Kespry CEO Explains How Drones and AI Transform Hurricane Roof Damage Assessment

Traditional insurance adjusters faced perilous climbs and painstaking measurements to gauge roof damage after hurricanes, with injury rates nearly four times higher than typical construction workers. They relied on ladders, cameras, tape measures, and manual extrapolation to estimate total property damage.

Kespry’s CEO, George Mathew, argues that drones can replace the ladder, delivering data faster, safer, and more accurately. “Enter the drone: you simply tap the area on an iPad, outline a polygon, and let the system capture every pixel needed for analysis,” he explains. This approach has already pushed Farmers Insurance’s daily capacity from three homes to three homes per hour—an 8‑ to 10‑fold productivity jump reported at Dreamforce 2024.

Beyond speed, drone imagery allows precise roof dimensions and feeds into AI models that detect hail impact spots and infer wind‑induced damage. The result is a richer, more reliable data set that feeds directly into insurers’ claims software, reducing manual labor and error.

Kespry’s roots trace back to 2013 when the founders focused on industrial IoT. Their first major win was in mining, where they used drones to photograph stockpiles, processed the high‑resolution images into 3‑D models via photogrammetry, and achieved 1‑ to 2‑percent accuracy—dramatically better than the 15‑ to 20‑percent variance of conventional survey methods. The same technology now measures construction materials and other heavy assets, opening a new frontier for computer vision in industry.

While tech giants like Amazon, Google, FedEx, and Uber have piloted drone programs, Kespry emphasizes practical, immediately deployable solutions. After the initial flight restrictions that keep drones out of damaged zones during the critical first 48 hours post‑hurricane, the airspace reopens, allowing systematic surveys to begin.

Mathew notes that the same methodology could apply to other catastrophes, such as fires and tornadoes, though assessment complexity increases with total loss scenarios. Even then, drones can map hollowed foundations, informing replacement cost estimates.

Looking ahead, Kespry sees “specified AI”—targeted algorithms for concrete decision‑making—as the most valuable direction. Their hail‑damage detection, powered by deep‑learning neural networks trained on high‑resolution imagery, exemplifies this trend. The more focused the use case, the higher the value delivered to businesses.

With partners like State Farm and Farmers Insurance and a proven track record across industrial sectors, Kespry demonstrates that drone‑based aerial intelligence, combined with machine learning and computer vision, is reshaping how we assess and respond to disaster damage.