AI use cases › Manufacturing & Industrials

Unilever: Machine-learning modelling to cut ingredient experiments

Computer models that predict lab results so scientists run fewer physical experiments when testing new ingredients.

How it works

Scientists train a model on past results, and it estimates outcomes instead of running each test; it also studies skin bacteria data.

What it delivers

It replaces some of the roughly 1,000 experiments once needed per new ingredient; no total savings figure was given.

Modelled business case

Figures are a model built from published benchmarks for this type of AI, not the organisation's own numbers. Change any assumption in the app.

Source: AI, machine learning and data behind Unilever's latest launches

Create your free Intlectify account to bookmark use cases, model the business case on your own numbers and start the free AI course.