Predictive cross-sell for a top-10 Ukrainian retail bank
The bank's sales function was reactive. Client managers had no systematic way to know which customer to approach, when, or with what product — though the transaction data to answer that already existed, unused.
We built a predictive layer on the bank's own transaction data. The model reads spending patterns to identify where a genuine cross-sell opportunity exists, then an automated flow delivers it as a ready-to-action task straight to the responsible client manager.
Our role: diagnosis of the sales process, prediction logic, target operating model and automated task routing. Our technology partner built and deployed the technical layer.