Limits, deductibles, sub-limits and exclusions checked one by one against a catalogue built line of business by line of business. Every conclusion in the report carries the article it was drawn from, quoted in full.
Detected a € 50,000 sub-limit for losses other than business interruption, and exclusion of dismantling and reassembly costs.
The request is always the same: "a quote, a comparison." The real goal is different: never ending up with a coverage gap that costs you dearly. An exclusion you didn't notice, or an overlooked sub-limit, is not a technical detail: when a claim arises it turns into a coverage gap — and it can become a dispute on your shoulders. Generic tools read keywords, but they don't grasp how limits, deductibles and exclusions interact with one another.
"Pure Financial Loss"
A structured process to assist contract review, phase by phase.
The system processes the policy text to identify the key data, including changes made after the contract was issued.
Art. 2.4.6, p. 22 — Table of limits, pp. 8-12.
Comparison tables between the in-force policy, renewal and quotes, to highlight regulatory changes and coverage updates.
Art. 9 of the renewal endorsement, not referenced in the schedule.
Support in claim handling, cross-referencing the notice data with the contract text.
Notification deadline: 3 days from becoming aware of the event.
A line-of-business catalogue is not a keyword list: it is the set of clauses that decide, in that line, whether the claim gets paid — together with the criteria to judge them. It is built one line at a time, which is why there are eighteen and not one.
This is a real report, not a screenshot: scroll it. Names, VAT number, policy number and claims history have been replaced; the structure, the analysis and the quotations are what your client receives.
Real report — Environmental Liability, waste treatment plant. Policyholder, insurer, address, VAT and policy numbers and claims data replaced with fictitious values.
Three things. None of them is "faster".
Exclusions and sub-limits surface before signing. It is the part of the job where a mistake does not show up straight away: it shows up at the claim, when the client asks why you never mentioned it.
Every conclusion in the report cites the article it comes from. If the client challenges it, or their lawyer does, you have the text in front of you instead of a recollection.
A clause-by-clause comparison moves the conversation to where you win it. The competitor brings a quote; you bring the reason that quote costs less.
Maurizio Molinari — twenty-five years in insurance broking, registered on section B of the Italian RUI under no. B000252003. Engineering degree, master’s in Risk Engineering.
For twenty-five years I read policies the way you do: general conditions on the desk, special conditions alongside, endorsements to cross-check by hand. And every time the same feeling at the end, that something might have slipped past me. Not through carelessness: an insurance contract is not written to be read quickly.
Section B means I answer to the client, not to the carrier. It is the same place you are in, and it is why this tool does what it does rather than what would suit an underwriter.
Insurance-Lab does the mechanical part of the job: it opens the three documents together, finds the clause, checks whether an endorsement changed it, and puts the exact text in front of you. What to make of it stays yours — it is the reason the client pays you and not a piece of software. — Maurizio Molinari
Security is not an add-on feature: it is the foundation the entire system is built on
Tools to mask sensitive data. You decide what to anonymize before processing.
Servers in Europe (Oracle Cloud). AI processing via Vertex AI, with transfers governed by EU standard contractual clauses.
The original PDFs are deleted after processing. We do not train models on your data.
Full compliance with EU Regulation 2016/679. DPA available on request.
The AI reads, extracts and compares documents: it doesn't decide. Every report is support material the professional reviews and validates — the last word always stays human, as insurance distribution rules require.
Every analysis follows a privacy-first pipeline: sensitive-data masking before anything reaches the model, processing on European infrastructure, file deletion when the work is done. Control over what gets shared stays with you.
Four monthly credit-based plans, with no annual commitment. A policy analysis costs 2 credits; a coverage gap discovered at claim time has no price list.
Monthly credits, and variable costs. Plan credits renew every month; unused ones roll over to the following month up to a quarter of the monthly allowance. The included credits may be updated based on the costs of AI service providers, with sixty days' notice and the right to withdraw without penalty within thirty days of that notice. If those costs fall, the credit allowance rises.
It is not a mockup, nor a screenshot picked for the occasion: it is the document the pipeline produces, with the structure and level of detail your client receives. The data is invented; everything else is real.
Uploaded PDFs are deleted right after the analysis, and the generated reports are stored encrypted. We do not keep permanent copies of the documents nor use them to train models; data stays on servers in the EU, in compliance with the GDPR. Masking tools are also available to anonymize data before upload.
Yes. LLMs reason on a probabilistic, not deterministic, basis: completely eliminating the margin of error is technically impossible. That is why Insurance Lab is a support tool — it saves hours and flags critical points, but the review and the final decision stay with the professional.
We issue a proper Italian invoice. We accept card payments via Stripe. You can choose monthly or annual billing (two months included).
Yes — and it's a question you should ask anyone offering you AI. In Insurance-Lab the AI reads, extracts and compares documents: it makes no automated decisions about the client. Every output is support material the professional reviews and validates — the human remains the sole decision-maker. On top of that, the privacy pipeline: masking before the model, no training on your data, EU servers.
A generic model reads the text you give it and summarizes. Insurance-Lab works on catalogs built line by line — 363 clauses across 18 insurance lines — and checks the policy item by item: limits, deductibles, exclusions, sub-limits and how they interact. Plus: built-in sensitive-data masking and structured reports designed for professional use, from policy comparison to claim pre-analysis.
It depends on the file: a corporate policy with special conditions takes a morning to read, and the report takes you straight to the points that deserve your attention. The honest answer is to measure it on your own workflow, on the line of business you handle most and on a policy you already know by heart.
Write to us telling us the line of business you handle most: we will reply with the details on the plans, the credits and what the report actually contains for that line. No policy to send, no client data involved.