Ada Mode - AI in Action for Industrial Decarbonisation

By Christopher Pilgrim and Anna Riley

Summary

At the June 2025 IDAIC meeting, Jack Lewis, co-founder of Ada Mode, presented an insightful case study exploring how AI and digital solutions are helping heavy industry decarbonise in practice. Focused on their recent work within the Solent Cluster, Lewis’s talk offered a clear example of how data-led optimisation can drive both emissions reductions and operational improvements in complex manufacturing environments.

Outputs

• Ada Mode hosted tool developed to enable optimised production planning and process investment to drive down energy consumption
• Revised production schedule realising a 6% reduction in energy demand, ~500Te CO2 per year and £190,000 saving

Background

Ada Mode specialises in developing AI and software for industrial operators, primarily in the clean energy and nuclear sectors. Their work includes supporting current UK nuclear plants and decommissioning operations. The focus of this case study is from  their work within the Solent Cluster’s Local Industrial Decarbonisation Plan (LIDP). The Solent Cluster, covering much of southern England’s industrial base, represents a significant source of industrial emissions – from oil refineries to chemical plants and maritime activities.

Project details

As part of the LIDP, Ada Mode worked directly with a major chemical manufacturer – a high energy consumer producing over 200 chemical products. Their goal: explore how AI and digital solutions could reduce energy demand across the site. Lewis outlined how Ada Mode tackled this by developing a bespoke AI-driven scheduling tool, delivered via a secure, cloud-hosted platform. This tool allowed the client to visualise energy consumption linked to production schedules and explore “what-if” scenarios for operational and investment decisions.

A key challenge was the complex, analogue nature of the site itself: decades-old plant equipment, mixed IT infrastructure, patchy data quality, and significant operational constraints. With no option to automate data flows, Ada Mode worked within these limitations - handling data manually and securely to develop their models.

Using regression-based machine learning models, Ada Mode linked historic production volumes, process energy consumption, and ambient conditions to predict gas and electricity demand across the site. From this, the tool could recommend optimised production sequences that met customer orders while reducing overall energy use – without requiring expensive capital upgrades.

Critically, the tool also supported investment planning, highlighting where targeted upgrades (such as dual-fuel burners or expanded solar arrays) would yield the greatest efficiency gains. As Lewis emphasised, the AI itself was relatively straightforward – the complexity lay in the industrial context, data integration, and user needs. Ensuring accessibility for both engineers and finance teams shaped the tool’s final design.

Conclusion

Ada Mode’s contribution to the Solent Cluster’s decarbonisation roadmap stands as a practical demonstration of AI’s role in industrial transformation. Lewis’s reflections at IDAIC made clear that while AI is a powerful tool, success depends equally on understanding people, processes, and the operating realities of industry.

Their work offers a replicable model for other industrial clusters: blending pragmatic digital adoption with strategic decarbonisation planning – and underscoring the need to meet industry where it is, not where technologists assume it should be.