Based on insights from McKinsey & Company’s latest research on AI and operating models in real estate
Most real estate companies have tried AI for building management systems by now. Plenty have run pilots. Some have even seen demos that looked genuinely impressive. And yet, very few have seen real financial impact.
McKinsey & Company calls this “pilot purgatory” – and according to their latest research, this is where the majority of real estate firms are stuck today.
So what separates the 20% that are actually getting value from AI?
It has very little to do with which AI model you use – and everything to do with how your workflows are designed.
The problem is not the technology. It is the operating model.
One of the sharpest insights from McKinsey & Company is simple:
When a pipe bursts at 6am, your tenant does not care whether you are using Gemini or Claude.
What shows up is your operating model – your workflows, your response systems, your ability to execute.
The AI model is only as useful as the processes it is embedded in.
Right now, most real estate workflows were never designed for AI.
- Work orders come in through fragmented channels
- Maintenance coordination is manual
- Tenants are left wondering if anyone received their request
- There is no learning loop to improve over time
AI can absolutely help with all of this.
But only if the workflows are built to let it act.
Pick a domain, not a list of use cases
One of the most practical ideas in the research is the “Goldilocks mindset”.
Most companies go too small:
- collecting AI use cases like a checklist
Or too big:
- launching vague transformation programs that never ship
The sweet spot is a domain.
A domain is a connected slice of work that is:
- big enough to matter
- focused enough to actually change
- tied to clear business KPIs
Examples include:
- leasing and renewals
- maintenance and facilities
- asset management reporting
These are areas with high volume, messy handoffs, and clear business impact – exactly where AI can create value.
Automate the steps. Protect the thoughts.
Once you pick a domain, every workflow can be broken into two parts:
Steps
Repetitive coordination work:
- routing tickets
- scheduling vendors
- logging results
Thoughts
Human judgment:
- decision-making
- relationships
- high-stakes approvals
The goal is not to replace people.
The goal is to remove coordination overhead so humans can focus on what actually builds trust and drives value.
This is where agentic AI becomes real.
Software that does not just generate insights, but monitors, decides, and acts across your building systems.
From insight to action: what this looks like in practice
For example:
A peak shaving agent can detect an upcoming power spike, analyze its root cause, and automatically reduce load by adjusting EV charging or HVAC systems – before it impacts your energy costs.
No dashboard.
No manual coordination.
Just execution.
Trust must be designed into the system
Another key insight is that trust is not optional – it is a product requirement.
AI systems must include:
- traceability
- auditability
- clear escalation paths
- human-in-the-loop for exceptions
The most successful companies design governance from day one:
- clear automation levels
- defined guardrails
- continuous feedback loops
This is what allows AI to move from experiments to real operations.
The compounding advantage
Here is what makes this shift powerful.
Once you redesign a domain properly, you build something competitors do not have:
- proprietary data
- connected workflows
- learning loops from real decisions
AI tools will become commoditized.
But operating models will not.
The advantage will belong to companies that:
- connect their systems
- redesign their workflows
- and let AI act
From fragmented systems to autonomous operations
This is exactly the gap ProptechOS is designed to solve.
Not another AI tool.
But the execution layer that connects data, workflows, and actions across your building portfolio.
ProptechOS enables real estate owners to move from:
- dashboards → execution
- insights → actions
- manual coordination → autonomous workflows
From sense → reason → act.
The real question is not about AI
McKinsey & Company estimates that AI could unlock $430–550 billion in value across real estate.
The question is not whether the technology is capable.
The question is whether your workflows are ready to let it act.
See it in action
See how AI agents detect, decide, and act across your buildings in real time.
