Property insurance underwriting
An underwriting assistant for property submissions: broker emails and attachments extracted into a property record with every field sourced to the document, underwriting standards checked with external data and flagged, an AI scorecard against a chosen set of criteria with the PDF beside it, questions asked of the document, and a suggested response to the broker.
One working session with the people in the seat: which decisions, which systems, what a good week looks like.
The app you see above, running on your own exports instead of the simulated world. Every screen, every decision, your numbers.
Wired to your systems with your permissions, agents proposing, people approving, the ledger recording who did what.
Overview
A broker's submission is a long email with attachments; the underwriter re-keys the property, checks for an existing policy, looks up the flood zone and writes back asking for what is missing.
The assistant extracts the property record from the submission with each field traceable to the source, runs the standards checks against internal records, filings and news, and flags what fails. A scorecard scores the risk against a chosen criteria set with a rationale per criterion; the underwriter can ask the document questions. The suggested response to the broker asks for exactly what the flags need and is sent from the inbox.
Key features
- Data extraction from submissions and records with sourcing to the underlying documents.
- Automatic detection and flags with external data such as filings and news.
- AI scorecard: scoring criteria sets applied by an LLM with rationale, viewed beside the PDF.
- Suggested responses to the broker from the flags and human-in-the-loop interventions.
Decisions it produces
- Flag or clear a standard on a submission
- Score a submission with a criteria set
- Send the suggested response to the broker
- Open the property in the ontology
Data sources
| Submissions | emails and attachments from brokers |
|---|---|
| Policy administration | existing policies by address |
| External data | flood zones, SEC filings, news |
| Underwriting standards | checks and thresholds |
Simulated data throughout. Book a call to see it on your own systems.
Related demos
Claims coverage engine
A coverage engine for private liability claims: claims selected for checks, the checks running against coverage conditions written in natural language, results per condition with the reasoning and the identified customer, and the conditions themselves added, modified and tested on a claim, with the guardrail that no payout is refused without a human confirming.
Portfolio risk trends and management
Portfolio steering for new business: gross premium, technical result, capital usage and loss against plan by policy segment and by corporate versus personal, the capital mix and new versus renewed business, a line of business view with rate and exposure trends, breach alerts on the plan, and macro trends from news, reports and meeting notes that trigger portfolio actions.
Real-time customer intelligence
The customer's record, the live call transcribed as it happens, facts and intent extracted from it, a recommended product with its justification, the interaction history and a callback booked, all on the relationship manager's screen during the call.