered on top. Your rules stay yours, and never go into the shared library.
NDA first, then your data is isolated to your agency, encrypted in transit and at rest, never used to train anything, and never folded back into the shared rules library.
Either way the report tells you what it filtered and why, so you can adjust the minimum premium and the lines you write rather than wondering whether it worked.
Every install is scoped to your business, so the exact number comes out of a 20-minute call.
Gil will look at what your AMS can export, show you the synthetic book, and tell you honestly whether a review is worth running. Prefer a person? Take twenty minutes with Vasu.
If the pain is really about what the carrier issued versus what it quoted, that’s a different agent. Gil will send you there himself rather than sell you this one.