Consumer goods / Household products

Unilever: Scaling digital twins of factory production equipment

Company
Unilever
Country
Multiple countries
Adoption stage
In operation
Source published
Date basis
The date the source was published. It can differ from the date adoption started.
How the source was checked
Read the full source text

The work problem

Unilever said it is using digital twins to identify issues sooner and simulate scenarios faster across its global factory network. Each site had a different problem. A deodorant stick line in the United States faced process flow restrictions, a mayonnaise factory in Poland had to reduce viscosity variation and waste, and sites in India and Vietnam needed to reduce soap quality defects, thermal energy use in powder detergent production and the use of premium ingredients in liquid detergent.

Technology and data

In a partnership with Accenture lasting several years, Unilever builds digital twins, which are virtual models of equipment and production lines fed with live shop floor data. The twins use AI insights and agentic capabilities. The company said Accenture AI agents predict maintenance needs and can progressively take on certain adjustments automatically, with human oversight. At Haldia in India, an energy twin optimises fan speeds, temperature setpoints and moisture controls. On top of the twins already in use, Unilever plans to build more than 40 new digital twins over the next 18 months and roll the approach out globally.

Results

Unilever reported results by site. At Raeford, North Carolina, a twin used in making Dove, Degree and Axe deodorant sticks predicts 95% of process flow restrictions, and the site cut waste by 20% and raised capacity by 10%. At Poznan, Poland, minor stoppages on the mayonnaise line fell by up to 20% and waste by nearly 30%. At Gandhidham, India, Dove soap quality defects fell by 30% over four years, measured in distribution centres before delivery. At Cu Chi, Vietnam, the site saved 1 to 2% of premium ingredients in OMO liquid detergent. Haldia reported a tangible reduction in thermal energy consumption over two years without giving a figure.

Limits and open questions

Apart from the four years at Gandhidham, the site figures come without a measurement period or baseline, and Haldia gives no figure. The more than 40 new twins are a plan, and the adoption month of each twin is not disclosed. The page shows 16 June 2026 while its metadata gives 15 June, so the displayed date is used. No adoption month is given, so the record is ordered by the publication date.

Sources

Compiled from public sources. These are not results from ATF Works customers.

Read original (opens in a new tab)