Before you buy AI for corporate real estate, make sure:
✓ It solves a real decision problem you already have.
✓ You can trace every answer back to its source.
✓ It works with your existing systems.
✓ You understand what happens when it is wrong.
If those conditions are not met, you are not buying intelligence. You are buying risk.
Everyone is selling AI right now. Not everyone understands what they are selling.
Corporate real estate teams are under pressure to "do something with AI." Boards are asking about it. CEOs mention it in town halls. Vendors who sold you your lease system three years ago have suddenly become AI companies.
Before you buy anything, slow down.
The cost of buying the wrong AI solution is not just the license fee. It is months of implementation, data exposure risk, internal credibility burned when the tool does not deliver, and change fatigue for a team already stretched thin.
Lease obligations, operating expense recoveries, capital approvals, compliance reporting, board reporting — these decisions do not tolerate black-box answers.
This is not a checklist for finding the flashiest AI vendor. It is a set of questions to help you determine whether you need AI at all — and if you do, how to separate substance from sales pitch.
What Good Looks Like
Before we get into the questions, define the standard.
In corporate real estate, responsible AI should produce answers that are traceable back to source documents and systems. It should use deterministic logic where financial or compliance risk is involved. It should work across your existing systems without requiring a rip-and-replace migration.
Anything that cannot meet those three conditions is not enterprise-grade.
1. What Problem Are You Actually Solving?
This sounds obvious. It is not.
Most AI purchases in CRE begin with a mandate, not a problem. Someone senior says, "We need to be using AI," and the team starts taking demos. By the third demo, the original question has been replaced by whatever the vendors are selling.
Before you talk to anyone, write down three to five decision questions your team struggles to answer. Where is risk building in the portfolio? Which projects are off track and why? What obligations are buried in expiring leases? How long does it take to produce a board-ready report?
If a vendor cannot show exactly how their product answers one of those questions with your data, in your environment, the conversation is not worth continuing.
2. Can You See Where the Answer Came From?
This is the question most vendors hope you do not ask.
Ask them to generate a specific answer. Then ask them to show every data source used, every document referenced, and the logic steps applied. Not a confidence score. Not "based on your lease data." The actual trail.
CRE decisions get reviewed by finance, legal, audit, and executive leadership. If you cannot explain how a number was produced, you cannot use it — no matter how impressive the demo looked.
If the vendor cannot trace an answer back to its sources in front of you, that is your answer.
3. What Happens When the AI Is Wrong?
Every AI system will produce incorrect or misleading output at some point. This is not a flaw in one vendor. It is a characteristic of the technology.
The question is not whether it will happen. The question is what happens when it does. Does the system flag uncertainty? Can a user see where reasoning broke down? Is there a deterministic layer underneath that catches errors before they influence a decision?
If a vendor tells you their system does not hallucinate, end the meeting. They either do not understand their own product, or they are not being honest with you. Neither is acceptable when your data includes lease liabilities, capital commitments, and compliance requirements.
4. Does This Require Replacing Systems That Already Work?
Be very careful here.
Some vendors position AI as the reason to consolidate your entire technology stack. Move your lease data here. Migrate your project data there. Bring everything into one platform, and then AI will work.
You have heard this before. It was called IWMS consolidation. Many organizations spent years and millions attempting it — and still ended up with fragmented data and manual workarounds.
AI should work with the systems you already have.
If the first step in implementation is a large-scale migration or platform replacement, you are not buying AI. You are buying a systems integration project with AI on the label.
5. Who Built This — and Do They Understand CRE?
A surprising number of AI tools marketed to corporate real estate were not built for corporate real estate. They are general-purpose products with a real estate skin on top.
Ask who designed the product. Ask if they have worked inside a corporate real estate organization. Ask them to explain how CRE data structures differ from other enterprise domains.
If they cannot speak fluently about multi-entity portfolio structures, cross-system lease and financial reconciliation, or how a facility manager's data needs differ from a portfolio strategist's, the product was not built for your world.
It was adapted for your budget.
6. What Does This Look Like in Six Months?
Demos are designed to impress. They show best-case scenarios with clean data and scripted flows.
Ask what the first 90 days actually look like. Ask when a real user on your team can ask a real question and trust the answer. Ask what has to happen on your side to make it work.
Most importantly, ask whether you can pilot the product with your own data before making a commitment. A vendor who is confident in their product will let you test it against a real scenario in your environment. A vendor who needs you to sign before you can see it working with your data is asking you to take their word for it.
The best way to predict what something looks like in six months is to see what it can do in two weeks.
7. How Much of This Is AI — and How Much Is Just Software?
Some of what vendors call AI is simply good software engineering: dashboards, alerts, integrations, reporting. Those are valuable. But they are not AI, and they should not carry AI pricing or AI risk.
Ask the vendor to clearly define where AI is used, where deterministic logic is used, and how the boundary between the two is managed.
The strongest implementations use AI selectively — where it adds leverage — and keep high-risk calculations auditable and rule-based. If a vendor cannot articulate that boundary, they may not know where it is themselves.
8. Are You Ready for This Internally?
AI does not fix bad data. It amplifies it.
If your lease records are inconsistent, project data lives in spreadsheets, and teams do not trust the numbers in existing systems, AI will not solve that.
You do not need perfect data. But you need clarity about where it lives, what state it is in, and what it would take to connect it.
If that clarity does not exist, an honest data and architecture assessment may be a better first investment than an AI product.
Closing Thought
The pressure to adopt AI is real. The risk of adopting it poorly is also real.
The vendors worth your time will tell you what their product cannot do, show you exactly how it works, and be honest about what it takes to get value.
The ones who promise transformation without specifics, who cannot explain their outputs, or who require you to rebuild your technology stack before delivering results are not solving your problem. They are creating a new one.
Take your time. Ask hard questions. Trust the people who answer clearly.
If you are evaluating AI for your CRE organization and want a neutral, decision-focused scorecard based on the questions above, I am happy to share one. Reach out and I will send it over.