Why facilities leaders need to understand what AI can do, where it stops and who remains in control.
Depending on which headline you read, AI is going to transform every industry, eliminate millions of jobs, or end humanity altogether. At the same time, nearly every software company is adding “AI-powered” to its website.
A feature that summarizes a work order? There’s AI for that.
A dashboard that flags an unusual number? Use AI for that.
A workflow routes a request using predefined rules? Also, AI.
We’re using the same term to describe everything from basic automation to systems that some technology leaders warn could eventually escape human control. The public is being asked to believe AI can solve almost any problem while also accepting it could become one of the biggest problems humanity has ever faced. It’s no surprise tension is creating confusion, fatigue, and a lot of skepticism.
A 2026 Bentley-Gallup survey found Americans’ attitudes toward AI had cooled significantly after two years of momentum. Meanwhile, OpenAI and other frontier AI companies continue to acknowledge that increasingly capable systems could introduce severe risks requiring stronger safeguards.
Those concerns deserve serious discussion. But facilities leaders have a more immediate question to answer: Can I trust this technology with the work happening across my locations today?
For facilities teams, the trust problem isn’t new
Facilities leaders aren’t skeptical of AI only because of the headlines. Their skepticism also comes from years of technology that promised more than it ever actually delivered.
They’ve been told a new platform will eliminate manual work, create a single source of truth, and provide complete visibility across the portfolio. Then implementation begins, and the operational reality gets in the way.
Asset records are incomplete. Provider updates arrive by phone and email. Approval rules vary by location. Invoice data doesn’t match the work performed. Meanwhile, the information needed to make a decision is spread across systems, teams, and conversations.
Adding AI to that environment doesn’t automatically fix the gaps, even if it can make decisions from them.
AI: A label that tells you very little
The term AI can describe everything from a tool that summarizes a technician’s notes to a system that recommends how money should be spent. Those are not the same data sources or trained models, and they do not carry the same risk.
The label became even less precise in September 2026, when President Donald Trump directed the executive branch to replace “artificial intelligence” with “superintelligence” in official communications. The White House said the change was meant to reflect the technology’s advancing capabilities and promise.
But “superintelligence” already has a specific meaning in AI research: future systems that would be far more capable than humans. Applying that term to today’s entire AI category doesn’t add any clarity. If anything, it adds more ambiguity to the already murky definition.
Facilities leaders should look past the labels and ask what the technology can actually do. Does it summarize information? Predict a likely failure? Recommend a provider? Approve spending? Write information back to another system? The answers matter more than what the feature is called.
Regulation is raising the standard
As AI and its definition evolve, so do the questions surrounding it. Concerns and questions about AI are no longer coming only from customers and technology teams. We’ve already begun seeing regulators distinguishing between low-risk tools and systems that influence more consequential decisions.
The European Union’s AI Act, for example, takes a risk-based approach. Many everyday applications are considered minimal risk, while systems used in certain sensitive or safety-related areas face stronger requirements.
In the United States, the regulatory picture is more fragmented. The National Institute of Standards and Technology offers a voluntary framework for governing, mapping, measuring, and managing AI risk. States are also developing requirements for specific uses involving consequential decisions.
The companies building the most capable models are also asking for clearer rules. OpenAI has called for mandatory, capability-based national safety requirements and says it’s working with other frontier labs on voluntary standards. OpenAI, Anthropic, Google DeepMind and Microsoft also established the Frontier Model Forum to share safety research and develop responsible practices.
That doesn’t mean the companies agree on every rule, or that voluntary commitments replace independent oversight. But it does show concern about control, access, and accountability is also coming from inside the industry as well as outside it.
The details will continue to change, and not every AI feature in a facilities platform will be regulated in the same way. But the direction is clear: Organizations will increasingly be expected to understand what an AI system does, what information it uses, what risks it creates, and where human oversight belongs.
Facilities leaders need their own AI governance now
Facilities teams shouldn’t wait for a government mandate or a vendor policy before setting their own boundaries. If AI can reach a CMMS, ERP, building system, provider portal or data warehouse, the organization should decide in advance:
- What data the system can read, retain or change.
- Which systems and actions it can access.
- When human approval is required.
- How its activity and decisions will be logged and reviewed.
Start with a narrow, low-consequence use case in a controlled environment. Test whether results are repeatable, failures are visible, and if the controls work as intended. Expand access only when the system has earned it.
Facilities expertise still matters
Experienced facilities professionals understand that operational realities rarely fit neatly into a software workflow. A recommendation that looks efficient on paper may ignore a warranty, a landlord obligation, a site restriction, a safety concern, or a relationship that keeps a location running.
AI can organize information, recommend action, and carry out approved work, but accountability for the outcome should belong to people.
Good facilities AI should make expertise easier to apply. It should preserve the reasoning behind a decision, show the information it used, while leaving accountability with the people responsible for the work.
AI may feel like a dirty word. That doesn’t make it useless.
AI is a slippery topic right now. Its capabilities are changing quickly, public claims often outpace what people can verify, and governments around the world are still deciding how it should be governed. Only time will tell where the technology ultimately leads.
What facilities leaders should control now is how it enters their operation today.
Understanding a CMMS platform’s AI capabilities—and its reach into business data and connected systems—is critical to the health of the organization. Leaders should know what the technology can see, what it can do, how its results are checked, and how quickly its access can be stopped.
The question is not whether AI belongs in facilities management. It’s whether facilities teams can use it without giving up the control, context and accountability the work requires.
AI can still be remarkably useful when it works within clear boundaries. Use it responsibly. Start small. Keep people involved in consequential decisions.