As sales are mostly driven through resellers and distributors, world’s most valuable beer brand HEINEKEN® did not always have sufficient insights into outlet sell out data. Until now.
By building a smart data-sharing platform in the native stack of Microsoft Azure, our company enabled HEINEKEN EMEA Export to make data-driven decisions on stock and demand planning data derived from our distributors and partners combined and mapped with our internal data. With a little help from Artificial Intelligence (AI). By transforming different data files shared by multiple outlets seamlessly into a readable standard format, HEINEKEN EMEA Export can get a firm grip on sales data and channels that used to be a black box before.
- 70%Time saving
- 10Connected outlets
- 98%Probability
Smart data export to PowerBI
Every HEINEKEN reseller sells a different set of HEINEKEN products to different end customers. And although the information and insights into sales volume are of great value to HEINEKEN, the brewer did not have access to them.
Martijn Otten, IT Manager EE&GDF & Heineken Experience, worked closely together with Hypersolid, and they found an innovative solution. We built a smart data-sharing platform in the native stack of Microsoft Azure. Different types of data files from distributor ERP systems are uploaded via email and automatically transformed to standard HEINEKEN output tables in PowerBI.
Automation and efficiency with OpenAI
Although many developers and companies make use of OpenAI, not all applications are using the full potential of it yet. We, however, recognized the golden opportunities of the technology. We developed a tool with an automated product linking engine, using the power of AI to automatically create product links.
Using prompt engineering in OpenAI, the tool is provided with queries that measure the probability that a specific product matches another product in the database. This probability can be as high as 99.5 per cent, automating a significant amount of work related to manually mapping products. Daniel de Vos, Head of Data & AI at Hypersolid explains: “As each customer has its own unique database structure and product portfolio, OpenAI recognizes the need for flexibility and understands that solutions are not always straightforward.”
Continuously learning from human validation
Although the tool can automate most of the matching process, a phase of manual validation is still required. This validation takes place in a Power App in Azure, where users can easily indicate whether the automatic matching is correct or not. The results of these validations are stored, allowing the tool to automatically match future appearances of the same product. This way, the model is continuously trained by human validations, resulting in a significant reduction of manual work and a more efficient product-matching process within HEINEKEN.
Seamless data integration, handled with the greatest of care
Data integration is seamless thanks to the simple process of sending uploads via email, which are automatically incorporated into the workflow. The uploading process was purposely not built in AI, allowing customers to retain control and diligence over their data.
The cleaning takes place before product linkage, showing an understanding of the diversity of outlets that may not follow the standard structure. The ITL process is fully adapted to different capabilities, with the dynamic model ensuring that no anomaly slips through the cracks. Regardless of the variation in columns, time zones, and the level of maturity of the files submitted.
Data cleaning and understanding the complexity of various files with different maturity levels have led to a dynamic model that, with the help of AI, significantly reduces the complexity of product mapping. The aim is not only to reduce duplication and errors, but also to gain valuable insights into what sells and what does not, enabling strategic decision-making and targeted deployment.
Up-to-date information available at marketeers’ fingertips
The scalable solution includes three different environments - development, acceptance, and production - allowing for a structured and controlled implementation. An important aspect of its success is the close collaboration with the customer of HEINEKEN EMEA Export, identifying the relevant use cases for specific customers.
HEINEKEN EMEA Export is now able to also share the data customers have provided back to them; helping them to gain better insight in their own situation. Within just a few months, the tool has already successfully connected 4 HEINEKEN outlets, planning on reaching 10 this by the end of 2024, enabling a seamless exchange of information. The latest data is now always available, which is possible thanks to the application of advanced AI technology. Helping HEINEKEN enhancing efficiency, reducing manual labor, and enabling more strategic decision-making. Cheers to that!
The makings of success
Keeping it simple
The beauty of this solution lies in its simplicity.
Seamless data integration
Data integration is seamless, thanks to the simple process of sending uploads via email, which are automatically incorporated into the workflow.
Control in your own hands
The uploading process was purposely not built in AI, allowing customers to retain control and diligence over their data.
Close collaboration
An important aspect of our success is the close collaboration with the customer, identifying the relevant use cases for specific customers.
Solid change starts here
Contact us and let's get started.
