A new machine learning operations system is delivering smarter production and improved sustainability metrics – alongside a range of customer benefits
AI is revolutionising industries worldwide, not least by turning technical capability into operational impact. At the same time, according to business consultants McKinsey, when AI initiatives underperform it is not because of flawed models, but because organisations struggle to integrate them effectively into day-to-day operations.
To avoid this pitfall, Sappi Europe has established an innovative system of data driven practices across its production, planning and environmental objectives. All in order to streamline operations, reduce energy cost and improve performance on a day-to-day basis in a way that also benefits customers.
Now, working in partnership with Orange Business, the company has taken its approach to scaling the use of AI one step further by implementing an ‘MLOps’ (machine learning operations) framework – to accelerate AI adoption and increase efficiency and sustainability right across the company’s business operations.
MLOps, machine learning and AI
Machine learning (ML) is a type of artificial intelligence that works by analysing data patterns in order to create models that automate tasks and enhance an organisation’s operations. Creating such models is a difficult and often laborious process, requiring the coordination of disparate teams, much configuration and significant computational power.
This is where MLOps comes in – as a way to manage machine learning to make it more efficient and reliable. MLOps sets up repeatable steps to get data in, build and test models, and then put them into daytoday use. When data or results change, MLOps triggers updates to the ML models.
Sappi, which ran its MLOps development project in conjunction with Orange Business from October 2024 to June 2025, set out to tackle problems including fragmented data sources, a reliance on manual control and a lack of real-time operational insights. A comprehensive MLOps programme was designed to overcome these challenges and improve automation, process optimisation and sustainability.
MLOps, energy use and sustainability
Sappi focused particularly on energy analytics – where fragmented data sources and manual processes have traditionally made it difficult to optimise energy use, reduce waste or get up-to-date visibility of resource consumption.
Now, by deploying the AI-driven MLOps platform, Sappi can access standardised, automated data and analytics across sites – so facilitating reduced energy consumption, minimised waste and more sustainable production.
For example, the company’s Maastricht mill is distinctive in that it not only draws electricity from, but also sometimes supplies electricity to, the Dutch grid. MLOps was introduced into this process to optimise energy utility management by automating end-to-end data processes and increasing the accuracy of cross-dependent operational and tactical processes.
All of this is important to improve the overall sustainability of Sappi Europe operations. By developing and integrating data and analytics into everyday operations, the MLOps system can generate new insights and increase flexibility in the use of energy assets at the same time as enhancing the efficiency of both utilities and energy-intensive processes.
MLOps, efficiency and customers
For customers, Sappi’s introduction of an innovative MLOps system has tangible value. With analytics embedded in production decisions, customers can benefit in a number of ways:
- Commercially, better forecasting and supply chain analytics improve Sappi’s view of demand and inventory – supporting clearer communication on delivery windows and reducing operational risk for customers
- Process and energy efficiencies support cost discipline – a more controlled cost base helps manage volatility over time
- Sustainability gains carry through the supply chain – customers seeking lower-footprint materials can credibly point to a supplier investing in energy efficiency and smarter resource management
- There are also brand signalling benefits – partnering with a manufacturer that treats AI as an operating capability can strengthen a customer’s own narratives about innovation and resilience
MLOps, scalability and the long term
Sappi also wanted the new MLOPs framework to be scalable across the whole business, including sales forecasting and supply chain management. The new system is configured to support the long-term deployment of AI-supported decision making throughout the business. It sets clear rules for who can access what, runs tasks on a schedule or when events happen, and keeps an eye on results so models stay accurate over time.
By centralising this setup, models are updated consistently and maintained uniformly across a range of different sites, which makes scaling far easier. That means the system can refresh and tune models automatically, so shortening the path from an idea to something impactful on the factory floor. It also connects teams across mills under shared practices.
Sappi’s MLOps strategy is today building a durable, future-ready AI ecosystem – one that benefits multiple business functions and customers alike.