From reports to decisions: property data and decision intelligence

Property companies tend to produce a lot of reports: occupancy, arrears, maintenance cost, renewals. Yet managers often still say they cannot get a clear answer. The problem is rarely a lack of data. It is a lack of connection between the data and the decisions people have to make.

Reports describe, decisions need more

A report tells you what happened. A decision needs three more things:

• A question, such as "Which buildings are at risk of missing renewal targets?"
• A threshold, such as "more than 20 percent of leases expiring in 90 days and not yet contacted".
• An owner and an action, such as "the leasing lead calls these tenants this week".

Without them, a dashboard becomes something people look at and then carry on as before.

Start with a small set of numbers

Pick ten or fewer numbers that matter to your business. For a property manager these might be occupancy, days to re-let, arrears over 30 days, maintenance cost per unit, time to close a request and renewal rate. For each one, agree on:

• The exact definition (what counts as "occupied"?).
• The single source it comes from.
• Who is responsible for it.

Many arguments in management meetings are really arguments about definitions.

One source of truth

When two systems report different occupancy, trust drops. Decide which system is the source for each number, and fix the others to read from it. It is dull work and it pays back every month.

Put the action next to the number

Instead of a chart of overdue requests, show the list of overdue requests, who holds them and how late they are. The moment a number leads directly to a name and a next step, it starts changing behaviour.

Where AI helps

AI can help people ask questions of their data in plain language, spot unusual changes and write a first summary of what moved this month. It cannot repair unclear definitions or messy records. That groundwork comes first.

A good first project

Pick one decision your managers make every month, for example which units to prioritise for maintenance. Define the numbers behind it, clean the data for those numbers only, and build a simple view that shows the list and the suggested action. Use it for three months, see what changes, then extend.