Sales Calls Classification

How Business Intelligence AI technology can help detect wrong sales and reduce the commissions paid.
sales report

The Challenge

Our client paid sales commissions to their teleshopping provider based on the monthly sales report sent by the teleshopping provider. The client did not have the internal resources to review the whole sales report, so they used to review just a very small percentage of sales. The manual review showed the report had some wrong sales. So, they were paying commissions for a small number of sales that had not been made.
The client did not know the incorrect extra commissions they were paying but estimated that from 2 to 3% of the reported sales were wrong.

Sigma’s classifier product has the capacity of learning from the historic and current data to detect wrong sales.

The Solution

Sigma’s team prepared a subset of the historic database, so it could be used by the classifier to learn how to detect wrong sales. The team filtered, cleansed and annotated it properly and helped to change the data collection process, so that current data could be used to improve the performance of the classifier.
Once the data was ready, Sigma’s classifier learnt from the examples and started to work on the detection of wrong sales.

The Outcome

Sigma’s classifier reviewed the calls and detected that, on average, 5% of the reported sales calls had not ended up in sales, which reduced the commissions paid by our client and provided a method to verify the sales report. The ROI was one month.