Customer activity was inconsistent throughout the network of 3,000 salespeople worldwide, resulting in an inefficient pipeline that was not fed with real customer signals, but was still being manually managed. The forecasting process was less driven by data than by subjective judgment.

Remove manual data input first and then introduce AI. Behavioural data collected automatically ensured accurate machine learning based on customer interaction, not on the reporting style of salespeople.

Salesforce Sales Cloud with Einstein Activity Capture for automatic email and calendar sync. Propensity models tailored to segments instead of a general one, and explainable AI scores in the opportunity records. Redesigned management dashboards with scored pipelines and structured sales enablement.