Pragmatic use of AI to Facilitate Efficiency Improvement
The power of Generative AI to unlock vast data resources and deliver tangible value is acknowledged. The challenge is to evolve beyond incremental and piecemeal use of generic tools such as ChatGPT towards AI embedded within core operational systems and processes. With pressure on the power industry to rapidly accelerate production to meet the extraordinary escalation in demand companies need to rapidly assess how and where AI can be safely used to improve productivity whilst retaining critical regulatory compliance.
Achieving Safe Expansion
The planned growth in nuclear capacity is key to decarbonisation and energy security goals. With an endemic shortage of skilled nuclear engineers, however, a step-change in productivity is essential to accelerate innovation and build the new infrastructure required โ but power companies must also embed compliance and quality to minimise risk.
AI has an important role to play in accelerating innovation and adding capacity by unlocking the immense data resources in place across the energy industry. There are opportunities to transform everything from regulatory compliance and maintenance to innovation and workforce skills, with incremental gains in efficiency, collaboration and compliance quickly adding up to deliver quantifiable improvements in productivity.
Maximising Knowledge
Power companies have an extraordinary amount of information, with diverse data resources spanning hundreds, even thousands of projects. AI has the power to unlock the value of that data by improving visibility and extracting information in a way that can be used by diverse individuals across the business. The use of natural language models transforms accessibility, allowing individuals in different roles to engage with and gain value from the same data resources. Focusing on meaning rather than keywords, the use of natural language facilitates fast access to information without requiring complex or industry specific terminology. It can provide precise answers and/or highlight relevant sections of documents, using a conversational style if required.
Natural language models also support role-specific language, from the complexity required by engineers reviewing a new design concept to the legal team crafting contracts or the procurement team managing a complex supply chain. The entire organisation can gain value from cross-departmental insight and also collaborate in a way that supports different goals and information needs.
Workforce Transition
Furthermore, the use of natural language also supports the different generational experiences and expectations throughout the workforce. For an industry facing the challenges of retirement and the need to entice younger people, the use of natural language models introduces new opportunities for training, education and engagement. Training courses can be seamlessly designed in a more relaxed language style to appeal to a younger demographic. Workers can query User Guides and quickly access relevant summaries not only using their preferred terminology but also multiple languages, enhancing engagement across a multi-cultural workforce.
The speed and ease with which natural language models help individuals to locate pertinent information also improves employee onboarding and minimises the potential disruption associated with older colleagues being constantly asked simple questions. From immediate and up to date health and safety information, such as which PPE is required in a specific location, to simple contractual searches to understand specific supplier relationships, AI can build confidence as well as productivity for new recruits across the business.
Building Robust Data Resources
The power and value of AI is, of course, directly linked the quality and accuracy of data resources. Over the years, the use of manual metadata entry, which is both time-consuming and error prone, has undermined the quality of data for many power companies. The result is not only time wasted by individuals searching for the correct information but also a lack of data confidence. At best, this can lead to colleagues spending more time conferring to check accuracy; at worst, it can cause compliance breach.
Adding AI functionality to enterprise information management systems can transform the time taken to improve the quality of data resources by rapidly highlighting data inaccuracy. By analysing document content, AI can highlight potential version inconsistencies, reducing the manual effort required to check accuracy. While manual intervention is then required to reclassify and ensure metadata and tags match, such data improvement projects should never be required again: once AI is successfully embedded within core processes, automated validation and verification ensure data is correctly tagged, classified and stored.
By empowering staff to focus on skilled tasks, AI will safely support the improvements in productivity that are essential if the industry is to meet the ambitious expansion goals.
Enhancing Compliance
Not only is an AI-enabled data upload process incredibly accurate, it is also quicker and more efficient, reducing manual effort and allowing individuals to be far more productive. For example, new regulations can demand information updates across potentially hundreds of relevant documents. AI will immediately surface the appropriate documents and set priorities and timelines for updates, ensuring staff maximise their productive time.
Indeed, one of the most significant benefits of this enhanced data quality is regulatory compliance. In addition to the speed with which individuals in every department can access relevant procedures, the embedded validation and verification also transforms confidence that the individual is reading the latest, up to date information. From health and safety to environmental standards and critical infrastructure protection, transforming speed of access is key to reinforcing adherence, improving oversight and streamlining regulatory reporting.
Human in the Loop
AI is not, however, a panacea. It cannot be left to run unmanaged and unchecked. At every stage of the process, it is essential to include human oversight within AI-driven processes. By ensuring answers are verified and sources validated by professionals, especially within critical engineering and compliance contexts, power companies can achieve the tangible gains in quality and productivity, efficiency and collaboration required without incurring operational risks.
While enormously powerful, AI is not a replacement but a facilitator; a chance to maximising existing skilled resources and accelerate knowledge sharing. It is a tool that, used effectively, can reduce manual tasks, fast track access to information and, critically, ensure that information is up to date and valid. By empowering staff to focus on skilled tasks, AI will safely support the improvements in productivity that are essential if the industry is to meet the ambitious expansion goals.
By Andrรฉ Gunter,
Head of Technical Solutions at Idox