Artificial intelligence and new software promise to change how buildings get delivered. They deliver on that promise only under two conditions:

• The data underneath has to be trustworthy.
• The people using the tools have to be able to trust them in their work.

In the first article in this series, we highlighted that the hardest problem in construction and design has not changed. The right information still fails to reach the right people early enough, and projects still pay for it.

That is not an argument against technology. The tools are improving quickly, and they solve problems that were genuinely hard a decade ago. But if they have not closed the oldest gap, what has to be true for these tools to deliver?

Artificial intelligence sits at the center of that question right now. The honest answer is more encouraging, and more demanding, than the marketing suggests.

A tool raises the ceiling on expertise

The most useful way to think about new technology is not as a replacement for people, but as a multiplier of what skilled people can already do. Software and artificial intelligence do not supply judgment. They extend the reach of the judgment a team already has.

A specialist who understands openings can cover more ground, catch more conflicts, and coordinate across more of a project when the right tools carry the repetitive load. The same tools in the hands of someone without that expertise produce confident output that no one should trust.

Technology cannot manufacture that judgment. What it can do is help the experts who remain apply theirs more widely, and help newer people work closer to a veteran's level than they could alone.

With the introduction of AI and machine learning, these are not going to wave a magic wand. They enhance and bring up the human factor. They are enhancements to human experience and knowledge, and that raises what the whole industry has access to.

Shari A. Muscat, AHC, CSI/CDT, General Manager of Openings Studio, on technology as an enhancement to expertise

The teams getting real value from new tools are not the ones replacing people with software. They are the ones pairing experienced people with software that lets them do more of what they are already good at.

The bottleneck is adoption, not capability

The capability of these tools is rarely what holds them back. The obstacle is whether people will actually use them. New platforms arrive faster than any team can evaluate, and each one asks a professional to change how they already work.

Most will not make that trade unless the tool fits cleanly into their workflow.

You have hyper speed development of different applications and AI platforms and add-ins. It has to be seamless for the collective, because people do not want to learn new platforms and new programs. How does this work with what I am already doing. That can be a huge roadblock, and cybersecurity is a big one.

Ryan McLachlan, AHC, CM-BIM, CFDAI, LEED Green Associate, on why adoption stalls

This is the part the technology conversation tends to skip. A tool that is powerful but disruptive loses to a tool that is modest but seamless, every time, because the modest tool actually gets used. Integration, ease of use, and security are not secondary features. They are the difference between a tool that changes how work gets done and a license that goes unused.

A technology's measure is not what it can do in a demonstration. It is what it does inside a working team under a deadline.

Every tool runs on data it can trust

Underneath both points sits a single precondition. A tool is only as good as the data it runs on. Artificial intelligence makes this truth sharper, because a model trained or prompted on incomplete, outdated, or untrusted information does not fail loudly. It produces a confident wrong answer, which is worse than no answer at all.

The evidence is already clear, and it is not specific to construction. Gartner predicts that through 2026, organizations will abandon 60 percent of AI projects that lack the trustworthy, well-managed data those projects depend on. The tools are not the point of failure. The data underneath them is.

Construction feels this more acutely than most industries, because its data is fragmented by default. Information is scattered across drawings, schedules, models, and email, produced by different teams who do not always trust one another's numbers.

Feed a powerful tool unreliable data, and it will scale the error instead of catching it.

Software and expertise, working together

Put the three together and the path forward is straightforward, if not easy. Technology delivers when capable tools, built to be adopted, run on data that people can trust, in the hands of experts who know what good looks like. Remove any one of those and the value collapses. A brilliant tool on bad data misleads. Clean data in a tool no one will use sits idle. The best tool in the hands of someone without judgment produces answers no one should act on.

That is why the answer isn't software instead of expertise, or expertise resisting software. It is the two together. The people who understand the work decide what matters and whether the output is trustworthy. The tools let them apply that understanding across more of a project, faster, and earlier, while decisions are still open. That combination is what moves a project, and it is available now.

The final article in this series turns to the people who lead this work, and to what the shift asks of them. The through-line holds to the end. A tool is only as good as the data it runs on and the judgment that guides it. Technology raises the ceiling. Expertise is still what reaches it.

Openings made easy