Digital twins move from buzzword to enterprise manufacturing engine
Digital twins in manufacturing are real-time, virtual models of equipment, production lines, or plants that continuously ingest operational data to simulate behavior, predict outcomes, and guide decisions so enterprises can improve efficiency, quality, and reliability at scale. For years, many manufacturers kept digital twins in pilot mode, tied to one machine or one line. The current shift is toward digital twins as a core layer of enterprise manufacturing optimization, integrated with ERP and industrial data platforms. This evolution turns real-time digital models into a shared decision engine for operations, maintenance, and supply chain teams. As Unilever and RUSAL scale their programs, digital twins are no longer side experiments; they are becoming repeatable blueprints that connect AI production simulation with business systems, helping manufacturing leaders standardize best practices and respond faster to demand and process variation.
Unilever: AI-powered digital twins on an SAP and BTP backbone
Unilever is rolling out more than 40 AI-enabled digital twins across its global manufacturing network over the next 18 months, working with Accenture on SAP S/4HANA Cloud and SAP Business Technology Platform foundations. The twins connect live shop-floor data with AI production simulation so teams can spot issues earlier and run rapid what-if scenarios. In Raeford, a twin that supports Dove, Degree, and Axe deodorant production predicts 95% of process-flow restrictions, cutting waste by 20% and increasing capacity by 10%. At Gandhidham, another twin reduced quality defects for Dove soap bars by 30% over four years through real-time control recommendations. This program builds on Unilever’s move to a clean-core SAP environment, where extensions run on BTP rather than in heavily customized ERP instances, allowing digital twins to feed predictions back into execution, maintenance, and supply processes.
RUSAL: 100+ digital twins as a refinery-wide optimization platform
RUSAL is building a corporate platform to manage the full lifecycle of digital twins across its alumina refineries, expanding an existing program already active at sites including UAZ, BAZ, AGK, Friguia, and Ewarton. Between 2026 and 2028, the company expects the platform to support more than 100 digital twins, turning advanced real-time digital models into a standard way to stabilize processes, improve operational efficiency, and strengthen decision-making. Digital twins sit at the core of RUSAL’s Model Predictive Control framework, which uses live process data to determine optimal operating parameters and provide timely recommendations to operators. The unified platform is designed to raise equipment reliability and production safety while standardizing best practices across facilities. By scaling from individual deployments to a coordinated portfolio of twins, RUSAL aims to unlock more value from real-time optimization and advanced analytics across its alumina operations.
From point solutions to integrated manufacturing intelligence platforms
The common thread in Unilever and RUSAL’s strategies is a move from isolated digital twin pilots to integrated manufacturing intelligence platforms. Unilever treats twins as part of a wider industrial AI layer that sits on SAP S/4HANA and BTP, with Accenture as the systems integrator connecting plant data, analytics, and AI agents that can recommend or automate adjustments with human oversight. RUSAL, meanwhile, is standardizing twin development and operation on a corporate platform so models, control strategies, and analytics can be reused and scaled across refineries. In both cases, real-time digital models are tied to core business and production systems instead of existing as standalone tools. That shift turns digital twins manufacturing initiatives into enterprise assets: they inform maintenance planning, quality management, and capacity decisions, and they give leaders consistent, data-backed views of how process changes will affect output and risk.
What enterprise leaders should do next with digital twins
For executives, the lesson is that digital twins pay off when they are part of an end-to-end operating model, not a single factory experiment. First, clean up and standardize core systems so twins can draw on reliable, reconciled operational data; Unilever’s SAP consolidation and move of extensions to BTP was a key enabler. Second, plan twins as reusable patterns across sites, as RUSAL is doing with its corporate platform, rather than one-off builds. Third, choose partners that can connect shop-floor technologies with enterprise applications; consultancies like Accenture bring integration skills and repeatable templates that internal teams may lack. Finally, treat AI production simulation and real-time digital models as shared tools for operations, quality, and supply chain teams, backed by clear governance. That is how digital twins evolve from promising technology into a scalable engine for enterprise manufacturing optimization.
