81.6%
accuracy at identifying payments
Results
40.3%
of costs saved while still collecting all the cash in the dataset
Benefit
The solution removed costs from the operational process without missing any of the collections previously received.
Context
For this collections call centre, there were more accounts to process than capacity available. However, many of the accounts contacted were not converting to collections – just 6% of accounts ended up paying as part of the campaign.
Methodology
Emerge built an AI solution to predict at the beginning of each month which accounts were likely to pay on the back of collections interventions. The list was then rank ordered according to the collections score determined by the AI solution. The collections interventions were prioritised in this rank order. Initially all the accounts were called to assess the model’s accuracy.