Digital twins move from concept to factory standard
AI-powered digital twins manufacturing programs are industrial initiatives where virtual models of equipment and production lines use live factory data to simulate behavior, predict issues, and guide real-time production optimization decisions at scale. That definition is no longer theory for Unilever. Working with Accenture, the consumer goods group plans more than 40 additional digital twins in the next 18 months, on top of sites already running the technology. Each twin mirrors a specific line or process, turning sensor streams and enterprise data into a continuously updated simulation. For manufacturing teams, this means a shift from reactive troubleshooting to proactive, data-led decisions. Instead of waiting for a bottleneck or quality deviation to appear on the shop floor, engineers can test settings and interventions inside the model first, then apply the winning scenario in production. This is where AI industrial operations start to look like day-to-day plant practice, not a pilot demo.
From deodorants to detergents: measurable production gains
Unilever’s digital twins are anchored in clear production optimization goals: less waste, higher capacity, and more stable quality. The Raeford plant, producing Dove, Degree, and Axe deodorants, shows what this can look like in practice. According to Unilever, the deodorant twin “predicted 95% of process flow restrictions in deodorant stick manufacturing, leading to a 20% reduction in waste and a 10% uplift in capacity.” Elsewhere, a twin at Poznan stabilizes mayonnaise viscosity and has cut minor stoppages by up to 20% while reducing waste by nearly 30%. In Gandhidham, a quality-focused twin has reduced Dove soap defects by 30% over four years. Energy and ingredient twins at other sites fine-tune fan speeds, temperature, moisture, and dosing to save 1–2% in premium raw materials. None of these wins are flashy alone, but together they rewrite the baseline of Unilever manufacturing AI performance.
Why SAP S/4HANA and BTP sit at the core
The digital twin story at Unilever rests on a long-running SAP S/4HANA deployment rather than isolated data lakes. Unilever has collapsed about 200 local ERPs into four regional SAP landscapes and now runs SAP S/4HANA Cloud under RISE with SAP, with a deliberate move to a clean core. Innovation, including AI and advanced analytics, shifts to SAP Business Technology Platform (BTP). That matters because digital twins are only as reliable as the data that feeds them. SAP provides consistent, structured information on orders, materials, quality, maintenance, and costs, while BTP hosts simulation models and AI services. Accenture and SAP have co-innovated industrial twins that sit on BTP and connect to SAP Digital Manufacturing, treating the twin as the bridge between cloud execution and physical machines. In Unilever’s case, this SAP backbone turns each twin from a one-off model into an enterprise asset that can be governed, scaled, and reused across plants.
From pilots to a repeatable blueprint for AI industrial operations
Unilever and Accenture are making a multi-year commitment to AI industrial operations rather than running isolated proofs of concept. More than 40 new twins are planned in 18 months, with the explicit aim of creating a repeatable blueprint for global rollout. This means common architecture on SAP S/4HANA and BTP, shared AI patterns, and a catalog of use cases that local teams can adapt instead of reinventing. Factory staff gain decision-support tools that highlight probable issues, show predicted outcomes of different settings, and recommend actions in near real time. At the same time, Accenture brings cloud, AI agents, and analytics skills while relying on Unilever’s complex production lines as the test bed. For Accenture, these projects are proof points that industrial AI can become durable, scaled work rather than one-off advisory deals. For Unilever, they mark a shift toward factories where simulation is a default part of every major run.
What this means for the next wave of digital twins manufacturing
Unilever’s program signals a new phase for digital twins manufacturing strategies: industrial AI is moving from experimental slides to operating discipline. By tying twins to SAP S/4HANA deployment standards and BTP-based innovation, Unilever avoids data silos and builds a shared foundation that other plants can tap. The variety of use cases—throughput, quality, energy, ingredient savings—shows that twins are not a single application but a method for continuous production optimization. As more sites adopt similar patterns, AI-enabled simulations will likely become as routine as line changeover checklists. The implication for manufacturers watching this move is clear: the value lies less in branding a solution as “AI-powered” and more in building the plumbing so predictions, simulations, and recommendations can feed back into everyday planning and shop-floor control, at scale and with measurable impact.
