What we build
An underwriting stack that holds up at submission speed — and in a regulator's review.
Each capability is a production component — not a proof-of-concept — wired into your policy core, documented for your actuarial committee, and monitored continuously.
Building + asset data ingest
IFC BIM models, GIS coordinates, IoT sensor streams, and SOV schedules parsed into a single exposure schema — geocoded, validated, and ready for modelling.
Peril modelling — flood, wind, quake, wildfire, subsidence
Location-level AAL and PML computed against RMS / AIR / Verisk methodology. Multi-peril combined with correlation-aware aggregation, not naive summation.
Cat accumulation at portfolio level
Graph-neural aggregation across the book — PML-at-tail, regional concentrations, reinsurance retention analysis, and scenario stress by event set.
Loss-ratio + combined-ratio forecasting
Segment-level LR/CR forecasts calibrated to your book — by line, geography, and broker. Continuous rate-adequacy monitoring against the evolving benchmark.
Reinsurance treaty optimisation
Quota share, surplus, XOL, and facultative logic applied per submission. Treaty-terms engine surfaces the right cede before an underwriter needs to ask.
Underwriting-guideline AI — queryable in plain English
Your guideline library made retrievable by natural-language query. Broker-facing quote APIs return in seconds — with the guideline path logged on every bind.
Underwriting motions we power
One pricing engine, every line.
Same exposure ingest, peril models, loss-ratio forecaster, and treaty engine — tuned per line. Shared rating layer, per-line model weights, portfolio-level accumulation. Every motion below runs on the same stack; only the models and integrations change.
Commercial property — single-risk to portfolio
Per-location peril pricing, multi-location accumulation, and portfolio PML reporting. Built for specialty insurers and MGA facilities.
Industrial / engineering lines
Process-industry exposure, BI interdependence, and construction-period cover — with engineering-report OCR and site-imagery classification.
Marine — cargo, hull, specialty
Voyage-risk scoring, port exposure, conflict-zone overlays, and general-average apportionment — with trans-Asia and Suez-sensitive routing.
Commercial motor + fleet
Fleet-level telematics ingest, route-risk scoring, loss-ratio forecasting, and driver-pool vetting. Priced per vehicle and per route, not a flat rate.
Broker-facility quote APIs
Sub-second quote-to-broker APIs for MGAs and delegated-authority programmes, with rate-adequacy and binder-authority guardrails enforced in-flight.
Reinsurance treaty desks
Cedent-facing portals, treaty-terms automation, ceded-premium tracking, and renewal-cycle analytics — for reinsurance teams serving multiple cedents.
Model families we deploy
No single model prices every risk. So we stack them.
Each model covers a distinct part of the underwriting motion — perils, accumulation, loss ratio, rate adequacy — and the outputs combine into one auditable quote.
Physics-grounded + ML peril models at location level. Returns AAL, PML, and event-set impact for each insured property, calibrated to your historical book.
Portfolio-level accumulation with correlation-aware aggregation — catches regional concentrations and treaty-exhaustion risk that location-level sum misses.
Gradient-boosted LR/CR forecasts by line, geography, broker, and occupancy class — updated continuously and stress-tested against economic regimes.
Watches rate vs benchmark vs loss trend — alerts when a segment drifts below adequacy floor. Feeds renewal triage and pricing-review cadence.
Data sources wired into every submission
Every input that moves a quote — integrated.
Pulled in parallel, normalised into a single exposure schema, versioned alongside the models that consume them.
Explainability, not just predictions
A quote alone doesn't pass compliance. A trail does.
Every quote, bind, or decline is accompanied by peril-model outputs, loss-ratio forecast, treaty-clauses fired, rate build-up, and rate-adequacy score — generated at decision time, indexed for audit, and available for regulator and broker review.
- SHAP-style feature contributions per quote
- Peril-model outputs + treaty-terms lineage logged
- Broker-facing rate justification (multi-language)
- Aligned to MAS RBC, Solvency II, NAIC ORSA, IAIS
Frameworks we align to
Why Axccelerate for underwriting
Not a rating engine bolted onto a spreadsheet.
A pricing stack.
A pure rating tool gives you one number. Our stack gives you peril modelling, cat accumulation, loss-ratio forecast, treaty optimisation — the infrastructure a real underwriting team needs to price and cede at the same time.
Pricing
Priced to the perils and the portfolio, not the submission volume.
Underwriting deployments are custom — we scope against your perils, treaties, and integrations before quoting.
Glossary
The vocabulary behind every quote.
A quick reference for the terms that show up in underwriting operations — the vocabulary your actuarial team, regulator, and audit documentation will all use.
- Peril
- Peril
A specific cause of loss covered by a policy — fire, flood, windstorm, earthquake. Priced individually and aggregated across the book.
- Cat model
- Catastrophe model
A physics + statistics model of rare, severe events — hurricanes, earthquakes, floods. RMS, AIR, and Verisk are the industry-standard providers.
- AAL
- Average Annual Loss
The expected loss per year over the long run, from a given peril on a given risk. The core input to technical pricing.
- PML
- Probable Maximum Loss
The loss expected from a severe but plausible event — usually at a specified return period (e.g. 1-in-250 years). Drives capital + reinsurance sizing.
- Exposure
- Exposure
The value, location, and characteristics of what's insured. Property TIV, fleet vehicle count, shipment cargo value — all forms of exposure.
- Accumulation
- Accumulation
The sum of exposures that could be hit by the same event — regional concentrations, single-occupancy-code clusters, supply-chain interdependence.
- Treaty / facultative
- Treaty vs facultative
Treaty = automatic reinsurance cover for a class of business. Facultative = one-off reinsurance on a specific risk. Most books use both.
- Loss ratio
- Loss ratio
Claims incurred as a fraction of earned premium. The core measure of underwriting performance — 60-70% is typical for well-run commercial books.
- Combined ratio
- Combined ratio
Loss ratio + expense ratio. Below 100% = underwriting profit, above = loss. Most healthy books target 90-95%.
- Rate adequacy
- Rate adequacy
Whether the current rate is sufficient to cover expected loss plus expenses plus profit margin. Measured against benchmark and loss-trend.
- Binder
- Binder
A temporary or permanent agreement granting a broker or MGA the authority to bind insurance on behalf of the carrier, subject to defined limits.
- Quote-to-bind
- Quote-to-bind ratio
The fraction of quotes issued that actually convert to bound business. A pricing-health KPI, sensitive to rate adequacy and competitor moves.
- Reinsurance
- Reinsurance
Insurance for insurers. Cedes part of the risk (and premium) to another carrier in exchange for reduced capital requirements and smoother results.
- Ceded / assumed
- Ceded vs assumed
Ceded = the portion passed to a reinsurer. Assumed = the portion you've taken on from another carrier. Net = retained after cessions.
Your underwriting desk, engineered.
30-minute scoping with a senior engineer and an underwriting-systems lead. You'll leave with a pricing plan, treaty-engine sketch, and realistic timeline — not a sales pitch.