by John Hutchins, Director Client Services, Slalom
In the past few years, the world has seen an AI revolution across all sectors, adding to an era already defined by digital innovation. Businesses across all industries are making the most of AI technology and exploring its possibilities, but the energy sector – as one of the world’s oldest industries – is failing to successfully adopt the technology, with only 33% of Utility & Energy businesses worldwide exploring AI.
The case for AI in the energy sector
AI can serve a wide variety of uses all the way up and down the energy value chain. In fact, research has found there to already be over 50 identified uses for the technology in the sector. Its algorithms can allow for predictive maintenance of generation and network assets, as well as load forecasting. It can also enable optimisation of energy distribution and grid management by forecasting demand patterns, identifying congestion and optimising energy flows, as well as being of use in customer servicing. Not to mention, it can also automate any admin processes.
However, even with these significant use cases, the energy industry is failing to successfully adopt the technology due to specific barriers that arise in this industry. At its heart, energy is a simple process with little fundamental change – even wind and water have been used for powering activities for centuries. With minimal differentiation nor new competition and disruption, cost is king, so the energy value chain and all its processes have been split into specialisms to concentrate on optimisation. This gives rise to a host of barriers to AI.
Barriers to adoption
AI thrives on demystifying complex relationships across processes. The deep silos in energy organisations don’t easily allow cross-collaboration to realise AI benefits. Information systems have historically not been valued as ‘core competencies’ to drive efficiency in energy processes. Hence, they’ve been outsourced or kept in yet another silo and the understanding of these systems has not been nurtured or retained. As a result, data and data skills – the lifeblood of AI – are undervalued and poorly maintained assets. In fact 40% of businesses in the energy sector find it difficult to hire data scientists, showing a lack of focus and interest. Significantly, data is not embedded in business processes – a huge hurdle because AI’s value increases with trial, error and learning – which means it must be an integral element of teams and products to realise benefits.
Risk and fear of failure is also high. Health and safety risks are real for workers, consumers and the environment, but legislation and regulation have followed the pattern of the industry; focusing on existing processes and prescribing ways things should be done, constraining the ability to test and learn with AI.
While these issues may seem too grand to change drastically, there are steps the industry can take to better and more successfully adopt AI.
Looking ahead to successful AI adoption
The industry needs to state a bold vision to transform and lead from the top. The deep organisational barriers in the sector require bold leadership and sponsorship to overcome. Therefore a leader has to emerge, which could be in the shape of a newly formed Chief AI Officer role, following in the footsteps of industries like finance and technology, or through shareholder push. Energy leaders must accept that data and AI will be part of BaU – not just innovation.
There also needs to be investment into data and platforms that bring siloed data together and, until such time as data becomes embedded, new teams that bring data and business specialists together to look at cross-cutting use cases. Taking your best operational practitioner out of their usual business and into this field may be difficult in the short term, but will prove to be essential. This can also mean investing in and valuing specialists who understand the business data and can bridge the business / IT divide, while cultivating trust in the data.
As a foundational point, the energy industry needs to upskill and enhance the existing community of its data professionals, recognising that a range of roles and career paths exist from data engineering, to data management, to data science to infrastructure engineering. It’s not as simple as ‘data analytics’, especially when it comes to AI; a comprehensive data strategy shaped around the people is required, which involves investing in data literacy across the workforce. Ultimately, there should be no divide between ‘business’ and ‘technology’ in a modern organisation.
Finally, data needs to be a normal part of teams and processes. Training models, testing recommended changes to processes and learning the result need to become a natural part of daily business and is crucial for successful integration of AI.
Final thoughts
The energy industry has been slow to make use of new technologies over the years and unsurprisingly there has been no real deviation from this when it comes to AI adoption. Although this is not hugely surprising, it should be corrected – for the good of consumers and, crucially, the planet. Transformative technology such as AI holds so many important benefits and potential use cases which are constantly being developed and the opportunity for innovation and commercial success is ripe. To ignore it would be more than just a fleeting opportunity missed.
