Why corporate real estate needs an ontology, not another dashboard

Most real estate teams we meet already have dashboards. Occupancy in one, lease events in another, headcount in a third, and a spreadsheet that tries to hold it all together before every quarterly review. The dashboards are not the problem. The problem is that each one only knows what its source system knows.

A portfolio decision – consolidate two floors, renew early, open a new market – never lives inside one system. It lives in the relationships between them: which teams sit on which floors, which floors sit under which leases, which leases carry which options and costs. That is exactly the layer a dashboard cannot see.

What an ontology actually is

An ontology is a shared model of the things your portfolio is made of – buildings, floors, leases, teams, badges, sensors – and the relationships between them. It is not a data warehouse and it is not another integration. It is the agreed definition of what a “seat”, an “occupied floor”, or a “lease expiry” means across every source, applied consistently every time the data lands.

• Entities are the nouns: sites, buildings, floors, leases, teams, people.
• Relationships connect them: a team occupies a floor, a floor sits under a lease, a lease has a landlord.
• Rules keep it honest: a floor cannot be more than 100% assigned, a lease cannot expire before it starts.

Once that model exists, every question becomes a traversal rather than a project. “Which leases expiring in the next 18 months cover floors below 40% utilization?” stops being a two-week analysis and becomes a query anyone can ask.

Why this matters more with AI

Large language models are extraordinary at language and unreliable at arithmetic across disconnected tables. Point an agent at six systems with six definitions of “occupancy” and it will produce a confident, wrong answer. Point it at a governed ontology and every number it quotes is traceable to a defined entity, a source, and a timestamp.

The ontology is what makes an AI answer auditable. Without it, you are asking your team to trust a paragraph.

This is why we built the Trebellar ontology first and the agents second. Agents are only as good as the model they reason over.

Where to start

1. Pick the three questions your leadership asks every quarter.
2. List the systems each question touches. If the answer is more than one, you need a shared model, not a report.
3. Define the entities in those questions once, and hold every source to that definition.

You do not need to model the whole enterprise on day one. You need a model that is right about the things you decide on most often, and a platform that keeps it right as the data changes underneath it.