During her keynote presentation at the Johnson Controls Digital Technology Conference, Dr. Koitz shared her vision for the next evolution of AI in facilities and workplace management. She envisions a future where AI moves beyond being an intelligent add-on to become an active operational partner.

According to Koitz, rather than simply providing insights, AI will increasingly work alongside people, helping them spend less time managing systems and more time delivering value.

AI is becoming a member of the team

To illustrate this shift, Dr. Koitz encouraged attendees not to think of AI as the technology behind better dashboards, search functions and analytics capabilities. Rather, it is an always-available teammate that understands context, recommends actions and carries out routine tasks.

“It's a team member that becomes an all-seeing and ever-learning ally. It has read all the manuals, knows your data, never sleeps and can interact with you using natural language,” said Dr. Koitz. “It can continuously optimise and it can even help our infrastructure self-heal. Last but not least, it can deliver that amazing occupant experience that boosts performance, productivity and outcomes.”

That shift in thinking is significant because it reframes AI from a tool that people use occasionally into an operational capability that works continuously in the background.

From data fatigue to driving business outcomes

Before digging into her vision of the future of AI, Dr. Koitz described how its application in facilities and workplace management has evolved to this point.

She pointed out that ten years ago organisations were overly focused on connecting systems and harmonising data to gain greater visibility into building performance.

“We got connected, but dashboards kept on getting very busy,” Dr. Koitz explained. “There was quite a bit of data fatigue.”

Today, organisations are building on those capabilities to enable predictive insights, where AI transforms data into forecasts, recommendations and operational intelligence.

"We know how space is utilised and we can forecast things, which is great," Dr. Koitz said. "But that’s not enough. We should absolutely not stop here, because we can do so much more."

The next phase will see AI take a more active role – turning insights into meaningful action. Rather than simply presenting information and leaving people to determine the next step, AI is increasingly able to understand context, recommend actions and execute routine tasks. With these enhanced capabilities, organisations can move away from solely focusing on equipment-level metrics to looking at broader business objectives.

“My personal view is that we should be able to give these systems a business KPI such as a sustainability, energy efficiency or workplace experience goal,” said Dr. Koitz. “That should take us into autonomous mode.”

People have grown accustomed to technology that leverages vast amounts of data to anticipate their needs and make their life easier, often with little or no effort required. Navigation apps reroute around traffic automatically. Shopping platforms provide personalized recommendations. Digital assistants respond conversationally.

People increasingly expect the technology they encounter inside buildings – whether it’s a workplace, hospital or classroom – to be as intuitive and responsive as the digital experiences they enjoy in their everyday lives.

The role of agentic AI 

The phrase "agentic AI" seems to be everywhere. However, many people still don’t understand what it means. Dr. Koitz pointed out that it comes down to allowing AI to complete routine work that would otherwise consume valuable time.

Sometimes a person delegates a task to an AI agent, such as creating and distributing reports. Other times, an AI agent may identify an emerging issue and alert the appropriate person to take action. Finally, AI agents may delegate tasks to one another to initiate actions autonomously. Advanced agentic AI implementations may do all these things concurrently, based on the situation and guardrails the organisation has put in place.

Though their capabilities may seem impressive, Dr. Koitz stressed that these systems should be viewed as junior members of the team rather than replacements for experienced professionals.

“These agents are not facility managers. They can take over your tedious stuff, but they cannot take over your strategic thinking,” explained Dr. Koitz. “So, don't be afraid of them. Embrace them as somebody who can execute what you ask them to, and they should integrate in daily workflows.”

How we build agentic AI is just as important as what we apply it to. As AI becomes more capable, trust becomes even more important. Dr. Koitz emphasised that responsible AI is defined not only by what it can do, but by how it is designed.

That includes, for example:

• Ensuring that organisations retain ownership of their data
• Selecting algorithms that fit specific business needs rather than applying AI indiscriminately
• Keeping humans firmly in control
• Establishing clear guardrails around autonomous actions
• Making AI-generated insights and recommendations transparent
• Collecting feedback for continuous learning and improvement

"We build AI with a lot of care," said Dr. Koitz. "There's no shortage of technology and resources available to develop an endless number of AI solutions. We could dedicate enormous amounts of compute power just so we could say, 'Look, we're using agentic AI.' But that's not very smart. That's why our first step in designing an AI solution is to think about how it addresses the true business needs.

These principles help organisations adopt AI with confidence while ensuring the technology remains aligned with operational goals,"concludes Dr. Koitz.