The labor math decides what AI is actually for
The trades are short roughly 110,000 licensed technicians in 2026, and about half of employers report they cannot find skilled applicants. That gap does not close by recruiting harder, because a licensed tech takes years of apprenticeship to produce. The pipeline is the constraint.
Which is part of why private equity is buying your competitors and, frankly, buying them for their people. When a licensed technician is the scarcest asset in the trade, acquiring a company becomes a hiring strategy. Construction is living the same arithmetic and needs 349,000 net new workers this year.
So when a vendor pitches AI to your team as a way to run leaner, they have misread the market. Nobody in this trade is trying to need fewer technicians. You could not hire more if you wanted to.
Every hour a licensed tech spends on paperwork, driving, or a callback is an hour of the scarcest resource in your industry spent on something unlicensed.
What I bring to the room
At Yembo I built AI that field teams in more than 20 countries use to assess a property from video, and produce the scope and inventory without sending someone to look first. That is the same problem your dispatch board has every morning, in a different uniform.
I talk about what worked, what did not, and the part most speakers skip: how to tell a crew that software is going to touch their job without losing the best of them to the shop down the road. Technicians have heard the word automation before, and they did not hear it as good news.
If your team wants to go deeper than a keynote, the agentic workflows workshop takes operations leaders through their own processes and ranks them by where automation pays back first. Contractors doing remodel and build work usually want to read about AI for construction as well.
