One of the world’s largest international manufacturers and distributors of soft drinks and liquor, herein known as the “company”.
The company sources ingredients from suppliers all over the world, as well as supplying products and services to thousands of clients who then distribute the products to bars, shops, supermarkets, vending machines, and more.
Every month the company is confronted with a huge number of overdue buyers' invoices amounting to millions of US dollars, imposing a burden on the company's finances and operating systems.
The company wanted to turn to AI to find a way to predict which of its customers are likely NOT to pay their invoice on time, in order to take preemptive measures to increase collection.
However, the team did not have a dedicated data scientist, nor access to the data science team within the enterprise. Therefore, they needed a platform,which is quick and user-friendly.
The company conducted extensive market research and chose Pecan‘s predictive analytics platform to address this challenge. They did so by building two deep-learning-based models to:
The customer’s BI team used Pecan to build the two models within 10 days, compared with 4-6 months required by alternative solutions which consume huge and expensive resources, i.e. data scientists. Pecan’s solution produced much faster results, allowing the team to be within budget, and to quickly understand where they stand.
It is important to reiterate that all this was done without the use of a data scientist! The person who built the model, analyzed the results and then used them was a business analyst with no previous AI experience.
The predictions generated using Pecan were correct 94.7% of the time. The team working on invoicing and collection were simply stunned. Furthermore, using Pecan, the customer was able to predict 346% more overdue cases, that otherwise would have “surprised” the collection team. This drastically reduced uncertainty, allowing the customer's finance department to take preemptive measures and focus on the most relevant accounts.
Finally, using Pecan the customer was able to single out which factors contributed most to the invoice being overdue, for each specific customer. Knowing these contributing factors allowed Pecan's customer to design a personalized course of action to preempt the overdue payment, raising the likelihood of such action's success.
Pecan’s technology saved, and brought in, millions of dollars. All this was done fast, and much less expensive than other platforms or the “old-fashioned” method.