The Radiology Lesson: Why AI Won't Replace Facade Inspectors
- Hammer Missions
- 1 day ago
- 3 min read
In 2016 Geoffrey Hinton told a room of computer scientists that people should stop training radiologists. A decade on, the opposite happened: US diagnostic radiology residency programmes offered a record 1,208 positions in 2025, up roughly 4% year over year, with vacancy rates at all-time highs and average compensation up sharply since 2015. There are now over 700 FDA-cleared radiology AI models — around three-quarters of all cleared medical AI devices — and demand for human readers has still gone up.
This matters to the built environment because facade, parking and structural inspection share radiology's core profile: a regulated professional, a visual assessment task, a digital input, and a signed deliverable carrying liability.
Why the automation didn't land
Three technical reasons, all of which transfer directly to inspection work.
Benchmark performance is not field performance. Models trained on standardised datasets have been shown to lose up to 20 percentage points of accuracy when deployed at a different hospital. The inspection equivalent is obvious to anyone who has run a defect model across two building stocks: a crack detector tuned on precast panels in a temperate climate degrades on rendered masonry in a marine environment. Domain shift is the rule, not the exception.
Interpretation is a minority of the job. A study following staff radiologists found only around 36% of their time went to direct image interpretation. The rest was protocol review, communication with clinicians, supervision and teaching. Substitute "QEWI" for "radiologist" and the ratio is familiar: site coordination, access planning, client communication, SWARMP classification arguments, repair scoping and report defence.
The constraint was supply, not demand. Imaging volume kept climbing with an ageing population while residency growth lagged. AI didn't reduce the queue — it made the queue addressable.
💡 Enjoying these insights? Subscribe to the Hammer Missions newsletter for expert tips, real-world drone inspection case studies, and the latest updates on how AI is transforming building assessments:
The same arithmetic in the built world
The inspection backlog is structurally identical. New York's Facade Inspection Safety Program covers buildings over six storeys on a five-year cycle, with roughly 12,500–18,000 buildings in scope depending on how you count sub-cycle assignments. Singapore's Periodic Facade Inspection regime, live since 1 January 2022, was projected to bring about 30,000 buildings into a seven-year cycle. In the US, facade ordinances exist in Boston, Chicago, Cincinnati, Cleveland, Columbus, Detroit, Jersey City, Milwaukee, New York, Philadelphia, Pittsburgh, San Francisco and St. Louis — and that list has only grown.
Now add everything not covered by an ordinance: parking structures, bridges, industrial assets, portfolio-wide capital planning surveys. The demand is latent because the old delivery model — swing stages, rope access, boom lifts, one engineer sequentially — could never price it low enough to unlock.
What AI actually changes, and what it doesn't for facade inspectors
The baseline moves. What was a high-leverage task (manually marking every defect on a 40-storey elevation) becomes a low-leverage task. The new high-leverage tasks are the ones AI cannot do: deciding which conditions are Unsafe versus SWARMP, choosing where to spend the hands-on budget, defending an engineering judgement, and carrying the stamp.
Firms that price to hours will feel this as fee compression. Firms that price to outcomes will run four or five times the volume through the same professional headcount. Empowered by AI, the facade inspectors themselves will see four or five times as many structures per year, which compounds their judgement faster than any training programme.
Interested in learning more about drone-based facade inspections or seeing how AI can enhance your workflows? Reach out to the Hammer Missions team — we’d love to show you how to bring this process to your next project.
About Us
Hammer Missions is a software AI firm helping companies in the built environment leverage drones and AI for assessing existing conditions. Having seen 5000+ projects, we're pleased to be working with leading firms in AEC to streamline and scale the process of facade inspections. If you're looking to learn more about how AI can automate and accelerate your building assessment projects, please get in touch with us below. We look forward to hearing from you.
