Data
Versioned, queryable canonical records and engineered variables.
Build AI, evaluate risk and find growth opportunities with governed data and calibrated models. The Inference Institute turns fragmented public, licensed and customer data into versioned products, matching services and licensed exports.
A data product is more than an endpoint. It is a governed, versioned body of records, evidence and models—built for a defined decision and delivered with its provenance, quality and limitations intact.
Versioned, queryable canonical records and engineered variables.
Match customer records to canonical entities and assets, with confidence scores and visible evidence.
Structured, textual and geospatial retrieval across each data product.
Calibrated scoring and prediction endpoints with model versions, uncertainty and limitations.
Licensed snapshots in Parquet, GeoParquet, CSV, JSON or SQLite.
Seismic Asset Intelligence combines earthquake hazard, asset exposure, building characteristics, observed damage evidence and calibrated models to support portfolio risk, post-event assessment and operational prioritisation.
Featured product · SAI-01
Asset-level earthquake exposure, event impact and damage intelligence for insurers, corporates, governments and investors.
Portfolio enrichment, accumulation analysis, post-event impact and claims triage.
View use case → 02Site resilience, critical-facility prioritisation, operational disruption and supplier exposure.
View use case → 03Scenario planning, affected-area assessment and inspection prioritisation.
View use case → 04Physical-risk screening, portfolio concentration and asset-level exposure.
View use case →Trust is release furniture, not marketing copy. Every product states where its evidence came from, what changed, how quality was measured and where a result should not be used.
Read our standardsThe brand line
Reading signal from risk.
The decisions that matter most are made on noisy, incomplete evidence. Our work narrows that uncertainty without hiding the assumptions, missingness or limits.
Papers, benchmarks and open samples demonstrate the methods behind commercial products. Open research remains open; commercial products carry separate terms.
We work with design partners to build and validate data products before moving them into recurring API and dataset delivery.
Discuss a data product →