AI Adoption
What AI Actually Does
By John J. BakerJune 26, 20266 min read
Most contractor owners we talk to fall into one of two camps.
The first camp is ignoring AI. They have heard enough about it to know it is moving fast, decided they do not have time to figure it out, and moved on. This is understandable. Running a $15 million painting operation leaves very little room for experimentation.
The second camp is trying to keep up with everything. They are subscribing to newsletters, watching demos, sending their estimator to a half-day AI class, and wondering why nothing has changed three months later.
Both camps are solving the wrong problem.
The actual bottleneck is not capability. It is attention.
Here is what we observe consistently across commercial contractors in the $10M to $50M range:
The business is not short on capable people. It is short on mental capacity.
Your estimator is capable. But she is also fielding calls while building quotes, chasing down RFI answers, and mentally tracking which jobs are waiting on materials. When a quote goes cold because no one had time to follow up, it is not a skills problem. It is an attention problem.
Your GM is capable. But he is also the person everyone comes to when they cannot find an answer. When he spends 25 hours a week reviewing quotes for margin before they go out, that is 25 hours he is not spending on the work only he can do.
Your superintendent is capable. But when he is manually building the crew schedule by calling eight people every Sunday afternoon and updating a whiteboard that is wrong by Monday at noon, that is not a competency problem. It is a design problem.
AI does not fix the capability gap. It fixes the attention gap.
What AI actually removes
The work that consumes mental bandwidth without producing anything genuinely new is exactly what AI handles well. Specifically:
Repetitive decision sequences. Your estimator applies the same logic to every quote. Your bookkeeper runs the same five reports every Monday. Your PM assembles the same closeout packet at the end of every job. These are not creative acts. They are rule-following acts. AI follows rules extremely well.
Information retrieval and formatting. When your team spends time hunting for information that already exists in the business and reformatting it for a different audience, that is friction AI eliminates. Voice notes from the truck become structured field logs. Five separate reports become one consolidated view.
Follow-up that falls through the cracks. Not because anyone forgot. Because everyone was busy doing something else that felt more urgent. Quote follow-up. Collections outreach. Job status updates to clients. These are high-value activities that consistently get deprioritized. AI handles the trigger and the first draft.
What AI does not remove
This matters just as much.
AI does not replace the judgment call on a difficult bid. It does not handle the conversation with a client whose expectations need to be reset. It does not manage a subcontractor who is not performing. It does not know when a job has a dynamic that your estimating model has not accounted for.
The work that requires relationships, pattern recognition built over years, and real-time situational judgment is not going anywhere. That is the work your best people should be doing more of, not less.
The goal is not to automate your team. It is to give your team back the hours they are currently spending on work a system could handle, so they can spend more of their day on the work only they can do.
Why most AI training does not stick
We have seen the pattern often enough to describe it precisely.
A business sends someone to a half-day AI course. The course is well-designed, run by smart people, and covers real tools. The attendee takes notes. They come back to the office, try to apply what they learned to their actual workflow, and run into friction immediately. The tool does not connect to the data they need. The workflow they want to automate is more specific than what the course covered. They have a question with no one to ask.
Within two weeks, they are back to the old process. Not because they are resistant to change. Because the gap between "I understand how this works" and "I have this running in my workflow on Monday morning" was never bridged.
The information was real. The application never happened.
What bridges that gap is building the solution on the actual workflow, with the actual data, before training begins. So that when a team member sits down for the first session, they are not watching a demo. They are watching their own work run better in front of them.
That is what changes behavior.
A practical way to think about where to start
If you are trying to decide whether and how to move on this, the question is not "what AI tools should we adopt?" The question is: where in our business is the most capable person spending the most time on work a system could handle?
That is the workflow to start with. Not because it is the most technically interesting, but because it produces the clearest before-and-after that your team can see. A visible win with one person in one workflow does more for adoption across your organization than a company-wide training initiative.
Start there. Get it running. Let that person's results make the case for the next one.
Three questions worth sitting with
Before talking to anyone about AI, including us, these are worth thinking through:
Where is our most capable person spending time on work a system could handle? Be specific. Not "admin" or "reporting." The actual task. How often. How long.
Who on our team would be genuinely excited to try something new, and already owns a workflow that could be improved? AI adoption stalls when it starts with skeptics. It spreads when it starts with someone whose results are visible.
What would change about our business if that person had two extra hours every week? If the answer is "not much," find a different workflow. If the answer is meaningful, you have found your starting point.
We work with commercial contractors who have identified a workflow that is costing them. If you are at that point, a 30-minute conversation is the right next step. No pitch. We will tell you honestly whether we can help and what that looks like.
Prefer a one-pager?
We put the key ideas from this article into a single printable page.