Property-level damage evidence
Compare pre-event and post-event imagery, score visible damage, attach evidence, and show confidence at the parcel level.
Ray by Rayford AI
Every property, ready and recoverable.
Ray turns local hazard history, post-disaster imagery, and property records into auditable damage evidence for emergency managers, local governments, and recovery teams.
Ray Assess
Hurricane Milton sample
Damage score
0.78
Confidence
High
Action
Inspect
Roof edge deformation, debris pattern, and street-level evidence indicate moderate structural damage.
Interactive demo
Watch remote and street evidence become a review-ready decision.
Remote sensing gives disaster teams coverage at scale. Street-view capture adds ground confirmation. Ray Assess turns both into a damage score, evidence trail, confidence, and inspection guidance.
Upstream
Remote + street capture
Output
Property evidence packet
Users
Emergency management, GIS, recovery teams
Start here
Damage assessment and claims triage.
Rayford AI starts with one urgent workflow: help human teams decide which properties need attention first, and why.
Claims and inspection triage
Rank properties for adjuster review and package imagery, metadata, and explanations for faster claim workflows.
Mitigation and grant support
Connect damage evidence with local hazard history, mitigation priorities, and FEMA-aligned recovery documentation.
Beachhead
Starting with emergency management teams that need property evidence fast.
2026 pilots
Pilot pricing available.
We are opening a small number of early pilots for historical event validation and property-level assessment workflows.
Discuss a pilotEmergency management offices
Prioritize preliminary damage assessment when a disaster creates more properties than field teams can inspect immediately.
Local governments and GIS teams
Convert street-level and aerial data into property-level damage layers for preliminary assessment, recovery planning, and public assistance workflows.
Adjusters and insurance partners
Use the same evidence package for claims triage once the public sector assessment workflow is validated.
Why now
Disaster losses are rising, but property evidence is still slow.
U.S. billion-dollar weather and climate disasters in 2024.
NOAA ↗Estimated U.S. damage from 2024 billion-dollar disasters.
NOAA ↗Global insured natural catastrophe losses reported for 2024.
Swiss Re ↗Industry coverage of Yifan Yang and Dr. Lei Zou's Texas A&M Hurricane Milton street-view damage assessment research.
Competitive edge
Not another imagery vendor or pre-event risk score.
Rayford AI focuses on post-event decisions: what changed, how severe it is, what evidence supports the score, and what a human reviewer should inspect next.
Street-view plus remote sensing
Imagery-only tools show pixels. Ray Assess compares ground and overhead evidence at the property level.
Built for disaster response
Pre-event underwriting tools forecast risk. Ray packages post-event damage evidence for recovery decisions.
Research-backed audit trail
Model arbitration, confidence, and paper-backed methods make the output easier for public teams to review and defend.
Innovation
A research-backed engine for property-level resilience.
Rayford AI translates peer-reviewed disaster GeoAI research into a practical evidence layer for damage assessment, claims triage, and long-term resilience planning.
Ray
From hazard data to trusted recovery decisions.
Ray is designed to read a property across time: what hazards it has faced, what changed after an event, what evidence supports the model's judgment, and what a human reviewer should do next.
Multimodal damage evidence
Fuse street-view, satellite, parcel, and local hazard context so property damage is assessed with more than one sensor.
Generative missing-view recovery
Explore satellite-to-street synthesis when post-disaster ground views are delayed, incomplete, or unsafe to collect.
Model arbitration and audit trails
Compare model signals, expose confidence, and preserve property-level evidence for human review.
Research foundation
Papers behind Rayford AI.
The company starts from Yifan Yang's research track in street-view disaster assessment, visual-language models, generative GeoAI, and spatial intelligence.
Evidence layer
From imagery to review-ready decisions.
-
01
Link property context
Parcel records, local hazard history, and pre-event imagery.
-
02
Compare post-event evidence
Street-view, satellite, drone, and field imagery where available.
-
03
Arbitrate model signals
Damage scoring, multimodal reasoning, and confidence estimates.
-
04
Export audit trail
Property-level evidence packages for human review.
Team
Built from disaster GeoAI research with technical mentorship.
Founder
Yifan Yang
Technical lead for Ray, focused on street-view disaster assessment, visual-language models, multimodal arbitration, and autonomous GeoAI.
LinkedIn Profile ↗
Scientific and technical advisor
Dr. Lei Zou
Advisor for the GeoAI and disaster resilience foundation behind Rayford AI's research-to-venture path.
LinkedIn Profile ↗
Technical advisor
Dr. Zhengzhong Tu
Committee advisor supporting computer vision, multimodal model design, and validation strategy for Ray's disaster assessment engine.
LinkedIn Profile ↗
Technical advisor
Dr. Heng Cai
Committee advisor supporting built environment context, infrastructure resilience, and product-risk review for property-level recovery decisions.
LinkedIn Profile ↗2026 Roadmap
Ten weeks from research prototype to pilot-ready demo.
-
In progress
01
Complete 40 customer discovery interviews.
-
In progress
02
Build a Ray Assess demo for a historical disaster event.
-
Planned
03
Create two property-level validation case studies.
-
Planned
04
Secure three serious pilot or LOI conversations.
Rayford AI
The resilience AI for every property.
We are scheduling customer discovery and pilot conversations with emergency management, local government, GIS, and recovery teams.