Johnson Matthey and Enginira collaborate to advance AI-enabled process engineering

Johnson Matthey and Enginira collaborate to advance AI-enabled process engineering

In one of the first deployments of agentic AI in process engineering, Johnson Matthey’s Catalyst Technologies licensing business put Enginira’s Facility Pro platform to work on real engineering data - improving the outputs in safety reviews and giving engineers back hours each week for higher-value work.

• Johnson Matthey’s Catalyst Technologies business completed a pilot with Enginira in April 2026, applying agentic AI to the engineering behind the plants it licenses to customers worldwide.

• The platform surfaced additional independent items for the engineering team’s review, complementing the findings of a completed HAZOP - strengthening design assurance and safety.

• Custom AI agents executed around 80% of selected end-to-end workflows autonomously with engineering oversight on the output, representing 7-12x gains in throughput on targeted manual tasks - increasing the high-value engineering each person can deliver.

• Engineers demonstrated the potential to reclaim on average around fifteen hours each per week (roughly 75 engineer-hours per week across the pilot team), enabling this time to be redeployed from manual tasks to higher-value design, optimisation and assurance work.

• Demonstrated how compliance and standards checks can now be run continuously and on demand rather than as one-off exercises, so issues are caught earlier and more often.

Engineering's AI moment

Across the economy, AI has reshaped how knowledge work gets done. Yet engineering - one of the most information-intensive disciplines - has remained largely untouched. Process engineering still depends on teams manually interrogating dense technical drawings, datasheets and standards to deliver complex projects, a way of working that has changed little in decades.

Johnson Matthey’s Catalyst Technologies licensing business set out to challenge that. The business produces the engineering packages that customers use to build their own plants, and as demand for licensed process technology grows, optimising how that engineering is delivered has become a strategic priority. To explore what was possible, Johnson Matthey partnered with Enginira to test how AI-enabled digital engineering could accelerate project delivery without compromising technical rigour.

Inside the pilot

Over the course of the pilot, Johnson Matthey engineers trialled Facility Pro across a range of engineering activities, using a subset of real engineering data held outside live customer projects. Rather than searching through static documents, engineers interrogated the underlying plant design directly - asking detailed technical questions, assessing engineering intent, checking compliance against internal standards, and producing structured engineering deliverables. Custom AI agents were configured to carry selected workflows from entirely manual to agent driven work with key human-in-the-loop checking.

The results pointed to a step change in what engineers can achieve with AI. On targeted manual workflows, custom AI agents delivered 7-12x gains in throughput and executed around 80% of selected end-to-end tasks autonomously, with average weekly usage at three hours per engineer the time saved on manual tasks can now be reinvested in higher-value design, optimisation and assurance work. Most striking, the platform improved the output of HAZOP recommendations for the team’s consideration to supplement a completed HAZOP, pointing to AI’s potential not only to accelerate engineering but to make it more thorough and to strengthen safety and design assurance.

What makes Facility Pro Different

Facility Pro is built specifically for engineering, going beyond the document-based and general-purpose AI tools that struggle with the structure and rigour of industrial design. The platform structures engineering information around the underlying plant design and its ontology as a knowledge graph which users can query, letting engineers interact directly with complex process systems and capturing engineering knowledge for future operation and design.

By combining structured engineering data with domain-aware intelligence, Facility Pro supports faster decisions, greater consistency and scalable delivery through automated end-to-end workflows and custom AI agents - turning hard-won engineering knowledge into a living, reusable asset rather than a static set of documents.

In their own words