Over the past year, I’ve had more conversations about Artificial Intelligence than I’ve had in the previous ten years combined. Every modular and offsite construction conference I attend seems to feature at least one software company promising to revolutionize production, eliminate waste, solve labor shortages, improve scheduling, and somehow make every factory dramatically more profitable almost overnight. The excitement is real, but so is the confusion surrounding what AI can realistically accomplish inside an offsite factory.
Some factory owners are convinced AI is the next industrial revolution, while others quietly believe it is simply another expensive shiny object that will disappear after a few years, joining the long list of software systems and production technologies that were once marketed as industry-changing breakthroughs. Personally, I believe AI will absolutely become part of the future of offsite construction, but I also believe many factories are approaching it the wrong way by focusing on the technology before understanding the operational problems they are trying to solve.
The Wrong First Question
The first question should never be, “How do we bring AI into our factory?”
The first question should be, “Where are we currently losing money, time, quality, efficiency, or communication?”
That shift in thinking changes the entire conversation. Every factory already knows where many of its weak spots exist. Rework, scheduling confusion, missing materials, engineering bottlenecks, delayed approvals, inconsistent quality control, inventory problems, equipment downtime, and communication gaps between departments are all common issues throughout the industry. AI should not be viewed as magic capable of curing every problem overnight. It should be viewed as another tool designed to help solve specific operational problems that already exist.
Unfortunately, many factory owners are already being approached by software companies selling “AI-powered solutions” before the factory has even clearly identified what needs fixing. That approach often leads companies into expensive systems that create more confusion than improvement.
AI Will Magnify Good Systems… and Bad Ones
One thing decades in construction and manufacturing have taught me is that technology rarely fixes dysfunction on its own. In many cases, technology simply exposes dysfunction faster and on a larger scale.
If a factory struggles with weak communication, inconsistent estimating, poor scheduling discipline, unclear procedures, or management indecision, AI will not suddenly repair those weaknesses. In fact, it may accelerate the chaos by pushing out faster information based on poor inputs. The old phrase “garbage in, garbage out” still applies, no matter how advanced the software sounds in a sales presentation.
Before investing heavily in AI, every factory should take the time to carefully map out how work currently moves through the organization. How do drawings move from sales into engineering? How are production schedules created and updated? How are defects documented and tracked? How are change orders approved? Where do projects routinely slow down or break apart?
Many factories will discover that some of their biggest problems have very little to do with AI at all. Instead, they are process, communication, accountability, or management discipline problems. Identifying those weaknesses early can save a company tens of thousands of dollars before they ever purchase a new software package.
Start Small and Boring
One of the smartest approaches I’ve seen discussed by manufacturing experts is to start with small, low-risk AI applications rather than attempting to automate the entire factory at once. That means resisting the temptation to immediately buy robots or install massive systems that disrupt production while employees struggle to understand them.
Instead, factories should begin with practical pilot programs that solve one specific issue at a time. AI-assisted estimating reviews, predictive equipment maintenance, production scheduling alerts, inventory tracking, quality-control photo analysis, and searchable AI databases for SOPs, codes, and project documentation are all useful starting points.
None of those applications sound particularly glamorous, but they solve real operational problems that affect profitability every day. The factories that ultimately succeed with AI will probably not be the ones making the biggest announcements online. They will more likely be the factories quietly improving one process after another while their competitors are still chasing buzzwords.
The Production Floor Must Be Included
One of the biggest mistakes management can make is treating AI as an office-only initiative. If the production floor believes AI is simply another management tool designed to monitor workers or eventually replace jobs, resistance will begin immediately and quietly spread throughout the factory.
That is why respected production employees need to be involved from the very beginning. The people framing walls, installing MEP systems, moving modules, handling materials, and solving problems on the line every day often understand operational issues that office staff may never fully see.
I’ve watched too many expensive software systems fail because nobody bothered to ask the people actually using them whether they made practical sense in real-world production. AI cannot become another management experiment disconnected from the daily realities of manufacturing.
Data Is the Hidden Challenge
Every serious AI discussion eventually runs into the same obstacle: data. AI systems depend entirely on accurate and organized information in order to provide meaningful recommendations.
That includes production times, labor hours, material usage, purchasing records, defect reports, maintenance logs, warranty claims, schedules, drawings, inspection notes, and project photos. The challenge is that many factories still have critical information scattered across spreadsheets, handwritten notes, whiteboards, emails, ERP systems, and individual employees’ memories.
Before AI can provide meaningful insight, many factories may first need to improve how they gather, organize, and manage information. That work is not exciting and will certainly not generate flashy LinkedIn posts, but it is often the most important part of preparing for AI integration.
Keep Human Judgment in Charge
This may be the single most important lesson of all. AI should assist decisions, not replace experienced judgment.
Offsite construction remains a complicated business involving engineering, transportation, weather, labor availability, inspections, production limitations, customer expectations, and constantly changing building codes. An AI recommendation may sound perfectly logical on paper yet completely impractical in an active production environment.
That is why experienced managers, engineers, supervisors, and production leaders must retain control over final decisions. The best factories will eventually use AI the same way skilled craftsmen use power tools. The tool increases efficiency, but the knowledge, judgment, and responsibility still belong to the person operating it.
The Factories That Will Benefit Most
I suspect the factories that ultimately gain the most from AI will not necessarily be the largest or the wealthiest. They will likely be the disciplined factories willing to measure performance honestly, admit weaknesses, improve gradually, and focus on solving real operational problems rather than chasing hype.
Those factories will understand that AI is not a replacement for leadership, accountability, communication, culture, or experience. Instead, it becomes another valuable tool that helps good organizations operate more effectively and make better decisions faster.
Over time, those small improvements can become major competitive advantages.
Modcoach Observation
Right now, AI in offsite construction reminds me of the early days of factory software systems decades ago. Some owners rushed into technology blindly because they feared missing out, while others ignored it entirely because they feared change. The factories that usually succeeded landed somewhere in the middle. They stayed curious, moved carefully, tested small ideas first, and focused on solving real operational problems instead of chasing industry buzzwords. I suspect AI will follow that exact same path for the offsite industry.



























