What a 90-day connected property pilot should prove

Connected property intelligence is a big claim, and big claims invite big programmes. That is the wrong way to start. The right way is a bounded pilot that proves the model on one decision with one cohort.

A good pilot is scoped tightly. One cohort of properties. One measurable decision, such as buy-to-let case completeness or valuation routing. A defined set of sources. And a short list of participants who will act on the outputs.

Success criteria should be agreed before the first record is matched. Match quality. Evidence completeness. Decision impact. Auditability. If the pilot cannot show movement on these, scaling would only scale the problem.

The pilot should also prove the governance model, not just the data. Can every field show its source and freshness. Can every recommendation show its logic. Can every override be recorded with its outcome. These are the properties that make the intelligence defensible later.

Ninety days is enough time to prove the join, the decision and the controls, and short enough to keep the scope honest. What it should not try to do is boil the ocean. Servicing, retention and portfolio intelligence come after the first decision is proven.

Start with one decision. Prove the value. Scale the intelligence.