As Unilever prepares to extend its digital twin technology across approximately 40 manufacturing sites worldwide, the company is placing business performance at the heart of its strategy. Its experience so far has highlighted the importance of setting clear objectives, defining measurable outcomes and monitoring progress from the outset.
Vicky Cuthbert, chief product supply chain officer for Unilever’s personal care business, acknowledged that the company did not always establish sufficiently precise measures of success during the early stages of implementation.
“When we started this work, we probably weren’t clear enough about ensuring that we were delivering benefits that could be tangibly measured in the results,” she told Supply Chain Dive.
Over the next 18 months, Unilever will work with Accenture, a consulting and technology services provider, to introduce artificial intelligence-enabled digital twins at around 40 locations. The initiative builds on deployments already carried out at five manufacturing facilities: one each in the United States, Poland and Vietnam, and two in India.
Digital twins create virtual representations of physical equipment and production lines. By collecting data from operational systems, they allow manufacturers to monitor performance, anticipate potential problems and assess how effectively production processes are running.
For Cuthbert, the experience gained from the initial deployments has reinforced a fundamental principle: companies must identify the problems they want to solve before introducing the technology, then establish how improvements will be measured.
“Spending time upfront to make sure that’s very clear for all involved stakeholders is incredibly important,” she said.
Preparing staff for digital transformation
And it is not only about the technical implementation; Unilever is also thinking about the organizational and cultural change that comes with digital twins. The technology offers greater visibility into what is happening in the manufacturing operation, but it also impacts the employees who are responsible for production, product quality, and other activities in the plant.
“Managers must not forget this human side,” warned Cuthbert. Teams need the right skills to understand and apply digital insights in their day-to-day work.
To help make this change happen Unilever has created DigiEdge, a training scheme to boost employees’ digital capabilities in planning and factory operations. The scheme uses gamification to make learning more engaging and to help staff develop the skills they need to work effectively with digital technologies.
“Don’t underestimate the overall cultural and organizational impact this has, and how you need to take the organization with you,” Cuthbert said.
This approach is a reflection of the company’s belief that successful adoption is not solely about the technology itself but about employees’ ability to understand, trust and use it.
Predictive maintenance and operational enhancements
One of the most promising applications of Cuthbert is predictive maintenance. Digital twins can help predict when equipment is likely to need servicing, for example, by analyzing sensor data.
More reliable predictions could help factories plan maintenance better, limit production downtime and ensure the right spare parts are available when needed.
The five manufacturing sites already using the technology have seen improvements in waste reduction, production capacity, product quality and operating efficiency, Unilever said.
Digital twins have helped the company’s personal care factory in Raeford, North Carolina, cut waste by 20% and boost capacity by 10%.
In Poznań, Poland, the food manufacturing plant has been able to reduce minor production stoppages by as much as 20%. And a personal care facility in Gandhidham, India, reduced product defects by 30% in four years.
The technology has cut the cost of premium ingredients by 1% to 2% at the home care facility in Cu Chi, Vietnam.
These results demonstrate the level of operational improvements that digital twins can support when deployed in line with specific manufacturing challenges.
Balancing supervision by humans and autonomous operations
Unilever’s current uses of the digital twin can suggest corrective actions when incidents occur, but these actions still need human approval before they can be put into practice.
But the technology also includes agents that can autonomously perform in certain low-risk processes. This paves the way for increasing levels of automation over time as the company builds confidence in how people and AI-based systems can collaborate.
As Cuthbert observed, the ratio of human involvement to autonomous activity is constantly changing in relation to the area and process of production.
That literally is tweaking depending on the process area week on week at the moment as we learn about how the human and the agent can coexist,” she said.
Unilever can develop its AI capabilities with the right oversight from humans and with autonomy that suits the needs of its operations by going slow and steady.
Selecting factories where digital twins can deliver the greatest value
With several hundred factories across its global network, Unilever must determine where the technology can make the most significant difference.
Site selection therefore depends on the potential to improve particular operations and establish standardized processes that can subsequently be replicated elsewhere. The objective is not simply to introduce digital twins across as many facilities as possible, but to identify applications that deliver demonstrable benefits and can be scaled effectively.
Cuthbert said success should be directly connected to business priorities, whether through improved service performance, greater agility and responsiveness, or more traditional measures of productivity and growth.
The company’s preference is generally to begin with technically advanced factories. Such facilities are more likely to have the necessary technology infrastructure, cybersecurity capabilities and production-floor sensors to support digital twin applications.
A minimum level of data collection capability is essential, Cuthbert explained, because the technology must be able to gather information from the relevant stages of the process being modelled.
By combining these technical requirements with clear performance targets, employee development and repeatable implementation methods, Unilever aims to turn its initial digital twin experience into a broader manufacturing strategy.
The expansion over the next 18 months will test whether the gains achieved at the first five facilities can be reproduced across a much larger share of its global production network.
