A recommender system that clusters tenants by performance profile and surfaces ranked candidates from matching clusters — replacing gut-feel leasing decisions with consistent, data-backed ones.
Retail leasing decisions are largely manual and relationship-driven, with limited use of data to match tenants to locations. The result is longer vacancy periods, inconsistent tenant mix, and missed revenue — particularly when expanding into new markets where the team doesn't have established relationships to draw on.
Leasing managers were making high-stakes placement decisions based on intuition and whoever happened to be in their network. There was no systematic way to evaluate fit or compare candidates, which meant deals took longer and outcomes varied widely.
We built a recommender engine that uses proprietary AI similarity analysis across tenant constituency, category, and local demographic data to surface ranked candidates for any given vacancy.
For the leasing team, the shift was from blank slate to ranked shortlist — with enough rationale surfaced alongside each recommendation that managers could act on it quickly, even in unfamiliar markets.
Reduced vacancy time and improved tenant mix by replacing manual leasing decisions with ranked, data-backed recommendations. Deal cycles shortened, location performance became more consistent, and the team could move confidently in new markets without relying on pre-existing relationships. The recommender is now a core part of their expansion playbook.