A Modcoach Warning About Agentic AI
For years, modular factory owners worried about the same things every morning when they walked into the plant. Material costs. Labor shortages. Transportation delays. Backlogs. Interest rates. Finding enough set crews. Trying to keep quality up while speeding production up at the same time.
Those problems are still here and probably always will be. But something new is beginning to quietly move into the offsite industry that I don’t think enough people, especially older management, fully understand yet.
It’s called Agentic AI.
Now, before anyone stops reading because they think this is another article about ChatGPT writing marketing posts or AI creating goofy LinkedIn graphics, that’s not what I’m talking about. Those are toys compared to what’s coming next.
What I’m talking about are AI systems that can observe operations, learn from experience, make decisions, coordinate activities, predict outcomes, and slowly improve themselves over time with very little human guidance.
And if that sounds harmless, keep reading.
It Starts Out Looking Helpful
Like most things in construction, the first version usually arrives wearing a friendly face.
An AI company approaches a factory owner and says they can help reduce waste, improve production flow, tighten scheduling, predict maintenance issues before equipment breaks down, coordinate deliveries, improve estimating accuracy, and maybe even reduce warranty claims.
Most owners would gladly listen to that conversation today. Some would sign the contract before lunch.
Especially now, when margins are tight, and everyone is trying to figure out how to produce more housing with fewer people.
At first, the AI simply watches and learns. It studies how the factory operates. It monitors production schedules, labor efficiency, purchasing habits, delivery timing, downtime, engineering changes, supplier issues, and customer complaints. Over time, it begins to recognize patterns that no human being could ever track manually.
That’s where things begin to change.
The Real Product Isn’t the Software
Many older managers still think software is the product.
It isn’t.
The real product is the information the software collects.
Imagine one AI company connected to dozens of modular factories around the country. Add in builders, suppliers, transportation companies, crane services, developers, and installation crews. Every production problem, every delay, every warranty issue, every scheduling mistake, every labor shortage, and every successful solution eventually flows into the same intelligence system.
At some point, that software company may understand the offsite industry better than the people actually running it.
And that’s the part nobody seems eager to talk about yet.
Your Factory Could Become Someone Else’s Teacher
Let’s say two factories use the same AI platform.
The first factory spends years feeding the system data on estimating, labor performance, production sequencing, scheduling, supplier issues, and warranty issues. The AI studies it all and learns what works and what doesn't.
Now another factory signs on.
Without directly sharing confidential files, the AI supplier already knows which production approaches are more profitable, which workflows create bottlenecks, which scheduling systems break down under pressure, and which management habits lead to operational problems.
One factory unknowingly became the training ground that helped improve another factory.
That may not sound unfair at first, but what happens when the AI supplier begins investing financially in selected factories?
The Day the Software Company Wants Ownership
That day is coming.
Maybe not tomorrow, but eventually.
A startup modular factory struggling with capital gets approached by a sophisticated AI company offering operational systems, robotics integration, automated scheduling, estimating support, purchasing intelligence, and production optimization. The startup can’t afford the full implementation cost.
Then comes the offer.
Instead of charging millions upfront, the AI supplier takes partial ownership in exchange for the technology.
Many startups would jump at that opportunity immediately.
But now the software company may have access to operational data from dozens of competing factories while also having financial interests tied to one of them. Even if nobody intentionally manipulates anything, the temptation to steer the best insights, improvements, and innovations toward preferred partners could become enormous.
That’s no longer just a software relationship. That becomes an influence over the direction of the industry itself.
Older Management May Underestimate This
I still run into factory owners who barely trust cloud software, use weak passwords, or hand over operational information without thinking twice about where it ends up. Some still believe technology companies are simply “vendors.”
The younger technology world doesn’t think that way anymore.
To many AI companies, data itself is the business. The software is simply the tool used to gather it.
That difference in thinking could eventually become dangerous for offsite construction because this industry has always been fragmented, undercapitalized, and hungry for operational improvement. That makes it extremely attractive to technology companies looking to dominate a niche industry before the industry fully understands what’s happening.
And once a factory becomes deeply dependent on one AI ecosystem for estimating, scheduling, purchasing, engineering coordination, and production management, walking away from that system may become almost impossible.
The Scary Part is That AI Could Actually Work
This is what makes the situation so complicated.
Agentic AI could genuinely help solve some of the offsite industry’s biggest problems. It could improve quality control, reduce waste, predict maintenance failures, tighten production schedules, coordinate trucking more effectively, and help factories operate with fewer costly mistakes.
Some factories may become dramatically more profitable because of it.
Others may finally achieve the consistency the industry has struggled with for decades.
That’s why owners will adopt it. Not because they’re foolish, but because the operational advantages could become impossible to ignore.
And that’s exactly why factory owners need to start asking harder questions right now instead of five years from now.
Who owns the operational data? How is it being used? Can the AI supplier aggregate lessons learned across multiple factories? What happens if the supplier also owns part of competing operations? Can factories ever fully disconnect from the system after years of integration?
Most current software contracts aren’t even remotely prepared for those conversations yet.
This Industry Has Seen Versions of This Before
Years ago, factory owners worried about competitors driving by the plant, counting modules sitting in the yard to estimate backlog. Today, some companies are preparing to voluntarily hand over nearly every operational detail of their businesses to cloud-based intelligence systems that run 24 hours a day.
The buildings may still belong to the factory owners. The equipment may still sit on their production floors. The employees may still wear company shirts with the factory logo on the back.
But someday, the real power in offsite construction may belong to the companies quietly collecting the intelligence flowing through them all.
Modcoach Observation
The offsite industry has always loved innovation. New machinery, new software, new robotics, new materials, and new systems always attract attention because everyone wants to believe the next breakthrough will finally solve construction’s biggest headaches.
But every once in a while, an innovation arrives that doesn’t just improve the industry. It changes who controls it.
Agentic AI may become one of those moments.
And by the time many older managers fully understand what they signed up for, the contracts may already be signed, the systems may already be embedded into their operations, and the companies collecting all that intelligence may have become more powerful than the factories themselves.
No comments:
Post a Comment