Sales recommendation


A large Bank from Asia sells financial products to Small-Medium-Enterprise (SME) companies. They use to characterize and segment customers by BI tools. However, this approach demonstrated limitations and sales force does not have enough indications to really impact in their everyday selling process.

The Bank is convinced that a modern approach based on larger segmentation and deeper personalization is possible thanks to their large data set of customers’ characteristics and historical purchases. Probably, this information hide purchase patterns that could be used to predict behaviours and anticipate potential sales.

The Bank contact Terminus7 to evaluate potential approaches to exploit this information through advanced Machine Learning techniques.


To sell to SME companies with better and deeper segmentation


Terminus7 worked with the Bank team to define the final solution useful for the sales force. Giving the Bank is facing a professional B2B sale process, which will require always a later one-to-one contact, we finally decided to look for the following solution. The sellers will have a list of customers with the probability of buying each product during the following 30 days. The list is ordered under a probability criteria.  

We train the T7 machine with customers characteristics and past purchase patterns. We expect that similar companies with similar past purchase patterns would have the same financial needs. Therefore, we can recommend products based on these similarities. This is a typical e-commerce personalization approach close to Collaborative Filtering techniques.


Recommendation based on purchase pattern

Do you want integrate AI to your business?