How AI is transforming the factory floor

AI is rapidly transforming the factory floor, accelerating the shift toward smarter, more efficient operations. From predictive maintenance to quality control, AI-powered systems are optimizing production lines, driving cost savings and reducing emissions.

The World Economic Forum’s Global Lighthouse Network, which recently welcomed 22 new members, showcases how digital transformation and cutting-edge technologies like machine learning and digital twins are driving next-generation operational excellence.

These advanced sites are not only boosting productivity but also setting new standards for sustainability and workforce development, enhancing human-machine collaboration and unlocking new levels of innovation.

Here, six leaders share insights on how their Lighthouse sites are using AI to drive industry forward.

Through the integration of AI-driven innovations, we have not only optimized our manufacturing processes and design but also empowered our workforce.

Key AI applications include a smart machine learning powered control system that adjusts parameters in real time, reducing scrap and preventing defects in sheet metal forming, resulting in a 12.5% material cost savings.

A decision tree-based model prevents clinching failures from variations in sheet thickness, cutting defect rates by 66%. A closed-loop valve gate control using convolutional neural network algorithms optimizes plastic injection, analyzing over 150K data points and improving cycle time by 18%.

Advanced machine learning algorithms in cleaning cycle design reduced time to market by 46% and achieved 99% optimization in cleaning performance.

Training programmes covering basic AI principles to advanced machine learning applications have led to 3,160 training hours completed in six months. A global automation programme guides factory-scale adoption and use-case sharing, governed by central and local digital transformation offices, with plans to establish a lighthouse factory for each product group.

Today, AstraZeneca is using AI to revolutionize how we develop, make and supply medicines. In drug development, predictive modelling helps optimize the physical and chemical properties of our active pharmaceutical ingredients and predict the performance of formulated products during manufacturing.

Generative AI (GenAI), machine learning and large language models are already helping reduce development lead times by 50% and reduce the use of active pharmaceutical ingredients in experiments by 75%.

In manufacturing, AI-powered process digital twins optimize the conditions for yield and productivity while reducing the use of raw materials and minimizing tech transfer requirements. The digital twins simulate the relationship between drug substance properties, process conditions and product quality to optimize operating conditions.

Combined with continuous manufacturing, we’ve reduced manufacturing lead times from weeks to hours. And with GenAI-human synergy, we are accelerating regulatory filings, cutting the time to create some documents by more than 70%.

As we head towards a net zero carbon footprint, AI-powered tools will also help us achieve this by allowing us to interrogate life-cycle data for our medicines and providing visibility of deviations and emissions “hotspots” so that we can mitigate these across the whole supply chain.

At Jubilant Ingrevia, we’ve embraced AI and machine learning across all production stages to boost efficiency, reduce process variations and optimize yield and throughput.

We’ve widely deployed “digital twins” – virtual replicas of critical assets – to model, forecast and manage operations in real time. Specific AI or machine learning models optimize production parameters, leveraging historical and current data to ensure quality and resource efficiency.

Using insights from our Digital Performance Management model, we’ve reduced process variability by 63%.

Our manufacturing units are equipped with internet of things-based monitoring systems with predictive analytics – superior AI algorithms to predict equipment failures before they occur. This approach has reduced downtime by more than 50%, enhancing our operational efficiency remarkably.

Soft sensors, powered by AI, enhance data collection and analysis, improving product quality and optimizing process conditions. The AI-driven analytics system manages energy consumption, reducing operational costs and achieving a 20% cut in Scope 1 emissions, supporting sustainability goals.

Integrating AI throughout our production process enhanced automation and boosted operational efficiency, laying the groundwork for a more sustainable and environmentally friendly future. Diving in headfirst across all 50 of our plants, we plan to deploy 10-12 use cases involving emerging technologies throughout our global operations this year and next.

A key first step in doing so has been to ensure all our plants are connected and integrated with an Operational Data Lake to get a real-time and integrated view of data to help us deliver AI or machine learning-based interventions to improve the yield and throughput.

Our JUMP (Jubilant Model Plant) serves as a digital lighthouse to perfect AI models before broader deployment, while our Digital Centre of Excellence drives this transformation with AI experts and Digital 101 training for all employees.

Recognition is also key in this journey – our rewards programme nurtures digital champions while DigiScoop spreads success stories, helping scale innovation.

We believe our company-wide adoption of AI will redefine what’s possible in chemical manufacturing.

At Siemens, we experience AI's transformative impact on manufacturing daily, boosting productivity, efficiency and sustainability. With rising labour costs, skill shortages and a need for eco-friendly solutions, AI is a crucial part of our vision for the industrial metaverse.

AI applications deliver remarkable results across our entire value stream at Siemens Electronics Factory Erlangen. For example, machine learning optimizes testing procedures, significantly increasing first-pass yield and boosting efficiency.

AI-enabled robots that pick and place different parts and materials in our fully automated assembly lines reduce automation costs by 90%. Manual workers are also empowered with AI-guided systems, enhancing productivity and quality.

Our industrial-grade AI infrastructure, built on Siemens hardware and software, simplifies adoption and reduces change management.

Automated training and deployment pipelines minimize the efforts for updates, while continuous automated monitoring ensures reliability and trust in AI algorithms. This enhances scalability, lowers adoption barriers and fosters trust.

Mengniu’s “digitalization 1.0” focused on digitalizing dairy farms and factories to achieve comprehensive digital coverage on the supply side, from raw milk to production.

Digitalization 2.0 shifted to optimizing management by building a digital marketing and consumer-side operations system, creating precise consumer profiles to enhance service experiences and marketing.

With digitalization 3.0, Mengniu integrates AI across the supply and consumer sides to optimize supply chain processes and boost efficiency. Three key AI scenarios have emerged.

In its One-Stop Laboratory, AI modules like neural network image recognition and reinforcement learning-based intelligent scheduling replace manual testing, ensuring accuracy and efficiency in critical test stages.

For procurement and cyclic delivery, AI automates supplier order scheduling and vehicle dispatching, increasing inventory turnover by 73% and operational efficiency by 8%.

In predictive maintenance, AI algorithms analyze equipment data to forecast faults and prevent downtime. These systems have enhanced overall production decision-making and operational efficiency.

Midea washing machines explore and restructure end-to-end green and sustainable new capabilities, widely deploying a variety of digital technologies integrated with AI applications in product design, manufacturing – quality, equipment and energy – and logistics, promoting intelligent operation in various sub-scenarios.

We have achieved a 25% reduction in development cycles, a 53% reduction in poor quality and a 29% optimization of logistics paths. The company is witnessing factory-scale adoption through the use of AI.

The deep application of AI in the entire factory process covers 457 sub-scenarios, mainly through self-developed small sample intelligent algorithms and open AI cloud platforms, significantly reducing sample collection and training time, and lowering scale promotion and operation costs.