Mark Simpson, Co-Founder, WeBuild-AI
AI adoption is accelerating across the energy and utilities sector, but many initiatives remain trapped in experimentation and fail to deliver impact at scale. The UK government’s new vision for an AI-enabled clean energy system acknowledges that adoption continues to be held back by barriers including fragmented data, technical integration, regulatory clarity, skills and organisational culture.
This underlines a critical challenge for the sector. AI can only generate operational value when it is supported by robust data foundations, integrated systems and the right organisational environment. Looking ahead, energy and utility organisations need to focus less on the capabilities of individual AI models and more on the foundations that enable them to deliver value at scale. That means ensuring their data, systems and governance frameworks are ready to support adoption.
Generic models lack critical industry context
Today’s AI models are increasingly capable, but most general-purpose systems have been trained primarily on broad, publicly available information. That makes them useful for everyday tasks but less reliable when applied to specialised energy infrastructure, proprietary assets and organisation-specific processes.
The issue is that much of the knowledge and information that’s important in energy and utilities is not available on the public internet. It sits in engineering documents, asset databases, legacy maintenance systems and the experience of employees who have worked with infrastructure for years.
As organisations recognise the value of this domain knowledge, the market is moving towards more specialised AI. Gartner predicts that, by 2027, organisations will use small, task-specific AI models three times more often than general-purpose large language models, reflecting the growing demand for sector-specific context.
Without access to proprietary knowledge, a generic model cannot fully understand an organisation’s operating environment and will produce answers that miss essential technical, regulatory or operational context. In a sector where decisions can directly affect critical infrastructure, those inaccuracies can diminish trust and constrain AI’s operational value.
From experimentation to enterprise impact
Moving from successful pilots to organisation-wide adoption requires more than access to the right technology. Organisations need a clear strategy for applying AI where it can create operational value, backed by the data, processes and expertise required to integrate it into day-to-day workflows.
Grounding AI in the information employees use every day makes it more effective at supporting decision-making, improving productivity and producing outcomes aligned with operational requirements. This creates a stronger foundation for adoption than relying solely on models trained on external data.
However, success is not only about building more specialised models. Organisations must also identify where AI can have the greatest impact and prioritise specific operational challenges. Focusing on use cases such as navigating engineering documentation, supporting regulatory compliance or helping field teams retrieve knowledge can be more effective than attempting a broad, organisation-wide deployment. When AI is applied to well-defined use cases, it’s more likely to deliver measurable outcomes, build organisational trust and create momentum for wider adoption.
Trust requires explainability, not just accuracy
As AI begins to play a larger role in operational decision-making, organisations must give governance the same attention as model capability. For energy and utility organisations responsible for critical infrastructure, trust is non-negotiable and AI systems must meet strict standards for security, safety and accountability.
Organisations need to understand how AI-generated outputs are produced before they can inform maintenance planning, outage response or network management. That includes having visibility into the data sources used, the basis for recommendations and the means to validate outcomes.
Explainability and auditability are now operational requirements and cannot be viewed as compliance exercises. Without them, energy and utility organisations will struggle to move beyond isolated pilots because decision-makers cannot rely on systems they are unable to understand or verify.
Effective governance provides the oversight needed to assess outputs, maintain accountability and keep people responsible for final decisions and it also creates the confidence required to embed AI in everyday operations and scale its use.
Turning AI potential into operational impact
The future of AI in energy and utilities will depend on how effectively organisations adapt it to the realities of their operating environments. The organisations that turn pilots into scalable operational capabilities will combine industry expertise, proprietary data, targeted use cases and robust governance. These foundations ensure AI is deployed responsibly, trusted by the people using it and translated into measurable operational impact.
