Predicting Future Customer Behavior in CPG

The primary challenge was to utilize predictive analytics and AI to create a "Suggested Order" system for each client (store). This system was intended to guide the telesales team by providing recommended products, thereby enhancing sales efficiency and personalizing customer interactions.

AB InBev, a leading beverage company, sells through call centers to its 450,000 small and medium-sized customers (bars, restaurants, nightclubs, retail stores), comprising 85% of its inventory. A key priority for AB InBev was to boost the sales of related products across its distributors and to increase market penetration of two brands, Budweiser and Stella Artois.

PulseUp founder and a team of highly skilled Data scientists led this project in his previous company EPICA, in collaboration with the ABI team, processed historical transaction data from distributors using the EPICA® "Tailored Suggested Product Model". This model was pivotal in generating recommendations for new SKUs to suggest to a test group versus a control group. The insights gained from this analysis proved that data was an invaluable tool for suggesting new products. The solution integrated first-party sales behavior data with third-party data, including local weather conditions, local football matches, and public holidays that could dynamically affect beer consumption in different geographical areas of the country.

The final output was a client-specific suggested order, which the call center then used, tailoring the script based on the AI-generated insights.

The implementation was highly successful, with 75% of customers accepting the AI-generated suggested order as is. This AI-driven recommendation system accelerated sales growth by 3.2% in the first seven months for the product references that required increased market penetration within the portfolio.

This implementation exemplifies how advanced machine learning algorithms can be effectively applied in a commercial setting to predict future customer behavior and optimize sales strategies. By integrating a sophisticated blend of historical sales data and dynamic external factors such as local events and weather, the solution provided a tailored, data-driven approach to sales. 

The ability of the AI system to learn from past transactions and adapt to varying external conditions demonstrates its robustness and adaptability. Furthermore, the success of this project underscores the importance of targeted, AI-powered interventions in enhancing customer engagement and achieving significant gains in sales and market penetration. 

This case study not only highlights the practical applications of AI in streamlining operations but also its potential to drive substantial business growth and customer satisfaction.