Image source: Public Domain
Partly, the Austin-based company building frontier AI for repair, launched PartlyRepair AI at SEMA 2026. PartlyRepair AI is the agentic AI assistant for collision repair shops, powered by Interpreter, Partly’s proprietary AI model that understands and reasons across vehicles, parts and repair procedures. It automates the research and admin across the entire job, from intake and damage capture to procurement and job close.
In collision repair, the margin is in the details: a line left off the estimate, the wrong part for the trim, a supplement without the evidence to back it, a car waiting on parts. Each miss erodes margin and extends cycle time. PartlyRepair AI was shaped by conversations with more than 150 collision shops and is built to catch those mistakes before they cost the shop.
“Ask a shop owner where the day goes and you hear the same list: going back over estimates line by line, checking prices across supplier portals, and chasing the documentation to get a supplement paid,” said Bill Lopez, SVP, Commercial, North America at Partly. “Every hour spent on that admin is an hour away from the cars in the bays. That’s the time PartlyRepair AI gives back.”
PartlyRepair AI already has six specialized agents live, with more to come. Shops see value across their workflows, including:
Interpreter is what sets PartlyRepair AI apart. Partly has spent six years training it, with a team that pairs AI researchers with experienced estimators, parts specialists and technicians. Interpreter draws on licensed OEM build sheets and data to answer vehicle-specific parts and repair questions on more than 1 million vehicle configurations. Interpreter is already at work in shops throughout New Zealand and the United Kingdom, where repairers using Partly’s platform order parts nine times faster and have 2.4 times fewer returns.
“General-purpose AI can read an estimate. But understanding what a specific vehicle actually needs to be repaired is a different problem,” said Levi Fawcett, CEO of Partly. “We've spent six years building the intelligence to solve that problem. Now we're putting it directly in the hands of repairers, helping them protect their margins while keeping them in control of every decision.”