Why AI matters at this scale

Suncoke Energy is a major independent producer of metallurgical coke, a critical fuel and reductant for blast furnaces in the global steel industry. With operations centered on large, capital-intensive coke-making facilities, the company's core business involves precisely heating coal blends in slot ovens or heat-recovery ovens to produce high-quality coke. For a mid-sized industrial player like Suncoke (501-1000 employees), operational excellence, cost control, and asset reliability are paramount for competitiveness. At this scale, the company has sufficient operational complexity and data generation to benefit from AI but may lack the massive R&D budgets of conglomerates. AI presents a lever to achieve step-change improvements in efficiency, yield, and predictive capability without proportionally increasing headcount or capex.

Concrete AI Opportunities with ROI Framing

1. Predictive Maintenance for Coke Ovens: Coke ovens are subjected to extreme thermal and mechanical stress, leading to refractory wear and potential failures. Unplanned downtime is extraordinarily costly. An AI model trained on historical sensor data (temperatures, pressures, gas flows) can predict refractory issues weeks in advance. The ROI is direct: shifting from reactive to planned maintenance avoids production losses that can exceed $1 million per day per battery, while extending oven life.

2. Real-Time Combustion and Thermal Efficiency: The coking process is highly energy-intensive. AI-driven control systems can continuously analyze real-time data from thousands of points across a battery to optimize combustion in individual heating flues. By dynamically adjusting air and fuel gas flows, the system can maintain ideal coking temperatures while reducing natural gas consumption by 3-5%. For a facility spending tens of millions annually on fuel, this translates to millions in saved costs with a rapid payback period.

3. AI-Powered Blend and Quality Forecasting: Coke quality (strength, reactivity) depends on the blend of various coals. Machine learning can analyze historical blend recipes, raw coal properties, and final coke quality test results to recommend optimal, cost-effective blends that meet customer specifications. This reduces reliance on expensive premium coals and minimizes quality give-away, improving margin per ton. The ROI comes from both raw material savings and reduced risk of off-spec product.

Deployment Risks Specific to This Size Band

For a company of Suncoke's size, key AI deployment risks are multifaceted. Technical Debt & Integration: Legacy control systems (SCADA) and enterprise software (e.g., SAP) may not be easily accessible for real-time AI models, requiring middleware and creating integration complexity. Talent Gap: The company likely has strong process engineers but may lack in-house data scientists and ML engineers, creating a dependency on external vendors and potential misalignment with core operational needs. Pilot Scaling: A successful pilot on one oven battery may face challenges scaling across different plant sites with varying equipment and data landscapes, requiring customized adaptation. ROI Justification: In a cyclical industry, capital allocation is scrutinized. AI projects must compete with traditional capital projects for funding, requiring exceptionally clear, quantifiable ROI projections tied to core operational KPIs like throughput, fuel rate, and maintenance cost. Failure to tightly scope initial use cases can lead to abandoned projects and reinforced skepticism.