Clossara reads the email threads behind every placement, quote and renewal and tells your account managers which ones are slipping, why, whether they will bind, and what to send next, with the message behind every finding. Your own private model. No third-party AI. No training on policyholder data.
Commercial lines, corporate and specialty, employee benefits, broker placement teams and MGA distribution.
Every release is judged on 176 held-out sales threads it has never seen, finding by finding, against the full conversation. Before anything else on this page, this is how often the current release is right:
of the findings an account manager sees were judged true against the thread. What the model cannot support, it hides.
of citations point at a real message. The reviewer can open the message behind every claim.
invented numbers in the gated release. A premium, limit or date not found in the thread or your documents is flagged for review.
November 2026 release. The four-page evaluation report is sent with every briefing.
Clossara reads each placement's full email history against your book and writes a short diagnosis. Every line points at the message it came from. What the model cannot support, it does not show.
Why is this renewal slipping? The root cause and the signals behind it: a pricing question left hanging, documents the underwriter is still waiting for, a single contact, a client who went quiet, each cited to the message.
Will it bind, and how confident is that, from the conversation rather than the renewal stage. Your renewal book, ranked by what the threads say instead of by the date on the notice. A thread with too little signal says so.
A reply that answers what the client or underwriter actually asked, in your house tone. Any premium, limit, deductible or date that is not in the thread or your documents is flagged before it becomes a promise. The account manager reviews and sends from their own mailbox.
One renewal, as the email thread recorded it. The CRM said "in renewal" the whole time.
Renewal notice sent, premium up 18% on the loss history. The system moves the policy to "in renewal" and nobody looks at it again until the date.
The finance director asks what drove the increase and whether anything can be done. Message 6 of 22. It is a buying question, not a complaint.
"Looking into it, back to you shortly." No figures follow. The underwriter's request for updated loss runs is sitting in a colleague's inbox, unowned.
The client stops writing. Another broker answered the premium question in a meeting. On day 3 the email says "we've placed it elsewhere." Clossara flags the quiet on day 52, reason cited, 49 days to act.
The fourth reason renewals are lost is the AI tool compliance said no to: every conversation-intelligence product the team evaluated sent policyholder correspondence, claims and health details included, to a shared cloud model. Clossara does not. See the requirements.
These are the questions data-protection, outsourcing and IT-security reviewers at insurers and brokers have asked us. The answers are what is in place, not a roadmap.
| Your requirement | How Clossara meets it |
|---|---|
| Policyholder data stays in our jurisdiction | The Portal and a dedicated instance are hosted in the EU. For a book that may not leave the building at all, the Clossara Box runs the whole product on your own network. |
| No client correspondence to a third-party AI | Clossara's own model, no third-party LLM at runtime. In the cloud it runs on servers Clossara operates and sees redacted text only; on the box it runs on your hardware. |
| Special-category data in threads | Claims and benefits threads carry health and personal details. Names, addresses and identifiers are replaced by labels before any text reaches the model, and the model stores nothing. What is read and kept is listed on the security page. |
| No use of our data for training | Never. Not in the cloud, not on the box. The model is trained on Clossara's own development data and ships as a gated release; it reads your threads, it does not learn from them. |
| Can we check what it says? | Every finding cites the message it came from; a finding without a citation is not shown. Each release is measured on a fixed held-out set and the report is available to your reviewers. |
| Least privilege on mailboxes | Read-only mailbox permissions (Gmail gmail.readonly, Microsoft 365 Mail.Read). Clossara cannot send, delete or modify mail. Drafts are copied and sent by a person. |
| Access control and audit | Enterprise adds audit logging and an SLA; SSO and SCIM are coming soon. Every stored analysis is traceable to the thread and the release that produced it. |
| Outsourcing and DORA questionnaires | We complete your security questionnaire as part of the review and provide the DPA and subprocessor list. On the box there is no processing outside your organisation to register. |
| Independent certification | in progress. We say so rather than imply it. Ask for the current status. |
Details, data flows and retention are on the security page. Nothing here is specific to insurance regulation; it is what the product does.
For a broker team or a line of business that wants to start this quarter. Hosted in the EU, redacted before the model, per seat and per month. Import closed renewals as .eml files and check the model against what happened before anyone connects a mailbox.
For a carrier or a large broker with its own review. Enterprise gives you an isolated application and database in the EU with audit logging and an SLA. The Box puts the whole product, model included, on your own network with no outbound connection.
It reads the mailboxes you connect. Most teams connect the account-management and placement mailboxes, not claims. If a claims thread is in a connected mailbox, it is redacted like every other thread before the model reads it, and the model stores nothing. You decide which mailboxes are in scope.
No. Clossara's model is trained on Clossara's own development data and shipped as a gated release. It reads your threads; it never trains on them, in the cloud or on the box. That is in the terms, not only on this page.
It understands conversations. A renewal is a deal with a date, a price objection and a counterparty; a placement is a deal with a third party asking for documents. The model diagnoses those the same way it diagnoses any sale: from what was asked, what was answered and who went quiet. It will not quote cover, rate a risk or interpret a wording. It will tell you the underwriter is still waiting for the loss runs.
Email is the primary source and works with any system, because the threads are in the mailbox. CRM connectors cover the common sales CRMs today; for a broker management system without a connector, the pipeline can be loaded from an export, and the thread analysis does not depend on it. Ask us about your system in the briefing.
Yes. Import a set of renewals whose outcome you already know as .eml files from a mailbox export, and compare what Clossara says with what happened. That is the fastest honest pilot, and it needs no mailbox connection.
A small team that builds and trains its own model. We would rather tell you that than pretend otherwise, and it is why the box exists: an insurer should not have to trust a vendor's operations when it can keep the whole thing inside.
Thirty minutes with the people who built it, with your security, outsourcing and data-protection reviewers in the room. Tick the box below and the evaluation report comes with our reply.
Smaller team? Sign up for the Portal. Prefer email? Write to [email protected]