The metals industry stands at a critical juncture, facing mounting pressure to enhance efficiency, reduce emissions, and remain competitive in an evolving global landscape. As digital transformation accelerates, artificial intelligence (AI) is emerging as a key enabler, optimising operations, minimising waste, and improving decision-making.
AI’s impact is being felt across the metals industry, driving efficiencies in energy management, predictive maintenance, and production planning. “We’re witnessing a fundamental shift in how the steel, aluminium and other metals industries approach operational efficiency and sustainability,” Tarun Mathur, Global Portfolio Manager Digital Solutions, Metals Industry, ABB Process Industries, says. “As manufacturers face growing demand for high-quality materials, AI-powered tools are increasingly coming to the fore to help identify production gaps, analyse data, and support better decision-making.
“What often sets ABB’s approach apart in this space is the combination of advanced data analytics and machine learning with deep domain expertise – a crucial factor in an industry where process knowledge is as important as technological innovation. A key application is our ML-based energy demand forecasting, enabling manufacturers to optimise energy use and navigate dynamic pricing in continuous and batch processes. Additionally, AI is being integrated into traditional solutions to enhance operational efficiency. For example, we leverage ML-driven thermal models to predict heat losses in steel manufacturing, thus optimising energy consumption.”
These solutions exemplify how AI, when coupled with domain expertise, can translate into tangible benefits, enabling manufacturers to make data-driven decisions that enhance both productivity and sustainability.
Strategic considerations for AI implementation
For executives considering AI adoption, success hinges on more than just deploying the technology. “For metals manufacturers, implementing AI requires a clear strategy aligned with business goals,” Mathur adds. “The first step on this journey is ensuring data readiness. While many producers generate vast data streams, they often lack the infrastructure to make this data actionable. This isn’t just a technical hurdle; it’s a strategic priority. To stand the greatest chance of success, businesses must carefully consider how to collect, structure, and leverage their data to drive genuine business value.”
Bridging legacy systems with AI-powered solutions presents another challenge. “Integration strategy is critical, particularly when dealing with legacy systems,” Mathur continues. “Bridging these with new technologies demands expertise in both operational technology and IT. By viewing digitalisation as a long-term investment, with patience for gradual value creation, we’ve found businesses are more likely to achieve lasting success.”
Organisational preparedness plays a pivotal role in AI deployment. Successful implementation hinges on change management. According to Mathur this means developing comprehensive training programmes and defining roles and accountability structures for AI model performance. “In an industry traditionally resistant to technological change, by creating internal champions who can demonstrate its benefits and fostering a culture of adaptability, organisations can position AI as a collaborative tool that enhances human expertise rather than replacing it,” he says.
Overcoming implementation challenges
AI deployment in metals manufacturing requires a structured approach to navigate operational complexities. Implementing AI within complex metals manufacturing environments requires a methodical approach. “It may seem obvious but start by focusing on clear business objectives rather than the allure of the technology itself,” Mathur says. “It’s surprising how many instances we’ve seen of businesses deploying AI without understanding what it is they want to achieve.”
Data integrity remains a critical factor. The quality of data underpinning AI systems is paramount, particularly in the metals industry where batch processes and varying product grades add layers of complexity. “We have learned that proper data labelling and structuring must be tailored to these unique characteristics,” Mathur explains. “This isn’t just about collecting data – it’s about understanding what that data means in the context of specific processes and products.
A common challenge is unrecorded data from manual interventions, which are quite common in metalmaking processes. To address this, ABB implements internet of things (IoT) sensors to capture real-time insights, such as tracking critical equipment movements in steel mills. However, technology alone is not enough. Collaboration between domain experts and customers is essential to bridge data gaps and align AI systems with operational realities, ensuring practical and sustainable results.”
AI in action: real-world impact
The benefits of AI are most evident when applied to real-world manufacturing environments. AI’s impact in metals production is perhaps best illustrated through a recent implementation of ABB Ability Expert Optimizer, an advanced process control technology, at a major Indian steel plant. Acting as an ‘autopilot’ for sintering and pelletising, it uses Model Predictive Control to optimise processes. Video analytics manage pellet size, reducing rejects by nearly 50 per cent. In the indurating furnace, AI models balance variables like fuel gases, temperatures, and machine speed to improve pellet quality. Unlike standalone solutions, this system operates in closed loop with automation for real-time results.
Similar benefits have been realised in steel production. “Another example of this approach in action is our ABB Ability Smart Melt Shop solution,” Mathur continues. “At JSW Steel’s Dolvi Works plant, this integrated system, which combines multiple IoT sensors with AI-driven analytics, has delivered significant returns. The plant saw casting speed increase by 4-5 percent, leading to an additional 24,000 tons of annual output, while simultaneously reducing energy costs by $250,000. These results exemplify how AI, when properly implemented with industry expertise, can deliver tangible operational and business benefits.”
Balancing automation with human expertise
While AI’s role in manufacturing continues to grow, human expertise remains indispensable. “Human expertise will continue to play a critical role as AI matures in manufacturing,” Mathur says. “Currently, AI serves as an advisory tool, offering recommendations for operators to implement. This approach allows humans to retain control and provide valuable feedback, enabling the system to improve over time.
“A key challenge of fully autonomous AI in manufacturing is accountability. Since humans are ultimately responsible for AI decisions, outcomes must be reviewed and moderated by human experts, especially in unforeseen scenarios that historical data cannot predict. In these cases, human judgment remains crucial to correct and guide the AI system.”
The long-term vision for AI in manufacturing is an evolution, not a revolution. As trust in AI solutions grows, operators will gradually adopt its recommendations. Currently, it is difficult to quantitatively model complex variables such as safety and productivity in AI systems. Therefore, balancing multiple objectives, particularly when those objectives involve trade-offs between safety and efficiency, requires human expertise to make the final call. Thus, while AI will continue to evolve and become more capable, human involvement will remain essential to ensure that decisions are made with the appropriate context, oversight, and ethical considerations.
The future of AI in metals manufacturing
The metals industry is undergoing a profound transformation, with AI playing a central role in improving efficiency, sustainability, and operational excellence. AI-driven quality control using computer vision and deep learning models is expected to further automate defect detection, improving yields and reducing material wastage. Real-time AI-powered digital twins will enhance operational transparency, enabling predictive adjustments that prevent inefficiencies before they occur.
Additionally, AI-enhanced workforce management systems are expected to support the next generation of skilled workers, enabling intelligent task allocation and predictive workforce planning. Digital ecosystems will allow AI solutions to interact seamlessly with enterprise resource planning (ERP) and supply chain management (SCM) systems, creating a more resilient and agile manufacturing network.
As data ecosystems grow more interconnected, integrating AI with enterprise-wide digital strategies will become a defining factor in competitiveness. Those companies that take a proactive approach to AI adoption will gain a significant edge, not just in efficiency but in innovation and agility. By strategically implementing AI solutions and ensuring seamless integration with existing processes, manufacturers can unlock substantial business value while maintaining the crucial role of human expertise in decision-making.
