HomeSpark Sentinel

HomeSpark Sentinel (2026 Insurtech Award Winner) serves the full policy lifecycle rather than a single point of insertion. At underwriting, it supplies interior condition and equipment data no exterior model can see, improving risk selection and pricing precision. For policyholder engagement, carriers deploy a white-label/co-brand direct-to-consumer experience that gives homeowners a genuinely useful product and the carrier a persistent relationship outside of renewal season. At claims, a pre-loss digital twin transforms first notice of loss and self-inspection: contents documentation already exists, timestamped and spatially anchored, compressing cycle time and reducing disputes. At renewal, ongoing scan and sensor data makes retention an informed decision rather than actuarial guesswork.

Sentinel is the first carrier-grade risk intelligence platform built from inside the home. Every other model carriers buy stops at the front door - satellites, aerial imagery, parcel records, roof age. Losses don't. They happen at the water heater, the supply line, the panel, the appliance.

It is built by HomeSpark, Inc., a data and infrastructure company creating the labeled dataset and vision model for the interior of the home. The company's consumer iOS app, HomeSpark.AI, lets homeowners capture their property in minutes using iPhone Pro LiDAR - producing a 3D digital twin with every system, appliance, and finish identified and spatially pinned, maintained as a living record over time. That capture engine, HomeScan, produces the proprietary dataset Sentinel sells against.

The underlying framework is SAGE: Scan, Assemble, Guide, Evaluate. Scan captures the space. Assemble builds the structured twin and identifies its contents through a vision language model pipeline. Guide is the homeowner-facing layer - AI assistant, maintenance scheduling, contractor management. Evaluate is where carriers plug in, converting interior condition, equipment inventory, and change over time into risk signal.

A Matter and IoT integration layer extends the twin from a static capture into a live model. A leak sensor that reports water is useful. A leak sensor that reports water beneath a fourteen-year-old water heater on a finished second floor is underwriting-grade.

The carrier value proposition is risk intelligence, customer quality, and margin predictability - not simply claim avoidance, though the loss-prevention economics are real. Matteo Carbone of the IoT Insurance Observatory has estimated water-related interventions at roughly 21 combined ratio points and fire at roughly 15.

Security is architectural, not bolted on: AES-256-GCM encryption with an envelope key structure, per-home key scoping, and optional zero-knowledge BYOK.

HomeSpark won the InsurTech Award at Insurance Innovators USA in Nashville, May 2026. The goal is larger than any single carrier deal: the world's best labeled residential interior dataset, and the vision model trained on it.