Artificial intelligence is already becoming part of the industrial landscape. Most companies have moved beyond merely asking whether they should use AI and are experimenting with it somewhere in their organisations. But experimental pilots are not the same as operation-wide impact.
McKinsey & Company research shows that 88% of organisations use AI in at least one business function, yet only 6% qualify as AI high performers, reporting significant value and an EBIT (earnings before interest and taxes) impact of 5% or more from AI use.
So, what separates an AI pilot from AI that can deliver value at scale? For industrial companies, the answer increasingly comes down to context.
Giving AI an understanding of the physical world
Oil and gas companies already have a wealth of data generated by industrial software, automation technologies and systems across engineering and operations. But data alone isn’t enough.
For AI to support meaningful decisions, that data needs context. What does a change in operating conditions mean for a particular piece of equipment? How might an adjustment in one area affect performance somewhere else? Could a proposed action introduce a new reliability or operational risk?
Digital twins can provide that context by bringing engineering models, operational data and physical behaviour together. This gives users a better understanding of not just what is happening, but why it is happening and what could happen next.
Combining this foundation with AI creates opportunities to analyse complex systems, evaluate potential changes and accelerate decisions while maintaining a connection to real-world behaviour.
Putting AI to work in engineering
Siemens is taking this approach by embedding AI and co-pilots into the industrial technologies where knowledge is created, used and shared, including lifecycle management, product design, manufacturing operations, and simulation and testing software.
Technologies including reduced-order modelling (ROM) AI and Physics AI work alongside digital twin capabilities within the Siemens Xcelerator portfolio to help engineers analyse complex information faster.
Physics AI, for example, can accelerate simulation by up to 1,000 times. That speed gives engineers the opportunity to explore more scenarios and understand the potential effects of changes while maintaining a connection to the physics governing the real system.
The potential goes beyond making existing software faster. Combining AI with engineering knowledge and digital twins can help organisations accelerate analysis, evaluate potential solutions and understand their consequences before implementing changes. In this way, AI becomes part of the engineering decision-making process rather than a standalone technology stacked on top of it.
Moving beyond individual tasks
Embedding AI in engineering addresses one part of the challenge. Scaling it across an industrial organization presents another.
Industrial information can span multiple systems, disciplines and locations. Using that information effectively requires context and guardrails, because companies also need to maintain control over their data and govern how AI uses it.
Siemens is addressing this challenge with Intelligence Center X, a platform designed to help companies move from individual AI initiatives toward adoption at scale.
Intelligence Center X brings together enterprise knowledge graphs, purpose-built machine learning and low-code application development. Graph Studio connects information across sources and domains to give AI agents structured industrial context, while AI Studio provides predictive and prescriptive intelligence based on that context. Mendix enables companies to turn those capabilities into governed AI and agentic applications without giving up control or ownership of their underlying data.
The goal is to give AI agents the context to understand relationships across industrial data rather than relying on a general-purpose model to infer what that information means.
From AI recommendations to trusted decisions
Giving AI context is one part of the equation. The ability to evaluate its recommendations before acting on them adds another. This is where AI and digital twins can become particularly valuable together.
With access to current, contextualised operational data, AI can help companies evaluate potential actions through a digital twin before those actions are implemented. This can provide another layer of insight when decisions affect complex industrial assets and operations.
The approach is already showing up in the energy industry.
Assystem, for example, combined as-designed plant models with real-world sensor data to create a digital twin that helps engineers understand structural integrity and optimize inspection and maintenance schedules. Siemens’ AI technology provides the analytical backbone, supporting the flow of information between the physical plant and its physics-based models.
Aker Solutions embraced engineering simulation and predictive engineering analytics to ensure their equipment is optimised through the complete lifecycle. The company delivered its digital twin “Thermal Insight” to manage thermal performance even in locations that are largely inaccessible.
Siemens Energy manages a broad fleet of global turbines and managing spare parts catalogues and specific customer configurations for each turbine requires manual processing and labour. Knowledge Graphs – built by subject matter experts and augmented by intelligent machine learning algorithms and application development – are the ideal solution to organise, manage and query its machine data structures across the fleet.
Examples like this demonstrate how connecting AI with engineering knowledge, operational data and physical models can help turn information into more useful insight.
Moving from AI adoption to AI impact
As industrial AI matures, success is likely to depend on more than the quantity of models, copilots or agents a company deploys.
The bigger measure will be what those technologies enable people to accomplish. Can AI work with the industrial context behind the data? Can recommendations be evaluated against the behaviour of the physical system? Can successful applications move beyond isolated pilots while companies maintain governance and control?
Digital twins, embedded engineering AI and scalable AI platforms are bringing these pieces together, helping industrial companies move from experimenting with AI toward broader adoption across engineering and operations. Ultimately, the opportunity isn’t about using AI as a tool. It’s embedding it seamlessly to create faster engines, smarter execution and trusted outcomes.
About the author
Hermias Hendrikse leads Global Industry Development for Energy, Chemicals and Infrastructure at Siemens Digital Industries Software (DISW), where he helps energy and process companies turn digital twin and AI technology into measurable engineering and operational gains. Over 15 years with Siemens DISW technology, he has delivered more than 60 industrial simulation projects across Africa, the Middle East and the United States, progressing from implementation engineer to global strategic leadership.
(Main photo: Siemens)
