Take reheating furnaces, for example. Operators, worried about under-heating, crank up temperatures unnecessarily - burning far more energy than needed. Or the long delays before anyone even spots a maintenance issue when data is logged manually. These inefficiencies aren’t just inconvenient - they’re expensive. Plants that stay manual often run much thinner Earnings Before Interest, Taxes, Depreciation and Amortisation (EBITDA) margins than those that invest in modern AI-driven operations.

Here’s the fix: AI load balancing. It’s not about replacing people; it’s about replacing the boring, error-prone tasks they hate. AI can juggle hundreds of variables in real time: temperature, flow rates, crane schedules. That keeps operations smooth and precise. POSCO nailed it, boosting efficiency and cutting energy use across their operations.

The Old Way vs. The Smart Way

If you’re tired of firefighting inefficiencies, it’s time to rethink your approach. Let’s dive into how the biggest names in steel - Tata Steel, U.S. Steel, Big River Steel, and ArcelorMittal - are using AI to cut waste, boost output, and protect their bottom line.

Tata Steel: Using Simulations to Optimise Capacity

The Problem: Crane Bottlenecks and Manual Decisions

Tata Steel’s melting shop was dealing with a frustrating bottleneck caused by uneven distribution of cranes and ladles. The issue? Crane tasks were being assigned manually, without a clear strategy or the benefit of real-time optimisation. This manual approach struggled to keep up with the complex interdependencies between equipment, fluctuating processing times, and unexpected breakdowns. It became clear that a safer, more efficient way to test and implement process improvements was desperately needed [3].

The Solution: Using AI Simulations

To address these challenges, Tata Steel created a digital twin of its melting shop using AnyLogic simulation software. This virtual model, developed by a team including S. Choudhary, A. Kumar, and S. Kumar, replicated every crane movement and physical constraint with precision. They then used Microsoft Bonsai to train a reinforcement learning model aimed at reducing crane waiting times at LD converters.

This digital twin allowed them to conduct 270 virtual experiments, including unconventional scenarios like suspending empty ladles nine metres above the floor. These rigorous tests ensured that process changes could be implemented with confidence, achieving a first-time success rate of over 90% [1] [2] [4] [5].

The Results: Higher Output and Better Efficiency

The results were game-changing. By introducing AI-driven crane scheduling, Tata Steel boosted daily throughput by 8%, adding an extra two heats per day. In practical terms, that’s about 3.3 tonnes of additional steel daily, saving the company millions of pounds each year. Crane utilisation hit a steady 80%, while vessel waiting times dropped significantly.

This initiative was part of Tata Steel’s broader digital transformation programme (2015–2020), which included over 250 digital twin models managed through their Industrial Revolution Optimisation Centre (iROC). Altogether, this programme delivered cumulative cost savings of £1.4 billion [1] [4] [5].

U.S. Steel: Generative AI for Real-Time Scheduling

The Problem: Complex Scheduling Challenges

U.S. Steel, like many heavy manufacturers, struggled with time-consuming inefficiencies. Technicians spent hours combing through paper manuals to diagnose equipment failures, while production schedules were frequently disrupted by supply chain hiccups and unexpected breakdowns. These issues made it clear that a more efficient, AI-driven solution was needed to streamline real-time scheduling.

The Solution: Generative AI for Smarter Forecasting

To tackle these challenges, U.S. Steel turned to AI, drawing inspiration from other steel plants’ successes. In September 2023, the company partnered with Google Cloud to roll out MineMind, a generative AI system powered by Vertex AI and Document AI. This tool can instantly summarise repair instructions and create detailed diagrams, complete with validity scores. The initial deployment at Minnesota Ore Operations covered over 60 haul trucks, allowing MineMind to start delivering results right away [6].

Matt Wilding, U.S. Steel’s Senior Director of Financial Data, Analytics, and Enterprise Performance Management, highlighted the collaborative effort:

“We’ve been engaging in a partnership with Google Cloud to create the first generative AI applications for the steel industry. We take the expertise on the application side from the Google team and U.S. Steel’s expertise on the operations side, put our heads together and came up with some innovative solutions.” [7]

Beyond maintenance, the AI was designed to handle real-time data analysis and decision-making. By integrating data from sensors, Programmable Logic Controllers (PLCs), and legacy systems, MineMind evaluates thousands of production scenarios and adjusts schedules automatically when disruptions occur [6][8].

The Results: Faster Repairs and Improved Efficiency

The impact of MineMind has been immediate and measurable. Work order completion times have dropped by an estimated 20%, allowing technicians to focus on more critical tasks rather than being bogged down with paperwork [6][8]. David Burritt, President and CEO of U.S. Steel, described the benefits:

“Faster repair times, less down time, and more satisfying work for our techs are only some of the many benefits we expect with generative AI.” [8]

Looking ahead, U.S. Steel aims for a 20% boost in overall productivity as the AI expands into areas like logistics, supply chain management, and process automation. By layering AI on top of existing systems instead of overhauling them entirely, U.S. Steel has shown how legacy operations can embrace modern tools to achieve substantial improvements.

Big River Steel: Predictive Analytics for Yield and Downtime

The Problem: Inconsistent Yield and Unplanned Downtime

Big River Steel faced a familiar struggle in the steel industry: equipment failures and unpredictable yield levels. These issues led to production losses and delayed deliveries, leaving both operations and customers in a tough spot. Relying on reactive maintenance only made things worse, as problems were addressed after they disrupted operations. What they needed was a forward-thinking system to tackle downtime and yield unpredictability head-on.

The Solution: AI-Driven Predictive Maintenance and Load Balancing

To address these challenges, Big River Steel turned to predictive analytics, building on their AI-powered load balancing strategies. They implemented a “learning mill” architecture - known as Big River 2 - that crunches enormous amounts of data to spot potential problems before they become critical [11]. This system integrates data from over 14,000 sensors spread across key processes. The AI keeps a close eye on electrical signatures, temperature, and vibration, flagging anomalies in real time [9].

To make this even more effective, they layered AI analytics onto their existing Computerised Maintenance Management System (CMMS). This ensures that every predictive alert is automatically turned into a structured work order, streamlining maintenance workflows [9][10]. On the production side, the Endless Strip Process (ESP) uses live data feedback to fine-tune production parameters on the fly. This keeps each coil consistent and eliminates surprises between batches [11].

The Results: Enhanced Yield and Reduced Downtime

By combining predictive maintenance with AI-driven load balancing, Big River Steel shifted from reactive to proactive operations. Improved yield. Fewer rejected coils. Less emergency downtime. Longer equipment lifespans. Addressing issues before they escalate has transformed their production process into a smoother, more reliable operation.

The AI Revolution Nobody Noticed in the Steel Industry | T V Narendran | Tata Steel

ArcelorMittal: Sensor-Based Maintenance and AI Monitoring

ArcelorMittal has taken its AI capabilities to the next level by combining sensor-based maintenance with advanced monitoring systems, aiming to revolutionise how equipment reliability is managed.

The Problem: Reactive Maintenance and Costly Failures

Unexpected equipment failures were a recurring nightmare at ArcelorMittal’s plants. Emergency repairs disrupted production, particularly with oxygen lances in basic oxygen furnaces, which were notorious for breaking unpredictably. When a lance failed, it often contaminated molten metal, leading to expensive clean-up efforts and production losses [13]. The numbers painted a grim picture: emergency work orders made up 34% of all maintenance tasks, while unplanned downtime consumed 8.5% of production hours [12]. The reliance on reactive maintenance not only drained resources but also shortened the lifespan of critical equipment, leaving engineers constantly firefighting instead of focusing on long-term solutions.

The Solution: IoT Sensors and Predictive Monitoring

To tackle these challenges, ArcelorMittal developed its Sentinel platform, a rugged Industrial Internet of Things (IoT) solution built to withstand the extreme conditions of steel production - high heat, intense vibrations, and corrosive environments [12]. The system deployed thousands of wireless sensors to monitor key metrics like vibration, temperature, sound, and electrical currents across a wide range of equipment, including robots, motors, and blast furnaces.

With edge computing handling data locally, cloud bandwidth usage was slashed by 85% [12]. Machine learning algorithms, trained on years of failure data, provided teams with an average of 15 days’ notice before critical breakdowns [10]. At the Hamilton, Canada facility, computer vision added another layer of precision, tracking the usage of individual oxygen lances to predict the best replacement times and avoid sudden failures [13]. The integration of AI allowed these predictions to automatically trigger work orders in the Computerised Maintenance Management System (CMMS), ensuring issues were addressed before they escalated [12].

The Results: Smoother Operations and Cost Savings

Unplanned downtime dropped 40%, falling from 8.5% to 5.1% of production hours [12]. Emergency work orders fell by 68%, and maintenance costs per tonne decreased from £12.80 to £9.40 - a 27% reduction [12]. The improvements didn’t stop there: robots saw a dramatic increase in Mean Time Between Failures (MTBF), jumping from 620 hours to 1,150 hours - an 85% boost [12]. Bearings lasted 2.3 times longer, and early detection of 27 failures saved 31 hours of downtime [14]. At Hamilton, the computer vision system eliminated downtime from lance breakages entirely, saving millions annually [13].

One Reliability Engineering Director summed it up perfectly:

“Sentinel transformed our maintenance approach, enabling us to detect issues weeks ahead and address them proactively” [12].

Today, Sentinel monitors over 200,000 assets across more than 50 plants, processing an astonishing 3.2 billion sensor data points every day [12].

Key Lessons from AI Load Balancing in Steel Plants

The examples shared earlier highlight how AI can reshape operations in steel manufacturing. These insights reveal the critical elements that separate efficient, AI-driven plants from those still bogged down by outdated processes like spreadsheets and last-minute fixes.

Integration Beats Full Replacement

Scrapping existing systems for a complete overhaul is often too costly and risky. A better strategy is to integrate AI into current workflows. Tata Steel’s experience shows how layering AI onto existing systems can minimise both risks and expenses [4][16]. Starting with small-scale implementations to prove value before scaling up is a smarter way to avoid costly missteps. This phased approach also underscores the importance of thorough testing before rolling out major changes.

Test on a Digital Twin, Not on Live Kit

Making changes directly on live equipment is a risky move - one mistake could lead to disastrous consequences. Digital twins offer a safer alternative, allowing manufacturers to simulate and test adjustments in a virtual environment. For instance, Tata Steel used digital twins to evaluate 847 burden combinations for blast furnace optimisation in just two days - a process that would have taken months using physical trials [4]. This resulted in a 90%+ first-time success rate for changes and a 4–6% reduction in coke usage [4][16].

In another example, a European steel producer used a digital twin to detect a cooling water temperature issue 38 days before it caused physical damage, preventing a £3.3 million emergency shutdown [17]. Dr Petra Krahwinkler from Primetals Technologies sums it up perfectly:

“The advantage of AI is that it can do this analysis in real-time… rather than operators looking at vast amounts of monitoring data manually, these systems can guide them precisely to what they need to focus on” [15].

Real ROI: Downtime Slashed, Yields Boosted

The financial impact of AI-driven improvements is undeniable. By applying data strategically, steel plants see direct cost savings and efficiency gains. Beshay Steel, for example, cut downtime by 47% and saved £2.8 million annually, achieving payback in just 4.2 months. Meanwhile, JSW Steel reduced load tracking time from 45 minutes to just three seconds, freeing up two million man-hours annually [16].

A £790 million steel mill using AI for scheduling boosted production by 1%, adding over 1,000 tonnes of finished product each year, while cutting planning time from five days to just one hour [18]. Other benefits include energy savings of 8–12% and defect rate reductions of 30–40% when first-time quality exceeds 90% [4]. Given that unplanned downtime can cost over £39,000 per hour, even small improvements lead to substantial savings [16].

Between 2015 and 2020, Tata Steel’s CEO T.V. Narendran spearheaded the development of their Industrial Revolution Optimisation Centre (iROC), covering over 15 plants with more than 250 digital twin models. This initiative delivered approximately £1.1 billion in savings and a 775% return on investment, while cutting unplanned downtime by 22% [4][16]. These results highlight the importance of precise measurement and a gradual, well-planned adoption of AI to secure the future of steel manufacturing.

Start Small. The ROI Shows Up Fast.

AI-driven load balancing is changing the game for steel manufacturing - making production faster, more efficient, and environmentally friendly. Just look at the numbers: Tata Steel saved £1.4 billion, while Beshay Steel slashed downtime by 47% in less than five months. These aren’t just isolated wins; they’re proof that embracing AI can shift plants from merely surviving to thriving. The secret? Transitioning from reactive maintenance to proactive, data-driven efficiency.

The takeaway here is clear: you don’t need to go all-in from day one - just start. You can skip the headache of overhauling your entire ERP or committing to a multi-year transformation. Instead, begin small with a four-week audit to uncover your biggest inefficiencies, like energy drains or coordination hiccups. Focus on quick, impactful fixes in the first 90 days - whether that’s sealing air leaks or balancing furnace loads. AI can serve as an overlay, pulling real-time insights from your existing systems, rather than replacing tools that already do the job.

Platforms like GoSmarter make this process accessible. For instance, the MillCert Reader (£275/month, billed annually) digitises messy PDF mill certificates, saving over 120 hours of manual work each year. Meanwhile, the Cutting Plans module (£1,000/month, billed annually) reduces scrap rates by 50% and replans production in seconds. These tools turn mountains of disorganised data into decisions you can act on — in seconds, not days — without the hassle of a full system overhaul.

The steel plants of tomorrow are already taking action today, blending automation with measurable ROI and tying carbon reduction directly to operations excellence.

If you’re still stuck with spreadsheets and scrambling to fix problems at the last minute, you’re not just behind the curve - you’re losing money with every shift.

FAQs

The 8 KPIs every metals manufacturer should track: Overall Equipment Effectiveness (OEE), throughput rate, scrap rate, first pass yield, On-Time-In-Full (OTIF) delivery, energy per tonne, embodied carbon per tonne, and cost per tonne. Track these in real time and you can see exactly where you’re losing money — before the shift ends.

The Old Way vs. The Smart Way

Let’s break down the KPIs that matter most - OEE, scrap rates, energy efficiency, and cost per tonne - and how to build a dashboard that doesn’t just inform but transforms your operations.

Watch: OEE and KPIs Explained in Three Minutes

OEE, throughput, and yield are the three numbers that expose hidden losses most metals manufacturers never catch in time. This three-minute explainer shows why they matter and how to read them correctly.

Which Production Efficiency Metrics Should You Track?

Tracking production efficiency is the fastest way to find losses caused by downtime, slow cycles, or quality issues. These key performance indicators (KPIs) distinguish factories hitting their tonnage goals from those falling behind. They also set the stage for deeper insights into quality and waste management.

Overall Equipment Effectiveness (OEE)

OEE is the most critical number on your dashboard - it combines availability, performance, and quality into one percentage that reflects your actual capacity utilisation [4][5]. While 85% is considered world-class, most metals manufacturers operate between 60% and 75% [4]. For a steel plant producing 2 million tonnes a year, every OEE point represents about £10 million in revenue [6].

Display OEE in real time and you catch a problem in the first five minutes — not at the end-of-shift debrief. A loss waterfall visualisation can show exactly where capacity is being lost [7]. Often, the bottleneck lies in critical equipment like the continuous caster or hot strip mill - addressing these constraints can unlock higher throughput [6][5].

Throughput Rate

Throughput measures how many metric tonnes are produced per hour or shift, excluding scrap and rework. It’s a direct indicator of how well you’re using your equipment. For instance, if a rolling mill rated for 200 tonnes per hour only produces 140, you’re losing 60 tonnes of potential output every hour.

If throughput drops more than 8% below capacity for over 15 minutes, stop and find the cause [8]. That’s not a drift — that’s a problem. Real-time tracking helps uncover small stoppages and speed reductions that manual methods often miss. Automated tools typically detect 30% to 50% more performance losses than manual tracking [6][7]. Following throughput, assessing yield rate can give you a clearer picture of first-pass quality.

Yield Rate

Yield rate measures the percentage of metal meeting quality standards on the first pass, without requiring rework or repairs [4][1]. For example, if you produce 1,000 tonnes but only 850 are saleable, your yield rate is 85%, with 15% lost to waste. In multi-step processes like rolling and finishing, these losses can compound. For instance, a five-step process with a 95% yield at each step results in an overall yield of just 77.4% [4].

Pair yield rate with scrap and defect data and you’ll see exactly where quality breaks down and which process is to blame.

On-Time-In-Full (OTIF) Delivery

On-Time-In-Full (OTIF) measures the percentage of orders delivered complete and on the promised date. For service centres and fabricators, it is often the number that determines whether you keep a customer. Most metals businesses track it retrospectively — a spreadsheet updated after a delivery fails. By then, it is too late.

Real-time OTIF tracking means knowing today which jobs are at risk before they miss their date. That requires live visibility of what material is in stock, what is already committed to other orders, and whether the cutting schedule can deliver on time. When that data lives in disconnected spreadsheets, OTIF surprises are inevitable.

GoSmarter’s scheduling module shows live commitment status across all open jobs — which are on track, which are at risk, and which jobs are competing for the same material. Planners can act before a delivery slips rather than explain why it did.

Which Quality and Waste KPIs Matter Most?

Reducing waste and maintaining high-quality standards are constant challenges for any operation. Even when scrap metal prices are favourable, they rarely offset the combined costs of wasted materials and labour [9]. To gauge whether your operation is efficient or leaking profits, focus on these three KPIs. They complement production metrics by ensuring quality and minimising waste across shifts.

Scrap Rate

Scrap rate indicates the percentage of material that ends up unusable and cannot be salvaged. Ideally, most established operations aim to keep this below 5%, while top-tier plants often achieve rates under 2% [9][10]. A high scrap rate often points to issues with materials, equipment, or processes - such as misaligned fixtures, worn tools, or human errors in estimation [9].

Pareto charts can help you pinpoint which processes or machines are the main culprits for scrap [9]. For a deeper dive into benchmarks and reduction strategies, see the Scrap, Waste & Yield Optimisation hub.

For example, Midland Steel moved from manual cut planning to GoSmarter’s AI-driven Cutting Plans for rebar and structural sections. Scrap rate halved — recovering material per month that had previously been written off as offcut waste. Admin time dropped by over 120 hours a year: time that had been spent manually transcribing heat numbers and grades from PDF mill certificates into spreadsheets. With accurate, live stock data feeding their order commitments, the team also stopped over-ordering buffer stock to cover for planning uncertainty, cutting the working capital tied up in slow-moving bar. The whole change was live within a week [2].

Stop wasting raw material because someone guessed instead of measured [2].

Defect Rate

Defect rate measures the percentage of units with flaws, including those that can be repaired through rework. This metric is invaluable for identifying root causes, whether they stem from material inconsistencies, equipment malfunctions, or process deviations [14]. Real-time sensor data can detect issues like equipment drift or tool wear before they result in defects [9].

To reduce defects, investigate causes by machine, shift, or material batch. Standardising work instructions can help minimise variability, while preventive maintenance can address issues like misaligned fixtures or worn-out tools before they escalate [9].

First Pass Yield (FPY)

First Pass Yield goes a step further by measuring the percentage of products that pass quality checks on the first try without needing rework. An FPY above 95% is considered excellent, while anything over 90% is generally acceptable [11]. Achieving high FPY eliminates the “hidden factory” of rework, which drains extra labour, materials, energy, and accelerates equipment wear [11][12].

Consider a steel service centre processing structural sections and flat plate. By fitting IoT sensors at entry and exit points on each production line, the team identifies exactly which cut or forming step is generating the most rejects. Statistical process control (SPC) charts highlight tool wear trends before parts go out of spec [13]. Incorporating FPY data into your dashboards allows for immediate adjustments, bringing quality control in line with live production. For a five-step line with 95% yield at each stage, the overall FPY is just 77.4% [4] — tracking each step separately shows you exactly where to focus first.

Which Cost and Sustainability Metrics Should You Track?

Efficiency and quality only tell half the story. Cost and sustainability KPIs protect your margins and keep you on the right side of UK CBAM and ESG requirements. These metrics track production costs, energy consumption, and compliance — key elements increasingly required for ESG reporting.

Energy Efficiency (kWh per Tonne)

Energy efficiency measures how many kilowatt-hours are used to produce one tonne of metal. This metric impacts both your production costs and carbon footprint[3]. To calculate it, divide the total energy consumed by the tonnes of metal produced. If energy use increases, it could point to outdated equipment or poor production scheduling. Dashboards that break down energy consumption by shift or production line can help you identify and address inefficiencies[3][18]. Tracking embodied carbon alongside energy use ensures you stay on target for regulatory compliance and sustainability goals.

Embodied Carbon per Tonne

Embodied carbon measures the CO₂ emissions generated per tonne of metal produced. Fail a CBAM audit and you face import duties based on estimated — not actual — carbon content. Estimated carbon is always worse than measured. Companies relying on manual cert processing are one audit away from finding that out the hard way. Tracking embodied carbon per tonne is how you build the evidence trail before the auditor arrives[13][16]. Calculating embodied carbon manually from mill certificates can be slow and prone to errors.

For a full guide on automating certificate handling, see the Mill Certificate Automation hub.

GoSmarter’s MillCert Reader does more than extract numbers. When a PDF cert arrives — scanned, emailed, or downloaded from a supplier portal — GoSmarter reads the heat number, grade, spec, and mechanical properties, then checks them against your purchase order automatically. If something does not match, it flags the non-conformance before the material reaches the floor.

The cert stays linked to the stock record, the cut job, and the delivery note. When CBAM or a customer audit asks for material provenance, you are not scrambling through a filing cabinet.

Companies using AI-driven cutting plans have seen scrap rates drop by 20–50%, boosting margins while reducing embodied carbon per tonne of finished products[2][15]. Alongside emissions metrics, monitoring production costs per tonne is vital for maintaining profitability.

Cost per Tonne

Cost per tonne is a simple but powerful metric: divide the total production costs - including materials, energy, labour, and overhead - by the tonnes produced[17]. This figure is critical for protecting margins. Dashboards can break down these costs by shift or production line, helping you spot inefficiencies. For example, if one shift consistently incurs higher costs, investigate whether setup inefficiencies, excessive scrap, or energy waste are to blame. Companies focusing on these financial KPIs have achieved profitability increases of up to 20%[17]. Tying cost per tonne to First Pass Yield is also effective - products that meet quality standards on the first attempt use less energy and materials than those requiring rework[18][3].

How Do You Build a KPI Dashboard for Metals Manufacturing?

Creating an effective dashboard isn’t about cramming in every metric you can think of - it’s about giving your team access to the right numbers at the right time. The sweet spot? Around 8–10 key KPIs that align with your plant’s goals, whether it’s cutting downtime or slashing scrap rates [1]. Operators need live machine status and cycle times. Managers need OEE and cost-per-tonne trends. Build for both. These steps will help you customise a dashboard that works for everyone on your team.

Choose the Right KPIs

Your dashboard should reflect the needs of different roles within your operation. For example:

• Operations managers: Focus on OEE, capacity utilisation, and scrap rate.
• Maintenance teams: Track downtime, MTBF (Mean Time Between Failures), MTTR (Mean Time to Repair), and maintenance costs per unit.
• Quality teams: Monitor metrics like First Pass Yield (aiming for 98% in top-performing plants), defect rates, and customer returns [19].
• Leadership: Look at revenue per employee and manufacturing costs as a percentage of revenue to evaluate workforce efficiency and financial performance.

The key here is actionability. A KPI is only useful if it drives decisions - otherwise, it’s just noise [20]. Once you’ve nailed down the metrics that matter, ensure they’re powered by real-time data.

Connect Real-Time Data Sources

GoSmarter adds intelligence to the systems you already use — not replace them. You can be live in a day from a CSV upload. Connecting to an ERP, MES, or IoT sensors via API is available when you are ready, but never a requirement for day one [5]. GoSmarter runs in the browser, hosted in the EU, and your data belongs to you.

For real-time stock visibility, Metals Manager links your stock records, mill certs, and open orders — showing exactly what material is available, committed, and due for delivery. To see how AI cut-planning fits in, visit the Cutting Optimisation hub.

GoSmarter’s MillCert Reader uses AI to pull data straight from mill certificates — scanned or digital — without any manual typing [2]. Set up threshold alerts via SMS or email when critical metrics like downtime or scrap rates exceed acceptable limits [5]. Tackle issues as they arise, not after the fact.

Design Clear Visualisations

Once your data is flowing in real time, the next challenge is presenting it in a way that’s easy to understand. Use visual tools that make performance gaps obvious at a glance:

• Gauges: Ideal for real-time metrics like OEE.
• Line charts: Great for tracking trends over time.
• Pareto charts: Pinpoint the main causes of defects or downtime [1].

Add colour-coding (red, yellow, green) to flag urgent issues like production delays or quality problems. Always include a “Target vs. Actual” comparison to help teams see immediately whether they’re hitting their goals [1]. Make sure the dashboard is mobile-friendly so shop floor operators can access it on the go [21].

For plant managers, the dashboard should allow for a quick “5-minute check” of both financial and production performance. Reliability engineers, on the other hand, need tools for deeper analysis of failure modes and asset health [22]. The design should reflect these varied needs, ensuring everyone gets the insights they require to act effectively.

Getting Your Team on Board

A dashboard only works if people use it. The biggest reason KPI projects stall is not the technology — it’s the conversation that never happened. Before you build, agree on which three metrics the MD will look at each morning and what action each one triggers.

Start with the operators. Show them how the dashboard makes their shift easier — fewer audit panics, faster cert retrieval, less back-and-forth on material availability. If operators trust the data, they will flag when something looks wrong. That feedback loop is what makes dashboards improve over time.

Roll out in phases. Begin with one data stream — usually mill certificate processing, because it has an immediate, visible payback. A metals business processing 30 PDFs a week typically spends 8–15 minutes per document on manual entry. That is 4 to 7.5 hours every week, or up to 390 hours a year. At £30/hour for an administrator, that is up to £11,700 annually before any error correction. GoSmarter’s MillCert Reader handles the same task in under 30 seconds. Add scheduling and OTIF tracking in week two or three. You do not need a systems integrator or a project manager to get started.

Getting Started: The 30-Day Path

Not sure where to start? Most GoSmarter customers begin with one data stream — usually mill certificates, because that is where the most manual effort lives. Once certs are being read automatically and flowing into your stock record, the KPIs that depend on material data (scrap rate, yield, cost per tonne) start updating without anyone typing. That typically takes a day to set up. Scheduling and live commitment tracking come next, usually in the second or third week. You do not need a systems integrator, a data warehouse, or a project manager. You need a CSV export of your current stock and an hour on a call.

Stop Guessing. Build the Dashboard.

Pick 5–10 KPIs that match your plant’s goals: OEE, scrap rate, energy efficiency per tonne. That’s it. Everything else is noise[1]. Take ArcelorMittal as a case in point: by prioritising OEE, production yield, and cost per tonne, they achieved a 10% boost in OEE, a 15% increase in yield, and a 12% cut in production costs per tonne[17]. That’s the power of a dashboard that drives action, not just information.

Manual data tracking is a drain on resources. Tools like GoSmarter’s MillCert Reader eliminate this inefficiency, saving hundreds of hours annually by automatically extracting heat numbers and grades from mill certificates[2]. Similarly, AI-driven Cutting Plans typically reduce scrap rates by 20–50% — the largest gains come on long products like rebar and structural sections, where optimising cut sequences across mixed bar lengths can recover tonnes of material that would otherwise become offcut waste[2]. Your team can stop chasing data and start fixing the actual problem.

Modern dashboards go beyond recording metrics - they actively enhance performance. Real-time monitoring of metrics such as throughput, defect rates, and embodied carbon per tonne allows you to address potential issues before they snowball into costly problems. Companies optimising their financial KPIs have reported up to a 20% rise in profitability, while those prioritising operational efficiency have cut production costs by as much as 15%[17]. Raw metal alloy can make up over 50% of direct unit costs[23]. Even a 1% reduction in scrap goes straight to margin.

The metals sector is moving swiftly towards predictive maintenance, embedded analytics, and AI tools that turn raw machine data into decisions [5]. If you’re still relying on manual data entry and end-of-day reports, you’re not just outdated — you’re losing money. A dashboard that shows your team exactly what to fix — and when — is faster than any end-of-day report ever could be. Your competitors are already running these. Your spreadsheets are not a fair fight.

FAQs

Here’s the hard truth: metals manufacturing is one of the biggest contributors to CO₂ emissions, responsible for 9% of global CO₂ output. With energy costs skyrocketing and carbon taxes like the EU’s CBAM kicking in this year, every wasted tonne of steel or kilowatt-hour of energy is directly eating into your margins.

But there’s good news. AI is changing the game, helping metals manufacturers reduce emissions, cut costs, and protect profits - all without ripping apart your current systems. Tools like GoSmarter automate the boring, time-sucking tasks that slow you down, from digitising mill certificates to optimising cutting plans.

The Old Way vs. The Smart Way

Let’s look at how AI can fix the mess and help you stay competitive in 2026 and beyond.

AI Reduces Energy & CO₂ by 25%

How AI Makes Sustainability Affordable and Practical

AI is already helping metals plants save money and reduce emissions - without requiring a complete system overhaul or a team of data scientists. By analysing thousands of variables in real time, AI makes small adjustments that add up to big savings. Let’s dive into a few ways this works.

AI-Driven Energy Optimisation Cuts Costs and Emissions

Energy costs can make up 20% to 40% of total production expenses in smelting operations [4]. AI tackles this by continuously fine-tuning furnace temperatures, airflow, and fuel blends to maintain the ideal thermal balance [4]. Instead of waiting for monthly utility bills to spot inefficiencies, AI identifies and addresses issues as they happen - like spotting burner wear before it starts eating into your profits.

“One operator, who has been there for 30 years, told me that this tool increased operational speed fivefold, minimising operator errors” [3].
Osas Omoigiade, Founder of Deep.Meta

While energy optimisation reduces costs and emissions, AI-powered scheduling ensures even more efficiency by cutting down on waste.

Reduce Waste and Scrap with Better Scheduling

AI-driven scheduling is a game-changer for cutting patterns and production runs, reducing scrap waste by as much as 50%. Even before full AI integration, tools like scrap and emissions calculators can provide instant insights to help you make smarter decisions [5].

Automate Manual Data Entry and Focus on Production

Manually entering data from PDF mill certificates into spreadsheets is a drain on time and resources. AI takes over this tedious task with computer vision and advanced language models, digitising certificates and feeding the data straight into your ERP system [1]. For example, Gerdau used AI to streamline ferroalloy use, cutting alloy costs by £3 per tonne while also improving their carbon footprint by using fewer additives [7]. This kind of automation not only simplifies workflows but also supports leaner, more sustainable production.

“AI is not just another tool – it’s a transformative force that redefines how we approach industrial automation; it enables us to shift from reactive operations to proactive, comprehensive decision-making” [6].
Thiago Maia, Executive Vice President at SMS group

GoSmarter: AI Tools Built for Metals Manufacturers

GoSmarter offers a collection of AI tools designed to take the hassle out of running a metals manufacturing operation. By automating tasks like mill certificate data entry, production scheduling, and scrap tracking, these tools let you focus on what truly matters - keeping the shop floor running smoothly, without drowning in spreadsheets.

Here’s a closer look at how GoSmarter simplifies cutting plans, digitises certificates, and integrates effortlessly with your ERP systems.

Smart Production Scheduler: Smarter Cuts, Less Waste

The Smart Production Scheduler leverages Genetic Algorithms to sift through thousands of cutting combinations for products like rebar. It aligns open orders with your inventory to create efficient cutting plans, significantly reducing leftover steel and scrap. Instead of manually piecing together plans and hoping for the best, the scheduler provides optimised first drafts that can cut scrap waste by as much as 50%. The result? Lower material costs, fewer emissions, and no need to overhaul your production process.

MillCert Reader: Automate Mill Certificate Data Entry

The MillCert Reader uses AI-powered OCR technology to extract data from unstructured PDF mill certificates, turning them into organised, actionable information. It even creates single-page PDFs sorted by heat code, eliminating the manual errors that often lead to compliance headaches. This tool can save production teams over 10 hours a month, freeing up time for more important tasks [8].

Seamless Integration with Your ERP Systems

No need to scrap your legacy ERP system to take advantage of GoSmarter. The platform works with what you already have, integrating smoothly through tools like Microsoft Azure Logic Apps, Power Automate, and Power BI. Setup is quick - log in, and you’re ready to go. Plus, the pricing is straightforward: start for free with basic scrap and emissions calculators, and scale up when you’re ready. No surprise fees or forced upgrades [1]. Just practical tools that fit right into your existing processes.

The ROI of AI in Manufacturing: Results You Can Measure

AI isn’t just a buzzword in manufacturing anymore; it’s delivering measurable results that directly impact the bottom line. By cutting waste, lowering emissions, and improving efficiency, manufacturers are seeing margins improve dramatically. The numbers speak for themselves - traditional methods simply can’t keep up.

Here are some real-world examples that showcase the power of AI optimisation.

Case Study: Before and After AI Optimisation

Midland Steel put GoSmarter’s Rebar Optimiser to the test in a two-week trial in 2025. Across 193 jobs and 734 tonnes of steel, the AI-powered cutting plans saved an impressive 20.22 tonnes of steel scrap - material that would have otherwise gone to waste.

“Smart technology can have a direct, measurable impact on reducing carbon emissions in steel manufacturing. The integration of AI and digital tracking has significantly improved our operational efficiency.”
Tony Woods, CEO at Midland Steel [2]

Another success story comes from a 2.4-million-tonne integrated steel plant that used AI for energy management in its blast furnaces and rolling mills. Over 18 months (2024–2025), the plant cut its energy intensity by 16% and reduced CO₂ emissions by 18%, saving £4.2 million annually [9]. These results underscore the financial and operational benefits of AI adoption.

In Egypt, Beshay Steel adopted AI-powered predictive maintenance to tackle unplanned downtime. The result? A 47% reduction in downtime and annual savings of £2.8 million. Even better, they recouped their investment in just 4.2 months [10].

How Lower Emissions Improve Your Bottom Line

Every tonne of material saved and every kilowatt-hour conserved adds up. With rising carbon taxes in the UK, cutting waste and optimising energy use isn’t just good for the planet - it’s a direct boost to your finances. By reducing emissions, you lower costs, protect margins, and free up resources without needing additional investments. It’s a win-win: better efficiency and a stronger competitive edge, all while maintaining quality.

Conclusion: Time to Modernise Your Manufacturing Operations

AI-driven solutions are reshaping metals manufacturing, delivering real-world results. These tools help reduce scrap, lower energy costs, and minimise emissions - all while protecting your bottom line. The numbers speak for themselves, and case studies highlight how quickly businesses see returns on their investment.

With carbon taxes and rising energy costs here to stay, the pressure to modernise is only increasing. The good news? You don’t need to rip out your existing ERP system or endure a drawn-out implementation process. GoSmarter works with what you already have, automating tedious tasks like processing mill certificates, optimising cutting plans, and tracking scrap. This lets your team focus on what matters most: production.

Start with a simple four-week audit to understand your current operations. From there, you can secure meaningful improvements in just 90 days by tackling areas like sealing air leaks, balancing furnace operations, and digitising documentation [1]. Real-time dashboards make energy efficiency visible, empowering operators to make smarter decisions. With flexible pricing that grows with your business, there are no hidden contracts tying you down. Every tonne saved and every kilowatt-hour conserved strengthens both your efficiency and profitability.

The choice is clear: stick with outdated, wasteful processes or embrace smarter, greener manufacturing that directly benefits your bottom line. The tools are here, the results are proven, and the time to act is now. By moving forward, you position your business for sustainable success and a lasting competitive advantage.

FAQs

Rising energy prices stifle investment

According to the report, nearly 40% of businesses have been forced to cut back on investment because of a sharp rise in energy prices. The issue spans industries, from chemical producers to pubs and restaurants, all of which are struggling to cope with energy costs that remain significantly higher than pre-crisis levels. Business electricity costs in the UK are currently 70% above those seen before Russia’s invasion of Ukraine, while gas prices have climbed 60% during the same period.

The report warns that without targeted actions to lower energy costs, “the risk of job losses, production cuts, plant closures and offshoring will increase.” Additionally, the UK’s ageing gas and electricity networks, coupled with outdated regulations governing energy supply, are exacerbating the problem.

Competitive disadvantage in global markets

The UK’s industrial energy prices are among the highest in the developed world, standing nearly two-thirds above the median of countries in the International Energy Agency (IEA) and the highest among G7 nations. Medium-sized businesses in particular face electricity costs that are around double the EU median, according to the report. While non-domestic gas prices are in line with those in the EU, they remain considerably higher than in nations like the United States and Canada.

“This acts as a brake on ambitions for economic growth”, the report said. It also noted that businesses are being deterred from investing in the transition to clean energy, which could bring long-term benefits and align with the UK government’s net-zero agenda.

Calls for reform and government action

The CBI and Energy UK are urging ministers to take decisive action, including a comprehensive review of the UK’s energy needs and reforms to improve efficiency across gas and electricity networks. This review is seen as essential to spur investment and reverse the trend of deindustrialisation that is already becoming evident in some sectors.

Louise Hellem, chief economist at the CBI, emphasised the urgency of the situation, pointing to its impact on key industries. “You can see it already in the chemicals industry, which has seen several closures”, she said, describing the current period as a “pivotal moment” for shaping the UK’s industrial strategy.

The report highlights that even progress made to reduce energy costs for some businesses has been limited in scope. Last year, the government introduced measures to reduce electricity prices for 7,000 of the country’s heaviest energy users by up to £40 per megawatt hour, a move aimed at making costs more competitive globally. However, Dhara Vyas, head of Energy UK, expressed concern that this assistance does not extend far enough. “Thousands of businesses outside the ringfence would continue to be hampered by high energy bills”, she said.

Industry demands broader solutions

While recognising the government’s efforts to lower domestic energy costs, Vyas stressed that the support provided to industrial users was insufficient and carried costs for other bill payers. She underscored the need for systemic reform of the energy market. “Lowering prices for all businesses is fundamental to the UK’s growth story”, she said. “Our aim will not be just about how to reduce bills. It will be the first of its kind to take a fundamental look at the energy market and the regulations to see how it can become more effective.”

The government has acknowledged the issue, with a spokesperson highlighting ongoing efforts to tackle the energy cost crisis. “We’ll shortly publish the response to our consultation on eligibility for the British Industrial Competitiveness Scheme, which will reduce electricity bills by up to 25% for over 7,000 businesses, and our Supercharger package of support will cut businesses’ electricity costs by up to £420m per year”, they said.

Warning signs for UK trade

The challenges facing UK manufacturers are reflected in trade figures, with a record £248.3bn deficit in goods reported for 2025 – £30.5bn higher than the previous year. While a growing £192bn surplus in services partially offset this gap, the data underscores the vulnerability of the UK’s manufacturing base.

As the government pushes forward with its industrial strategy, the calls for broader and more inclusive reforms to address energy prices will likely shape the next phase of the UK’s economic policy. Without decisive action, the UK risks a decline in industrial output and global competitiveness, potentially jeopardising its position as a manufacturing powerhouse.

Read the source

This article unpacks Hzer’s transformative insights into energy design, sustainability, and leadership, offering actionable lessons for production managers, quality engineers, and operations directors eager to modernise their operations.

Rethinking Sustainability: From Expense to Opportunity

For many leaders in the metals and steel industry, sustainability is perceived as an unavoidable expense, driven by regulatory compliance or public relations. Hzer challenges this assumption head-on: “Sustainability can actually be very profitable from day one”, he asserts.

The key lies in reframing sustainability efforts as a driver of efficiency and innovation. For example, integrating renewable energy and efficient design systems into industrial operations can reduce operational costs and improve energy usage effectiveness - core metrics for profitability in high-energy-consumption industries.

Insight for leaders: By focusing on total cost of ownership rather than upfront expenses, sustainability investments can deliver measurable returns. For instance, using AI-powered cooling systems or renewable energy solutions can reduce long-term costs while boosting eco-compliance.

The Energy Grid Conundrum: A Catalyst for Innovation

European markets, especially key industrial hubs such as Frankfurt, London, Amsterdam, and Paris, are grappling with energy grid constraints. These regions frequently struggle to meet rising energy demands, driven by industries like metals and steel as well as data centres. While grid shortages may seem like obstacles, Hzer views them as opportunities for smarter strategies.

He explains, “In most of these locations, there simply is no power anymore, which leads to different strategies to deal with it.” Companies are innovating by:

• Designing microgrids to integrate renewable energy sources like solar and wind.
• Exploring new locations outside urban areas where power availability is higher.
• Collaborating with local governments to optimise zoning and infrastructure.

Practical example: Some projects are linking data centres with solar parks and battery storage to create “energy centres”, where renewable power sources work seamlessly alongside traditional grids. These systems not only lower costs but also reduce strain on overburdened infrastructure.

Sustainability in Practice: A Story of Synergy

One of the most compelling insights from Hzer’s interview is the potential for synergy between energy producers and energy-intensive industries. For instance, he describes a collaboration with a solar park operator who faced penalties for producing excess electricity during summer. At the same time, data centres required maximum energy for cooling during these hot months.

By co-locating the solar park and data centre, both sides benefited:

• The solar park avoided penalties by supplying energy directly to the data centre.
• The data centre gained access to cost-free or low-cost energy during peak demand.

This partnership illustrates how sustainability initiatives can simultaneously deliver financial and environmental advantages, debunking the myth that eco-friendly practices are inherently unprofitable.

Transforming Leadership through Constraints

It’s not just technical or financial systems that need innovation; leadership mindsets also play a crucial role. Hzer emphasises resilience and adaptability as critical traits for leaders navigating constraints. “Every imperfection has an opportunity”, he says.

Rather than accepting limitations at face value, Hzer encourages leaders to take a 360-degree view of challenges, seeking creative ways to turn barriers into benefits. For example:

• Collaborating across industries to share knowledge and resources.
• Engaging local governments and stakeholders to develop mutually beneficial solutions.
• Exploring emerging technologies like AI to optimise operations.

Leadership, Hzer argues, requires patience, persistence, and a long-term vision. By adopting these qualities, executives can unlock innovative opportunities and drive meaningful change.

Redefining the Future: From Data Centres to “Energy Centres”

One of the most forward-looking ideas Hzer introduces is the evolution of data centres into “energy centres.” With the rise of AI and other energy-intensive applications, the focus is shifting from just housing data to managing massive amounts of energy.

For industries like metals and steel, which similarly rely on high energy consumption, the concept of energy centres offers a roadmap for the future. By investing in systems that prioritise energy efficiency and sustainability, businesses can stay ahead of regulatory pressures and market trends while improving profitability.

Key takeaway: The future of industry is not just about managing processes - it’s about managing energy. Leaders who understand this shift will position their organisations for long-term success.

Key Takeaways

1. Sustainability drives profitability: Energy-efficient systems and renewable integration can lower operational costs.
2. Turn constraints into innovation: Grid challenges can lead to creative solutions like microgrids and energy centres.
3. Synergy creates value: Collaborations between energy producers and high-consumption industries can yield mutual benefits.
4. Leadership requires resilience: Adopt a mindset that rejects “impossible” and seeks opportunities in every challenge.
5. Focus on total cost of ownership: Investments in sustainable technologies may increase upfront costs but deliver significant long-term savings.
6. Energy is the key resource: Industries should prepare to manage energy as strategically as they manage production.
7. AI enables efficiency: Use AI and automation to optimise energy use and streamline operations.
8. Engage stakeholders early: Building relationships with governments, communities, and partners can accelerate project timelines and ensure alignment.

Conclusion

Alexander Hzer’s insights offer a powerful roadmap for leaders in the metals and steel industry who are eager to embrace sustainability without compromising profitability. By seeing energy design as an untapped opportunity, businesses can innovate, cut costs, and meet sustainability goals simultaneously.

The path forward requires not just technical innovation but also bold leadership and strategic thinking. By focusing on energy management, collaboration, and long-term vision, industry leaders can turn today’s constraints into tomorrow’s competitive advantages.

For the forward-thinking manager, engineer, or director, the message is clear: sustainability is not an obstacle - it’s a tool for transformation.

Source: “Why sustainability can be profitable from day one” - The CEO Magazine, YouTube, Jan 21, 2026 - https://www.youtube.com/watch?v=XXlhL9vC6RQ

Here’s the truth: AI isn’t here to replace your team; it’s here to stop the waste. From optimising furnace operations to predicting energy demand spikes, AI tools are slashing costs, cutting emissions, and making production smoother. Steel plants have saved millions annually, while aluminium casthouses are closing efficiency gaps with real-time data.

What’s the difference between the old way and the smart way?

If you’re tired of sky-high energy bills and inefficiencies slowing you down, it’s time to rethink your approach. Let’s break down how AI is transforming metals manufacturing - and how you can start saving today.

Case Study 1: Steel Manufacturing - Furnace Operations

The Problem: Inefficient Furnace Processes

Furnaces are a major cost driver in steel production. Electric arc furnaces, for instance, consume anywhere from 350 to 700 kWh per tonne of steel, while blast furnaces often suffer from inconsistent coke usage due to reliance on manual monitoring and operator judgement [1]. Many facilities still depend on outdated methods like shift handovers or monthly utility reviews to identify inefficiencies. By the time someone notices that Furnace #2 is burning 15% more fuel than Furnace #1 under identical conditions, weeks of costly fuel waste have already occurred [5]. Traditional systems often fail to catch subtle problems, such as worn burner tips, poorly calibrated air–fuel ratios, or the precise timing when steel reaches its ideal temperature. This oversight can leave 15–30% of potential savings untapped [1].

The AI Solution: Real-Time Monitoring and Process Control

AI has revolutionised furnace operations by providing real-time monitoring and control. In November 2024, Spartan UK’s Gateshead plate mill introduced Deep.Meta’s “Deep.Optimiser” platform. This system, powered by a digital twin built on 40 years of production data, alerted operators the moment steel hit its optimal temperature. This eliminated guesswork and avoided unnecessary heating [4]. Similarly, in April 2024, ArcelorMittal Asturias in Spain deployed an AI-driven image-based system on a 1.2 MW industrial burner. Using colour cameras and neural networks, it estimated flue gas oxygen levels with 97% accuracy, ensuring efficient use of low-calorific blast furnace gas [7]. Over in Wu’an, China, Puyang Steel integrated infrared thermal imaging and 6-axis robotic arms into its No. 2 Converter in 2023. The AI system analysed molten steel composition in real time, automatically initiating slag removal.

“Previously, we relied on experience to determine slag removal timing. Now, the AI analyses molten steel composition in real‑time, triggering automatic operations 15 minutes faster per heat.” [6]

These advancements have delivered measurable improvements in efficiency and performance.

The Results: Energy and Emission Reductions

The impact of AI on furnace operations has been striking. At Spartan UK, energy use dropped by 24 kWh per tonne of steel, while CO₂ emissions during reheating fell by 5% [4]. ArcelorMittal’s system cut energy consumption by 52.8 kWh per tonne and reduced CO₂ equivalent emissions by 13.2 kg per tonne of steel [7]. Meanwhile, Puyang Steel’s faster slag removal process resulted in ¥4 million in annual alloy savings (around £440,000) across three production lines [6]. On a broader scale, AI-driven furnace management has led to an 18% drop in CO₂ emissions and a 16% reduction in overall energy intensity [5]. One integrated steel producer, after implementing AI across its blast furnaces and rolling mills, saved £3.3 million annually by lowering energy intensity from 22.5 GJ/tonne to 18.9 GJ/tonne - achieving full payback in just 14 months [5].

These results highlight a major shift in steel manufacturing, combining cost savings with a more efficient and environmentally conscious approach.

Case Study 2: Aluminium Casting - Predictive Analytics

The Problem: Energy Loss in Casting Operations

Aluminium casthouses are notorious for their high energy demands. Processes like high-pressure diecasting, annealing, and continuous casting often run inefficiently, especially at elevated temperatures. At Ryobi Aluminium Casting in Carrickfergus, Northern Ireland, engineers uncovered a 13% energy efficiency gap between two diecast machines that were supposed to perform identically [12]. The issue? Data was scattered and unanalysed, leaving no centralised system to compare energy use against production output. Similarly, annealing furnaces used for heat-treating aluminium coils were set to fixed temperatures, ignoring real-time conditions. Without precise monitoring, operators kept heat levels unnecessarily high, wasting fuel [8]. These inefficiencies highlighted an urgent need for smarter energy management.

The AI Solution: Predictive Chemistry and Recovery Models

AI stepped in to revolutionise the way energy was managed. At EPFL and Novelis Switzerland, a machine-learning tool was trained on CFD and experimental data to predict heat transfer coefficients in an ACL furnace. This enabled the optimisation of a flue gas recycling strategy, using exhaust gases from hotter furnace zones to preheat cooler ones, significantly reducing fuel waste [8].

In the US, Arconic collaborated with Lawrence Livermore National Laboratory to apply machine learning to Direct Chill (DC) casting. By combining casting simulation data with numerical optimisation, they could predict defects like end cracks in a fraction of the time. Tasks that used to take days were now completed in minutes, allowing Arconic to produce ingots with fewer defects and less need for energy-draining recasting [9].

At Ryobi, engineers developed a custom dashboard that unified electricity usage data, production output, and tariff information. This made it easier to identify inefficiencies and act on them [12].

“Our factories generate vast amounts of data with the potential to unlock efficiency, cost savings and innovation. [But to achieve this] we needed a one stop shop for all our data, a modern, bespoke digital tool where we could visualise the opportunities to be more efficient, profitable and environmentally responsible.” [12]

• Ciarán Maxwell, Low Carbon Project Lead, Ryobi

These advancements demonstrate how AI can modernise aluminium casting, aligning economic goals with environmental responsibilities.

The Results: Lower Fuel Use, Lower Costs

The impact of these AI strategies has been substantial. Novelis’s flue gas recycling system reduced fuel consumption in aluminium annealing furnaces by 20.7% [8]. Arconic’s predictive modelling cut the ingot scrapping rate, potentially saving the US aluminium industry an estimated $60 million annually (around £47 million) in energy costs [9]. At Ryobi, Ciarán Maxwell anticipates a 20% reduction in overall energy consumption within the first year, freeing up funds to invest in renewable technology [12].

AI also boosted production quality. Specialised aluminium casting saw yields jump from 82% to 97%, an 18.3% improvement, while high-volume rod casting reduced scrap rates by 75%, dropping from 6% to just 1.5% [10]. Defect detection became faster and more accurate, with a 96% success rate compared to 72% with manual checks, and inspection times fell from 10 seconds to just 2 seconds [10].

For perspective, even a 1% reduction in scrap at a typical high-pressure diecasting facility can prevent 600,000 kg of CO₂ emissions annually [11]. These advancements are not only about cutting costs - they represent a shift towards cleaner, more efficient operations.

AI in Energy Management and Procurement

Energy Forecasting with AI

Predicting furnace demand spikes has become far more precise with the help of advanced AI models like Random Forest, k-NN, and Gradient-boosting. These tools analyse historical consumption patterns to achieve forecasting accuracy exceeding 97% [1][13]. This means operators can foresee cost peaks 4 to 20 weeks in advance [6][15], making it possible to shift energy-intensive operations, such as Electric Arc Furnace melting, to off-peak hours when tariffs are lower.

Dynamic Demand Response systems take this a step further by automatically adjusting production schedules based on fluctuating tariffs and the availability of renewable energy [13]. For instance, a steel manufacturer using C3 AI at a hot roll mill reduced utility demand charges by 40 MW per month over five months, while also increasing on-site power use by 1.8%. This combination led to an impressive $14 million in annual energy savings [2]. Similarly, between 2024 and 2025, a 2.4-million-tonne integrated steel manufacturer implemented Oxmaint’s AI platform, achieving $4.2 million in annual savings with a payback period of just 14 months [5].

AI also supports smarter procurement strategies by tracking a “Green Electricity Index” (GEI), enabling manufacturers to weigh both cost and the share of renewable energy in their energy mix [13]. Together, these forecasting and procurement capabilities create a solid foundation for on-site energy optimisation, reducing waste and costs.

On-Site Energy Optimisation

Once energy demand is accurately forecasted, AI fine-tunes operations on-site for maximum efficiency. Real-time monitoring reveals equipment issues, such as damaged furnace doors or worn-out burner tips, that would otherwise go unnoticed during manual checks [14][5]. A compelling example comes from ArcelorMittal’s Industeel factory in France, which saved €340,000 in just 12 months by deploying an Energy Management & Optimisation System (EMOS). This system processed 3,000 data points per second to monitor reheating furnaces, enabling operators to quickly identify and fix problems causing energy inefficiencies [14].

“Within the first month of real-time monitoring, we discovered our #2 reheating furnace was consuming 15% more fuel than #1 under identical conditions. That single finding paid for three months of the system cost.” [5]

• Plant Energy Manager, Integrated Steel Manufacturer

AI also synchronises reheat furnaces with rolling mill operations to minimise thermal losses and reduce demand spikes [1]. At Xinjin Steel in Wu’an City, China, an AI-driven system employing LSTM time-series algorithms for 4-hour advance forecasting optimised chemical dosing in water treatment. This improvement saved 3.8 million kWh annually - enough to power 1,200 households for a year [6].

Through intelligent scheduling, peak demand charges can be reduced by 18–22% [1], while predictive load optimisation can lower overall energy costs by 15–25% [1]. For mid-sized integrated steel plants, these strategies translate to annual savings of $2 million to $5 million (approximately £1.6 million to £3.9 million) [1].

Comparing Results Across Operations

Steel vs Aluminium: Key Metrics

AI is proving its worth in energy savings across both steel and aluminium industries, though the focus and scale of improvements vary. In Wu’an City, China, 12 major steel enterprises managed to cut energy use per tonne by 18%, dropping from 562 kg SCE to 461 kg SCE, while also saving over ¥200 million annually [6]. Spartan UK’s Gateshead plate mill provides another example, achieving a 24 kWh per tonne energy reduction and a 5% cut in CO₂ emissions through AI-driven optimisation [4][12].

For aluminium casting, the results are equally promising. Ryobi’s Carrickfergus facility used AI-powered dashboards to uncover a 13% energy efficiency gap between two identical diecasting machines. Ciarán Maxwell, Ryobi’s Low Carbon Project Lead, anticipates that full implementation could “reduce our overall energy consumption by up to 20% in the first year” [12]. While steel operations focus on furnace optimisation and slag removal, aluminium manufacturers target machine-level benchmarking and improving high-pressure diecasting efficiency [4][6][12].

The benefits extend beyond energy savings. Steel plants in Wu’an saw labour efficiency jump by 60% - from 42 to 67 tonnes per man-hour - and equipment downtime drop by 65.6%. Meanwhile, Ryobi’s optimised diecasting machines achieved an 11% improvement in Overall Equipment Effectiveness (OEE) [6][12]. On a global scale, deploying the Deep.Optimiser platform across 1,600 steel plants could slash CO₂ emissions by 500 megatons annually, representing a 20% reduction in the total emissions tied to steel production [4][12].

These results highlight both the financial appeal and environmental advantages of AI-driven solutions.

ROI Timelines and Environmental Benefits

AI investments in heavy industry deliver not only operational efficiencies but also quick financial returns and environmental improvements. In the steel sector, payback periods typically range from 6 to 14 months, making these systems one of the fastest-returning investments available [1][5]. For example, a 2.4-million-tonne integrated steel manufacturer using Oxmaint’s AI platform saved £3.3 million annually, achieving payback in just 14 months [5]. Similarly, at a hot roll mill, C3 AI’s energy management system delivered £11 million in yearly savings by cutting utility demand charges by 40 MW per month over five months [2]. Mid-sized integrated steel plants report annual savings between £1.6 million and £3.9 million through intelligent scheduling and predictive load management [1].

The environmental impact is just as impressive. Oxmaint’s deployment reduced CO₂ emissions by 18%, lowering output from 1.92 tonnes to 1.58 tonnes of CO₂ per tonne of steel [5]. Xinjin Steel’s AI-powered water treatment system saved 3.8 million kWh annually - enough energy to power 1,200 households for a year [6]. As Tarun Mathur, ABB’s Global Digital Lead for Metals, puts it:

“AI is making sustainability and decarbonisation more profitable by linking carbon reduction with operations excellence” [3].

These examples show that AI isn’t just about improving efficiency - it’s also a powerful tool for aligning profitability with sustainability.

Conclusion: Modernise Your Operations

The examples above leave no doubt: AI-driven energy optimisation delivers real, measurable benefits for metals manufacturers. These advancements not only sharpen your competitive edge but also deliver a return on investment in as little as six to 14 months [1][5].

The contrast between manual processes and AI-powered systems is stark. While traditional energy management methods hit only 60–70% of the theoretical efficiency ceiling, AI systems consistently achieve 92–98% [1]. A veteran operator at Spartan UK, with three decades of shop floor experience, shared with Deep.Meta founder Osas Omoigiade that the AI tool increased efficiency fivefold while significantly reducing human errors [4]. This is the difference between educated guesses and precise, data-backed decisions.

Getting Started with AI

You don’t need to overhaul your entire setup to get started. Begin with a four-week audit to assess your energy consumers and metering infrastructure, creating a baseline normalised to production [5]. From there, focus on quick, high-impact fixes in the first 90 days - such as sealing compressed air leaks, addressing furnace imbalances, and benchmarking machine performance - to gain early support from your team [5]. Using real-time dashboards can also make energy efficiency a shared point of pride on the shop floor [5].

For metals manufacturers, platforms like GoSmarter are tailored to handle the tedious and time-consuming parts of production. Whether it’s digitising mill certificates, refining scrap rates, or managing production schedules, these AI tools turn mountains of data into clear, actionable insights - helping factories run smoother, cleaner, and without unwelcome surprises. The tools are here. The results are proven. Taking these steps ensures you stay ahead in a fast-changing industry.

FAQs