Mud & Microchips · August 2026

What AI Still Can’t Do in Construction (And What That Means for You)

Full disclosure: there’s a whole category of construction work that AI can’t touch yet, and probably won’t for a while. Knowing what’s feasible, helps you make informed decisions on solutions that will actually be used and efficient in your business. If you’ve been burned by a software promise before, this is the post I’d want you to read first.

What AI Can’t Do: Run Your Jobsite

You still need boots on the ground. A jobsite is a live, physical problem that changes by the hour. Think about everything that can move on a single day:

Each one of those is manageable. The problem is they happen at the same time, they interact with each other, and they demand judgment in the moment. That’s the part AI can’t do yet: it can’t stand in the middle of all that complexity, weigh it, and solve it. Real problem-solving on a chaotic site is still a human job.

I want to be clear about the “yet.” I think automation on the physical side is coming sooner than we predict, hybrid job sites will be a real thing in the next 5-10 years.

What AI Can Do: Take the Weight Off the Back Office

Now here’s the other side and it’s a back office of every job. We are talking about reporting, quantities, summaries, and schedules. All of it requires coordination that’s repeatable and necessary to move the job to its next step.

That’s where the real opportunity is. You can’t automate the problem-solving, but you can support it by taking the repeatable back-end work off your plate. My favourite way to describe it: if a report takes ten steps today, the goal is to make it take two. You’re not removing yourself from the work. You’re removing the parts that were never worth your time in the first place.

The tools themselves change constantly, which is exactly why there’s no single “buy this” answer and why a custom approach matters. The rule of thumb is: automate the repeatable, protect the human judgment.

Why You Should Start With the Office, Not the Field

We sat down with two of the three partners at a construction company. The very first thing they said to us was: “Our field people won’t be willing to switch to any digital tool.”

We hear that a lot, and the resistance is real. Training people is genuinely hard, and I’m not going to pretend it isn’t. But as I sat across from them, I was thinking something they hadn’t considered: the field isn’t where you start.

The bottom line of that whole conversation was this: before any tool ever reaches the field, they needed a project management tool for themselves, for the office side, for full visibility into their own jobs. They didn’t have that yet. So the field resistance they were worried about wasn’t even the right problem to solve first. You can’t roll a tool out to the crews when the office doesn’t have its own house in order.

The same tool that gives the office visibility is the one that gives you back time - time you can spend overseeing and mentoring your field people instead of chasing paperwork. Start with the office. Field adoption becomes a separate, later question, and a much easier one once the value is already obvious.

What This Means for You on Monday Morning

So you’ve accepted that AI can’t run your jobsite but can take your office from ten steps to two. What do you actually do about it?

1. Find the task you drag your feet on

Look for the admin work that you, the owner, keep putting off. Anything that’s stalling the process and delaying the job should be first in line. That’s your best candidate, because that’s where the pain and the payoff are both highest.

2. Automate the doing, keep the approving

The model is simple: the task itself gets automated or delegated, someone gets assigned to oversee it or owns it from the start. You only approve it. You stay in control of the decision without being stuck doing every step by hand. If you can’t delegate, you don’t have a business - you have a job.

3. Where to look first

Some of the areas we explore with clients, where the doing can be automated or delegated and you keep the final sign-off:

Each of these still needs your approval at the end. None of them needs you doing all the manual work in the middle. (If you want the numbers on what that time actually adds up to in dollars, we break the labour-savings math down in “Can a Small Construction Firm Afford AI?” - worth a read alongside this one.)

The Real Point: It Was Never About the Technology

If you take one thing from this, let it be this. What we help our clients with isn’t AI. It’s getting time back.

The tools are always changing and that’s why a custom approach beats chasing whatever’s trending. But the whole point and the core mission behind what we do is much simpler. What we genuinely want for you is fewer headaches when you walk back into your shop on Monday morning.

My business partner Mat puts it better than I can: problem-solving is a lot easier when you can focus only on the challenging situations, instead of reminding all the trades what colour of hard hat they should wear.

Start with the office. Keep your boots on the ground. Give yourself your Monday back.

Mud & Microchips helps construction and trades businesses implement AI that actually works — starting with an operations audit that measures where your hours go.