Every email read. Every document filed. Nothing sent without you.
Briar is an AI operating system for independent life-insurance agents. It watches your inbox, reads carrier documents, files them against the right client, tracks every application through underwriting, and drafts your replies — then stops and waits for your approval.
Briar is an AI operating system for independent life-insurance agents. It connects to the agent's inbox and reads every message, classifies it, and drafts replies the agent approves. It extracts policy facts from carrier PDFs, files documents to the correct client, tracks applications from submitted to delivered, and surfaces opportunities a busy advisor would miss — most notably living-benefit riders clients already hold. Briar is carrier-agnostic and prepares work rather than sending it: nothing reaches a client without the agent's approval.
Overview
What is Briar?
Briar is an AI assistant for independent life-insurance agents. It connects to your inbox and reads every message, files arriving documents against the right client, extracts policy facts from carrier PDFs, tracks applications from submitted to delivered, and drafts your replies — which it holds until you approve them.
What it is
An AI operating system — and the system of record — for a life-insurance practice.
Who it's for
Independent life-insurance agents and small agencies in the United States.
What it does
Reads the inbox, extracts facts from carrier documents, files them to the correct client, tracks applications through underwriting, and drafts replies.
What it won't do
Send to a client on its own, or state coverage, eligibility or pricing. Those are licensed representations.
Carriers
Carrier-agnostic. Not built around one company's products or rules.
Price
$500–$800 per agent, per month.
Built for the actual workflow
Declarations page
Illustration
Policy receipt
Conversion deadline
Accelerated benefit rider
Term expiry
Cash value
Paramed exam
Attending physician statement
Beneficiary change
1035 exchange
Underwriting requirement
Chronic illness rider
Face amount
In-force illustration
Carrier amendment
Delivery receipt
Rating class
How it works
One message, start to finish.
Six stages run on every email that arrives. Briar works through them beside this text as you scroll — the same six it runs at 3am on a Tuesday.
01It arrives
How does Briar see a client's email?
Briar watches the inbox every fifteen minutes — inbox and spam both, because carrier notices land in spam and matter anyway. Each message is classified by free rules first: sender, labels, whether the address is already on the book. A paid model call happens only when the answer is genuinely ambiguous.
Junk is filed without spending anything. A known client's message is never buried by a promotional signal.
Proven on a real mailbox: 5,000 messages, 0 errors, 19 minutes.
02It reads the document
How does Briar read a carrier PDF?
An attachment is opened inside a sandbox that can return typed fields and nothing else — policy number, carrier, coverage, cash value, premium, conversion deadline, riders. There is no free-text field in the output, so an instruction hidden in a PDF has nowhere to live.
Eighteen document kinds are distinguished, because a policy receipt is not a policy and the agent does something different about each.
Living-benefit riders found on a real illustration, unprompted, at 0.95 confidence.
03It knows whose it is
How does Briar decide which client a document belongs to?
Filing happens only when two independent identifiers agree — the sender resolving to one client, confirmed by the document naming that same person, or a carrier policy number already on the book. A name alone files nothing. A name is what an attacker writes, and fourteen clients on one book can share one.
When signals disagree, Briar consults the relationship graph and asks a one-tap question — “this looks like your client's spouse” — instead of an open one.
Every automatic filing is logged with the evidence behind it.
04You ask
What happens when an agent asks Briar a question?
Ask in the app, or by text. Briar does not answer from memory — it queries its own records, walks the graph from the client to their policies, documents, and relationships, and collects the evidence as it goes.
Broad and shallow by default; narrow and deep only when a real question demands it.
Reads its own retrieved records, never the model's recollection.
05Every figure carries its source
How does Briar prove an answer is true?
Each number in the answer is checked against the record it came from before you ever see it. A coverage amount cites the document it was read from. A claim that cannot be traced is not shown — Briar says it does not have that instead.
The verification runs after the model and before the render, so an ungrounded answer never becomes a file, a letter, or a slide.
Failed verification shows the reasons, in words, rather than a confident guess.
06And then it stops
Does Briar send the reply itself?
The draft is finished, in your voice, with the right facts — and it goes no further. Briar prepares; you release. This is the one rule the entire system is built around, and it holds even when the convenient thing would be to skip it.
Ask by text and Briar will answer anything and prepare anything. It will finish nothing that leaves the building.
Nothing reaches your client without you.
Capabilities
A full assistant, not a feature.
Briar holds the book, works the inbox, reads the documents and prepares the output. Anything marked Roadmap is not built yet — said plainly here, because a product that overstates itself on its own website will overstate itself to your clients.
The inbox
Briar watches it so you do not have to.
Polls every 15 minutes — inbox and spam both
Classifies on free rules first; pays for a model only when genuinely unsure
Drafts a reply only where one is warranted
Rescues a known client's mail from the spam folder
Applications from submitted through underwriting to delivered
What every carrier is still waiting for, and what nobody has chased
A stale-case chase list, oldest first
Relationships — spouses, beneficiaries, referrals
Bulk import of an existing book from a CSV export
Judgement
What a great advisor would notice and never has time to act on.
Living-benefit riders clients already hold and were never told about
Conversion windows before they close
Renewal milestones at 5yr, 2yr, 1yr, 6mo and 30 days
Silence — a client nobody has spoken to in months
Beneficiary problems: a minor with no trust, an ex-spouse still listed
Every finding quotes the data point that triggered it
Ask it to do the work
Answers that carry their sources, or no answer at all.
Ask anything about a client and get it cited to the record
Draft a letter in your voice, with your sign-off
Compare a new illustration against current coverage
Build a proposal deck, a Word letter, a spreadsheet, a PDF
Every result states whether Briar could verify it
Nothing renders at all if it cannot be grounded
However you like
You should never have to open a dashboard to ask a question.
The web app — mail, clients, drafts, insights, applications
Text it and it answers — summaries, status, who someone is
Telegram, verified cryptographically
Notifications that say what happened, never what was in it
Phone — call Briar, and Briar calls carriersRoadmap
Appointments and schedulingRoadmap
Why Briar
Not AI-flavoured software.
01
It writes the way you write
Every time you edit a draft before approving it, Briar records the correction and updates its model of your voice. The drafts stop needing edits. And you can open what Briar concluded about how you write and change it — a profile that shapes every message and cannot be inspected is one you either trust blindly or distrust entirely.
What you type outranks what Briar inferred. Permanently.
02
It notices what you never have time to
Term policies from several carriers include terminal, chronic and critical illness accelerated-benefit riders at no extra cost. Most clients holding them have never been told. Briar reads the policy documents it already has and tells you which of your clients are sitting on benefits nobody has mentioned — the single highest-value conversation in a life book, and the one nobody has time to go looking for.
Found on a real carrier illustration, unprompted, at 0.95 confidence.
03
It is one workflow, not five tools
The insight, the draft, the document, the application, the send — one system. Briar is the system of record itself, so you are not paying for it alongside a CRM and reconciling two books that disagree. An existing book imports from a CSV export, and the inbox fills the rest in by itself.
Briar holds the book. It does not sit on top of somebody else's.
04
It is built to survive a regulator
Agencies and uplines are liable for their vendors, so Briar is built to be examined. Client data is isolated by the database itself rather than by application code. Every automatic decision is logged with the evidence behind it. Briar never states coverage, eligibility or pricing to a client, because that is a licensed representation and an E&O exposure — a wrong figure is professional liability whether a human or a model said it.
“Why did that end up there?” is always answerable.
5,000
messages backfilled in one run
0 errors · 19 minutes
47
tables with tenant isolation
verified automatically
25
typed fields per document
across 18 document kinds
0.95
confidence on living-benefit riders
on a real carrier PDF
The promises
A wrong answer that looks right is worse than none.
That sentence decided more of Briar's design than any other. It is why Briar refuses rather than guesses, why every figure carries a source, and why it is built to fail loudly instead of quietly.
Nothing reaches your client without you.
Briar prepares; you release. An inbound email can never cause an outbound one.
Briar never claims anything.
It relays what it holds and shows you where it came from. It will say “I don't have that” rather than guess.
Your data is yours.
No other agent can see it. Isolation is enforced by PostgreSQL itself, so a bug in the application cannot leak across agents.
Briar tells you when it cannot see.
A dead inbox connection is announced, not hidden behind a healthy-looking dashboard.
Briar asks rather than guesses.
It files by itself only when two independent identifiers agree — never on a name alone.
Briar can retrace its own footsteps.
Every decision it made on its own judgement is logged with the evidence behind it.
Pricing
Independent agents and small agencies.
Briar is carrier-agnostic. It is not built around one company's products or rules, so an agent writing with several carriers sees one book rather than one silo per company.
$500–$800
per agent, per month
Against a category that runs roughly $50 to $300. That premium is only justified if Briar genuinely replaces the administrative work rather than adding another tool to manage — which is the standard it is built and measured against, and the standard you should hold it to.
Straight answers, including the ones with a “no” in them.
What is Briar?
Briar is an AI operating system for independent life-insurance agents. It reads the agent's inbox, files documents against the right client, extracts policy facts from carrier PDFs, tracks applications through underwriting, and drafts replies. The agent approves anything that leaves the system — Briar prepares work, it does not send it.
What does Briar do for an insurance agent day to day?
It watches the inbox every 15 minutes, classifies each message, and drafts a reply only where one is warranted. It files arriving documents to the correct client, keeps applications moving from submitted to delivered, flags what each carrier is still waiting for, and surfaces opportunities in the existing book.
Does Briar send emails to my clients automatically?
No. An inbound email can never cause an outbound one. Briar prepares a draft and stops; the agent reviews and releases it. The single exception is an optional document-received receipt, off by default, using wording the agent wrote in advance — never text a model generated.
Is Briar a CRM, or does it sit on top of one?
Briar is the system of record itself. It holds contacts, policies, documents, applications in flight, commissions, tasks, and history — so an agent does not pay for Briar alongside a CRM. It can import an existing book from a CSV export, and it builds the book automatically from the inbox.
Which insurance carriers does Briar support?
Briar is carrier-agnostic. It is not built around one company's products, forms, or rules, and it reads documents from any carrier — declarations pages, illustrations, policy receipts, and carrier letters. Agents who write with several carriers see one book rather than one silo per company.
How much does Briar cost?
Briar is aimed at $500 to $800 per agent per month, against a competitor band of roughly $50 to $300. The premium is justified only if Briar genuinely replaces administrative work rather than adding another tool to manage, which is the standard it is built and measured against.
How does Briar keep one agent's client data separate from another's?
Isolation is enforced by PostgreSQL row-level security, not by application code. Every tenant table carries a policy the database applies to every query, so a bug in the application cannot leak across agents. Coverage is verified automatically across 47 tables on every check.
Can Briar read a declarations page or an illustration?
Yes. Briar extracts policy number, carrier, coverage amount, cash value, premium, conversion deadline, riders, beneficiary, and issue age as typed fields. On real carrier PDFs it identified terminal, chronic, and critical illness accelerated-benefit riders unprompted, at 0.95 confidence.
What are living benefits and why does Briar look for them?
Living benefits are accelerated-benefit riders — terminal, chronic, and critical illness — included on many term policies at no extra cost. Most clients were never told they have them. Briar reads the policy documents it already holds and flags clients sitting on benefits nobody has mentioned.
How does Briar avoid filing a document under the wrong client?
It files automatically only when two independent identifiers agree — the sender's address resolving to one client, confirmed by the document naming that person, or a carrier policy number already on the book. Never on a name alone. Anything ambiguous becomes a one-tap question instead of a guess.
Can Briar make up a policy number or a coverage amount?
It is built so that it cannot. Every figure Briar states must trace to a record it retrieved, and the check runs before anything renders. If a claim cannot be grounded, Briar says it does not have that rather than guessing. A wrong answer that looks right is treated as worse than a missing one.
How does Briar learn to write like me?
Every time the agent edits a draft before approving it, Briar records the correction and updates its model of that agent's voice. Over time the drafts need fewer edits. The agent can also open the profile Briar built, read what it concluded, and correct it directly — what the agent types outranks what Briar inferred.
Is Briar safe to point at a mailbox full of client information?
Untrusted content — email bodies, attachments, web pages — is treated as data and never as instructions, and document extraction runs sandboxed with typed output only, so an instruction hidden inside a PDF has nowhere to live. Documents are encrypted at rest and every automated decision is logged with its evidence.
Does Briar replace the licensed agent?
No, and it is deliberately built not to. Briar never states coverage, eligibility, or pricing to a client — that is a licensed representation and an E&O exposure. It handles administration so the agent spends their time on the work only a licensed human may do.
Early access
Briar does the admin. You do the part only a licensed human can.
Briar is onboarding a small number of independent agents. If you write life and you are spending your week on administration instead of clients, get in touch and tell us what your week actually looks like.