DATA & AI

“Every month, cluster leads, and business analysts produce detailed reports on the business performance based on syndicated data provided.

The data includes several metrics such as share, contribution, value sales, unit sales, volume sales, etc., for different levels of aggregation, such as channel, country, cluster, region; and also product aggregations, like brand, category, gfo, for different manufacturers in the last three years, for thousands of products.

Analysts then produce detailed reports (or narratives) for multiple combinations of the levels abovementioned, which is a very time-consuming task. Automating this task would leave more time for the analysts to interpret the data and focus on tasks of higher added value.

In collaboration with Genesis, we tested the usage of state-of-the art large language models (LLMs) provided by OpenAI to accelerate the production of business performance narratives, with the level of detail dependent on the user (country manager, brand manager, etc.), while also providing the user an efficient way to interrogate these data using natural language. For example, the user could ask “Write me a business report comparing our performance in Germany YoY” , or more directly “What are out top 10 brands in terms of growth?”. Using a mix of existing reports, surveys with business analysts, and the actual data (or a sample of it), Genesis were able to demonstrate that GPT could answer correctly, timely, and started delivering interesting insights.

In the end of this POC, the business analysts were very excited by the prospects.”

Isabel Rosa – Kenvue