Automation in Energy and Utilities
AI and ML are going to make the energy sectors more efficient and will address the current need for de-stressing the climate. An in-depth survey from the industry shows that the industry leaders are ready for it
The global energy and utilities sector is undergoing unprecedented change:
• The three “Ds” of decarbonisation, deregulation
and decentralisation are having a significant impact. Currently only 10% of UK’s electricity comes from coal-fired generators, and in 2019 the National Grid has logged more than 1,000 hours of coal-free electricity.
• The sector is moving from its conservative, regulated past to a new future where innovation is key. Its consumer base, which used to be largely passive, has now moved to a world of ‘prosumers’ who expect a sophisticated, service-based industry.
• Digitisation will be critical to capitalising on these shifts. Technologies, such as automation and artificial intelligence, are playing a pivotal role in managing
the balance between demand and supply, boosting efficiencies in all the entirety of the value chain, innovating the customer experience and transforming business models. The global research study, Reshaping the futur e:
Unlocking automation’ s untapped value , explored the intelligent automation landscape (by ‘intelligent automation’, it is implied a combination of rule-based technologies such as RPA and added intelligence through advanced analytics and artificial intelligence). Examining specific use cases and the benefits they can deliver, it drew on the views of more than 700 executives involved in implementing intelligent automation solutions. Building on what we learned from that cross sector program, this latest 2019 research takes a specific look at energy and utilities (oil and gas, electricity utilities, water utilities, energy services and electricity and gas utilities). We surveyed close to 530 business leaders in sector organisations who are experimenting with or implementing intelligent automation solutions. We also analysed more than 80 use cases, assessing their maturity, complexity and the benefits on offer.
Our research finds that the sector has under-estimated intelligent automation’s true potential. Though there has been progress in AI-driven transformation in core technical operations since 2017, many organisations have yet to scale-up their initiatives. However, an elite group of companies that are making significant progress in driving use cases at scale have been identified.
The characteristics and approaches of this highperforming group offer an insight into best practices for scaling-up intelligent automation.
This study focuses on four key areas:
• It begins by probing what value intelligent automation offers the industry, including whether organisations have under-estimated the value on offer and where the upside is
• It assesses the progress organisations have made and the challenges that are preventing many from reaching scale
• It profiles the use cases that offer the maximum potential and which should provide the focus for investments
• Finally, drawing on the best practices of a highperforming, elite leader group, it outlines key recommendations for driving intelligent automation at scale. Intelligent automation offers significant value to the sector, and its worth has actually been under-estimated by executives.
Sectoral Potential of AI
• US-based electric and gas utility, Xcel Energy, uses data from sensors on wind turbines to develop highresolution wind forecasts through predictive analytics and artificial intelligence. As a result, the company has been able to reduce costs to end customers by $60 million by increasing efficiency of generation.
• Gazprom, the Russian gas giant, used robotic process automation ( RPA) to automate verification of metre readings. In the first two weeks after the automation went live, an employee was able to validate about 130 invalid meter reads, saving 10 hours of work per employee.
• United Utilities, the UK’s largest listed water utility, recently tested an AI platform to analyse large data sets on factors such as weather, demand for water, pump performance and electricity prices. The information
THE ENERGY AND UTILITIES SECTOR COULD REALISE COST SAVINGS FROM $237 BILLION TO $813 BILLION IF IT WERE TO IMPLEMENT INTELLIGENT AUTOMATION IN ITS TARGET PROCESSES AT SCALE
is used to make decisions on the most cost-effective and efficient way to run pumps, detect burst pipes and minimize the risk of discoloured water. During the trial, the utility saw energy savings of 22%. • Offset Solar, a US-based solar company, generated $1.2 million revenue within six months using a simple homepage messenger chatbot. In fact, when companies implement these technologies, they often find they deliver greater benefits than were expected. This confirms a tendency highlighted by Roy Amara, co-founder of Palo Alto’s Institute for the Future, who says, “We tend to overestimate the effect of a technology in the short run and under-estimate the effect in the long run.”
Nearly half of our respondents say that they have underestimated the true potential of intelligent automation. 47% say that the cost savings were underestimated and many said the same of customer satisfaction (48%) and
revenue gains (45%).
Artificial intelligence is on the rise, though critical challenges remain in achieving scale. Only a minority are able to scale up their intelligent automation initiatives
We define “scaled adoption” as deployments that go beyond pilot and test projects and are adopted to a significant degree across business units, functions, or geographies. However, scaled adoption in the sector is rare. This is true at both a global cross-sector level (where in 2018 we found that 16% have reached scale) as well as for energy and utilities specifically (15%).
Nikolai Lyngo, senior vice president corporate strategy and innovation at Equinor, believes that moving to greater scaled adoption will require a mindset change. “I think one of the challenges that we have as an industry is our ability to reach full-scale implementation,” he says. “I think it is partly because this is something new and requires a different kind of thinking as well as execution. We need to have the basics right – the data. I think sometimes we are trying to do too many things at the same time and that is a problem.”
Reaching enterprise-wide adoption after a pilot requires foresight, strong governance, long-term strategic planning and a determination to keep the change management angle front and centre. While most organizations understand the importance of automation and AI, they are not always clear on the exact role of these technologies and how they fit into the overall organisation strategy.
Abhijeet Bhandare, chief automation officer, GE Power says, “The first thing is finding the purpose – why automation? People end up doing automation just because it is cool. We need to realise that these are business dollars that are being put at stake.”
AI has Matured
Energy transition trends – such as generation decentralisation, auto consumption, local load-demand balance with energy management systems, and smart
devices creating significant volumes of structured and unstructured data – support a variety of transformative AI use cases. AI is helping the sector in a variety of ways, from boosting efficiencies to contributing to the fight against climate change. The rise of renewable power mix is also enabling AI-based use cases, such as forecasting, demand and supply management and energy trading.
In 2017, we surveyed more than 120 senior executives from energy and utilities organizations that are already implementing AI. Today, we have made a like-to-like comparison with our current research, involving the 373 executives that are experimenting with or implementing AI. We found significant progress: • Two years ago, the majority – 55% – were just deploying pilots • Today, the majority – 52% – have actually deployed a number of use cases In 2019, 52% of energy and utilities organisations have deployed a few or multiple AI use cases.
The sector has made significant strides in deploying AI technologies in core technical operations across the value chain, from generation to trading. We analysed more than 20 use cases in the core value-chain functions and found that AI technologies dominate.
WE TEND TO OVERESTIMATE THE EFFECT OF A TECHNOLOGY IN THE SHORT RUN AND UNDER-ESTIMATE THE EFFECT IN THE LONG RUN — Roy Amara, Co-Founder, Palo Alto’s Institute for the Future
Exploration/Production/Generation
• NextEra Energy, a US-based Fortune 200 energy company, is applying machine learning to optimize operating parameters for its wind turbine fleet. The aim is to maximise output and perform predictive maintenance. “We operate at $3 to $4 [per MWh] better, including availability and operating costs, on the wind side than anyone else in the country,” says Jim Robo, CEO, NextEra Energy. • French Oil and gas major, Total, has signed an agreement with Google Cloud to jointly develop AI solutions that will be applied to subsurface data analysis for oil and gas exploration and production.
Transmission/Distribution
• US-based startup AppOrchid is deploying deep learning and natural language processing to understand grid behaviour under variable wind conditions. • Canadian multi-utility provider, Utilities Kingston, is piloting AI and geospatial analytics to optimize its leak detection activities. The solution is expected to reduce the time and cost of detecting leaks by more than 60%. • Pacific Gas & Electric has employed machine learning to increase the accuracy of load reduction forecasts for demand response.
Supply/Energy Services/Retail
The UK’s EDF Energy, serving approximately five million business and residential consumers with electricity or gas, uses Amazon’s Alexa as a service channel, helping consumers in areas such as account balance inquiries, learning next payment dates and submitting meter readings.
Aidan O’Sullivan, head of University College London’s energy and AI research says, “The proliferation of virtual assistants in homes, combined with data, could fundamentally disrupt the way we buy and use electricity. The integration of energy data with products like Alexa and Google Home may lead to AI home energy management systems where, for example, rather than turn on your
washing machine yourself, you schedule it to run when the electricity price is going to be lower.” The use of AI and predictive machine-learning algorithms is enabling the consumer to have foresight over their energy profile.
Trading
• UK-based start-up, Open Utility, has spearheaded peer-topeer activity using its algorithm-based platform, Piclo Flex. This is the UK’s first online marketplace for renewables. Working in partnership with licensed suppliers, Open Utility provides commercial energy users and generators with a simple, intuitive, and transparent way to buy and sell power. Rule-based technologies dominate in support functions In line with global averages, rule-based technologies – IT process automation and robotic process automation (ITPA/RPA) – are also prominent in the energy and utilities sector.
Three out of every four organizations are adopting ITPA/ RPA technologies, with most implementations in support functions rather than core technical operations. UKbased United Utilities has been a pioneer in using RPA
in its back-office processes. Since 2017, the utility has automated more than 20 processes, with another 12 in development.
“The benefits are not just time savings, but making the process robust,” says Genevieve Wallace Dean, the company’s head of robots about the importance of RPA. “We know the robots aren’t going to go on holiday or be off sick.”
Even though rule-based automation is most prominent, scaled adoption is still rare. Of those organisations that
have deployed rule-based technologies (397 organisations of the 529 we surveyed), only 17% have deployed them at scale. Organisations should ensure that they stabilize and optimise their processes prior to applying automation.
Organisations are missing a big opportunity by ignoring high-impact use cases.
Our analysis of more than 80 use cases shows that only a minority are focusing on use cases that are not
only easy to implement but have a high-benefit upside (which we call the “quick wins”). Neglecting these automation and AI quick wins – which span core and support functions – is a missed opportunity. Within core functions, organisations are focusing on low-complexity and low-benefit use cases.
Our research shows that over a third of the energy and utilities organizations (38%) are focusing a lot of effort on use cases that are easy to implement but which have a low-benefit upside. On the contrary, fewer than one in five (18%) are focusing on the quick wins.
Organisations are tackling the most complex support function use cases.
In support functions as well, only 11% of the organisations are focusing on the quick wins and just over one in four (26%) are focusing on use cases that are not only complex to deliver, but also have a less-compelling upside. Forty-six percent are focusing on high-complexity and high-benefit use cases.
The road to intelligent automation at scale
The traditional utilities business model is under pressure worldwide, as technology changes and increased competition make their presence felt. However, despite the monumental changes happening, the sector has only just started to advance beyond its existing conservative mindset.
We see this tendency in driving intelligent automation at scale.
Our research shows that only a few organisations have been able to break the conservative stranglehold and create truly breakthrough initiatives. We call this group the “Automation Frontrunners”.
Use cases are the foundation of intelligent automation strategy. Finding and developing viable automation and AI use cases gives leaders a clear understanding of how these technologies fit with the organization’s business strategy, competencies and current and future technology capabilities.
When utilities kick off automation and AI initiatives, they need to consider the following areas: 1. Ensure they have the basic minimum technical talent in place to collaborate with functional teams. Without this expertise, use-case selection and implementation effort will suffer. 2. Cover the underlying fundamentals before proceeding, such as data accessibility, legal and ethical implications, risk assessment and success criteria/KPIs. 3. The potential scope of impact should also be a key criterion while choosing, initiating, and scaling up use cases. 4. The time and resources invested in AI initiatives require a delicate balancing act. AI applications require time for gestation and optimisation. For example, an application may not be fully optimised until it has enough data. Therefore, you need to strike a balance between ensuring AI initiatives have enough incubation time while also ensuring that they do not take up too much of your talent and resources, which will often be thin on the ground.
Five Recommendations
By studying what these automation frontrunners do differently; we have developed five recommendations for companies looking to join their ranks.
Utilities tend to have a ‘conservative’ approach in evaluating benefits from automation and AI. As we saw at the beginning of this report, around half of the respondents to our survey said that they undervalued the impact of implementing these technologies, despite the fact that the sector has driven significant benefits compared to other industries. This could reflect the fact that the sector lacks the internal expertise to select appropriate use cases and measure benefits.
1. Develop a pragmatic approach to evaluate and choose intelligent automation use cases
Utilities tend to have a ‘conservative’ approach in evaluating benefits from automation and AI. As we saw at the beginning of this report, around half of the respondents to our survey said that they undervalued the impact of implementing these technologies, despite the fact that the sector has driven significant benefits compared to other industries. This could reflect the fact that the sector lacks the internal expertise to select appropriate use cases and measure benefits.
Abhijeet Bhandare, chief automation officer at GE Power, highlights the importance of developing a strong approach towards use case selection. “We have a very clear filtering criteria defined for automation use cases,” he explains. “We have close to 200 automation ideas in pipeline, and on an average about 50% to 60% of them will be rejected. It is important to focus your attention on the remaining 50%, as they will give you the most value. And you must have the right criteria – whether it is value, efficiencies, cost savings or the opportunity cost. Organisations should focus on quality over quantity of use cases.”
While automation and AI is a valuable way to transform an existing approach, maximum potential often occurs in new applications. Take, for example, renewable power microgrids.
With the advent of battery storage, coupled with predictive analytics, it is possible to build virtual micro grids that can run totally on distributed renewable energy. A high-potential use case like this would not be unearthed and considered if organisations only focus on reformulating existing functionalities and processes.
“Digital is fundamental to being able to produce the energy, because you want the energy to be increasingly clean and renewable,” says Morag Watson, Chief
THE FIRST THING IS FINDING THE PURPOSE – WHY AUTOMATION? PEOPLE END UP DOING AUTOMATION JUST BECAUSE IT IS COOL. WE NEED TO REALISE THAT THESE ARE BUSINESS DOLLARS THAT ARE BEING PUT AT STAKE — Abhijeet Bhandare, Chief Automation Officer, GE Power
Digital Innovation Officer, British Petroleum. “Digital is fundamental to renewable. You cannot do renewable energy without digital. You can’t do it. It’s fundamental to solar. It’s fundamental to wind. It’s fundamental to distributed energy.”
Using the complexity vs. benefit assessment can help to organise and prioritise intelligent automation use cases to ensure they are aligned with the organisation’s digital goals.
2. Invest more effort in integrating and optimising the right processes for intelligent automation deployments
Intelligent automation can remove considerable overhead in both support and core business functions. However, they require considerable integration and re-engineering of process flows. Force-fitting these solutions to existing structures carries risks: financial, safety, or reputational. In our experience, organisations usually over-emphasize the difficulty of technology execution and underestimate the importance of process re-engineering and workforce impact.
Our survey also shows that considerations around security and legacy system integration are still stumbling blocks for the utilities sector. Critical challenges highlighted by our research include: • Complex IT security requirements • The significant investment required • Integrating automation technology with existing systems and tools. Automation frontrunners are aware of these challenges.
We found that over half of them (51%) are planning to set up dedicated teams over the next two to three years to investigate the impact and adoption of automation, compared to just 11% of their peers.
3. Put greater emphasis on technology, backed by larger budgets
Automation frontrunners make greater budget commitments than their peers - they spend 19% of their current IT budget on intelligent automation, compared to 13% in other organisations. Moreover, they intend to increase this by 31% over the next two to three years and aim to automate around 17% of their current business processes. In both these areas, they exceed their peers.
Automation frontrunners are clearly more focused on unlocking untapped potential rather than replacing
existing business processes. “I often talk about artificial intelligence as augmented intelligence as opposed to artificial intelligence,” adds Morag Watson, chief digital innovation officer in BP. “It really is about bringing something else to the table, such that our experts and people can focus even more on the stuff they were trained to do and on the higher order stuff. Everybody seems to think there’s a limited amount of high order stuff to be done, but I don’t think we know. I think that there could be a whole new class of jobs, and a whole new range of things to be done in the future, that we didn’t know existed.”
The increased budget commitment of automation frontrunners reflects their determination to tackle more complex technologies such as machine and deep learning.
While these technologies are high-cost in terms of talent and resources – and can have a higher risk profile – they tend to deliver greater rewards compared to rulesbased automation. As we saw earlier, we also found that most, if not all, high-benefit use cases for core functions are based on AI applications.
The International Energy Agency (IEA) predicts fossil fuel share in global supply will fall to 40% by 2030, with renewable contributing 65% by 2040.
On the one hand, this means that utilities will need to adopt highly volatile energy sources such as solar and wind to achieve better economics. At the same time, they need to supply the growing number of high power applications such as electric vehicles and electric home heating. This could lead to highly volatile demand and supply, potentially overloading or oversupplying a local grid. In this environment, capabilities such as local grid management, weather forecasting, data-driven demandresponse strategies, efficient energy storage utilization and dynamic pricing will be essential to ensure a reliable operation while ensuring competitiveness. Climate change regulation is also a growing issue. For example, San Diego Gas and Electric prevents wild fires by utilising sensor data – along with satellite weather monitoring – to
ensure distribution assets are maintained. These sorts of sensors and analytics will also play a critical role in ensure the organisation can meet California’s goal of generating 100% of its electric power by 2045 through clean energy.
4. Put in place dedicated and centralised leadership – as well as governance – for intelligent automation
With many organisations struggling to find critical digital skills, a centralized and dedicated taskforce can be critical to achieving automation goals. Over time, this centralised approach will also ensure the organisation builds the experience and expertise it needs.
As Figure 20 shows, front-runners are more likely to have a centralised approach than decentralised governance or a hybrid approach – 51% take a centralised approach, but this drops to 36% for the rest of the sample. In addition: • Thirty- nine percent have a dedicated leader for automation • Fifty-five percent have their management board or CEO sponsoring all initiatives.
This top-down approach provides focus and has helped frontrunners deliver multiple projects, at scale, while achieving significant upside.
5-Involve your workforce, invest in their capabilities and drive a dedicated change management programme.
Resistance can be a significant barrier to automation initiatives, but 80% of frontrunners have employees who are open to it. In addition: • Seventy-eight percent say automation has raised employee satisfaction levels, compared to 65% of the rest of the sample. • Seventy-six percent say it has created new job profiles, counteracting the fact that 71% also say that it has led to workforce reductions. The sector’s organisations need to design and implement a comprehensive upskilling strategy to realise their goals.
It will be critical to achieving scale, encouraging employee adoption and bridging the talent gap. Hugh Mitchell, chief HR and corporate officer at Royal Dutch
Shell, says, “Once we’ve recruited good people, we ensure they get the training and experience they need to do their challenging jobs effectively and deliver on strategic priorities. We do a careful segmentation of our work force, looking at which types of jobs and skills are likely to give us a competitive edge. That requires an indepth look at the needs and priorities of the people in these skill pools.”
Recent research we conducted on workforce upskilling shows that organisations that are in the midpoint of their upskilling program have: • More successful intelligent automation initiatives • Higher employee morale • Support of a significant majority of employees (90%) for automation and AI initiatives. For energy and utilities, upskilling is a significant issue. As we saw earlier, for example, “lack of talent skilled in automation technologies” was a challenge for 57% of organisations.
To design and implement a successful upskilling program, six factors are critical: 1. Assess your tech investments and to what extent they
will impact on the workforce 2. Define the skills you need and when you need them 3. Make the upskilling program a win-win for your people and organisation by making them relevant to the needs of the company 4. Align learning with organizational strategy 5. Equip leaders to communicate effectively 6. Ensure you have a change management strategy in
place
Conclusion
The global energy industry is undergoing monumental change.
The traditional, centralised provision of power from sources such as coal, gas and nuclear is facing an economic, political and social backlash – a result of climate change, health and safety and security concerns. At the same time, growing demand from developing countries, as well as new usages like electric vehicles, need to be satisfied. All of these trends mean that distributed and weather-dependent renewable energy is an increasingly critical and growing part of the energy portfolio.
Embracing RPA, advanced analytics and artificial intelligence is absolutely instrumental in meeting climate change goals and the growing demand for clean, cheap, reliable energy.
Driving scaled adoption in support and technical functions will play a key role in driving greater efficiency and customer satisfaction while also helping rethink operating and business models. However, our research shows that many organizations struggle to make a success of intelligent automation, failing to achieve scale and drive value. Organisations are also missing out on critical use cases in core and support functions that can deliver outsized benefits. The benefits of intelligent automation are consistently undervalued and misjudged, which reveals a significant gap in capabilities, talent and ambitions.
To address this gap, there are a number of priorities. First, organisations are capturing most automation value from the most mature and accessible use-cases in the downstream business, such as chatbots. But, as our survey shows, more value resides in the upstream business (generation, trading) and midstream activities (networks). Organisations therefore need to focus efforts towards this domain, as well as new businesses, such as renewables and beyond the meter. It is easier to grow the value from a digital enabler in these new deep techs than refurbishing an existing asset (coal or gas plant for example). Second, organizations need to take a pragmatic approach to choosing which automation and AI use cases to focus on, while also making sure they maximize ROI by re-engineering processes prior to implementation. Finally, they must drive change management to create and instil a culture and mindset that welcomes intelligent automation as a necessary component to complement the human workforce, creating the smart, augmented utility of the future.
The traditional business models of stable long-term agreements backed by large tickets capex projects is slowly being replaced with small players and smaller projects – shale oil, rooftop solar panels, micro-grids are a few examples of this trend. If the energy and utilities sector is to grow alongside this era’s challenges, it requires an ambitious adoption of automation and AI. We believe organisations who develop such capabilities and culture will be able to innovate more aggressively through new business models and grow vibrantly at the expense of the organisations who fail to change.
About the authors: Philippe Vié is VP, Capgemini Group Energy, Utilities & Chemicals Sector Lead; Alain Bollack is VP, Capgemini Invent UK; Dr. Adam Bujak is VP, Intelligent Automation, Capgemini Business Services ; Jerome Buvat is Global Head, Research & Head, Capgemini Resear ch Institute; Nancy Manchanda is Manager,Capgemini Invent India; Shahul Nath is Consultant, Capgemini Resear ch Institute; Abhishek Jain is Consultant, Capgemini Research Institute
Source: Capgemini Research Institute; Global Automation Research Series: Energy and Utilities.