Quick answer: Usage is up, belief isn’t. ServiceTitan’s 2026 State of AI in the Trades polled 1,017 contractors: only 11% place high or complete trust in the AI tools they use, 58% trust them little or not at all, and yet 66% of AI users already save 3+ hours a week. We use it. We just don’t believe it. Azati’s answer is vetting, not hype: our P&ID agent, built with DIGATEX, was vetted by KBR and rolled out across 20+ operating plants and 10+ warehouses. It processed 120,000 engineering documents in under 24 hours at 98.8% accuracy, and the technology has handled 500,000+ legacy isometric drawings for PETRONAS. So the real question: who has vetted the AI your team uses every day?
The shallow lesson: better models will fix trust
Accuracy matters – it’s the top concern among AI users in that survey. But a higher benchmark score doesn’t make a site engineer believe the output.
The real lesson: trust transfers through people and proof
Years of accumulated expertise and experts willing to put their name on the result are what move AI from “interesting” to “approved.” It works the same way in a ballroom: three strangers become colleagues in minutes once someone vouches. Trust, sheet by sheet.
The sobering part
Saving three hours a week with a tool you don’t believe usually means someone is quietly double-checking it. Until trust exists, part of every AI productivity gain is spent on verification nobody budgets for.
Full disclosure: Azati built that P&ID agent with DIGATEX, so yes, we’re partial to “vetted beats viral.” See our engineering AI work.
FAQ
Adapted from a LinkedIn post by Sergei Ivanov, CEO, Custom Software & AI Solutions in GCC – read the original, written after The Exchange: Leaders in Construction Middle East (Construction Week Middle East). Source: ServiceTitan, The Update: 2026 State of AI in the Trades, September 2026.
