Given the ubiquitousness of artificial intelligence (AI) today, it shouldn’t be surprising that even Eskom has been using AI, albeit in its simplest form, for four years.

South Africa’s energy utility introduced Arthur, its ‘ever-present, amped-to-assist, AI-enhanced chatbot’, back in 2021 to help customers report faults, check their accounts and submit meter readings.

Speaking at a conference last year, Eskom board chairman Mteto Nyati emphasised AI’s importance to customer service. ‘If you combine a human being with AI, you can attract work the continent could not attract in the past. Instead of seeing AI as a problem, it’s actually helping us operate at a much higher value, moving our people to do work they may not have been able to do in the past,’ he said.

Utilities, however, are notoriously risk-averse, so their growing uptake of AI is a game-changer.

‘Utilities have been risk-averse for good reason, but it’s clear that in today’s fast-moving environment, slow-rolling innovation is far riskier – and could spark fierce competition from new entrants,’ Raghu Madabushi, a director at Silicon Valley-based VC group National Grid Partners, writes in Fortune magazine. ‘OpenAI’s ChatGPT has upended legacy tech in just a few years – utilities could easily be up next,’ he says.

The benefits of AI go way beyond customer-service chatbots to include improved regulatory compliance, better transmission and distribution planning and new electrification opportunities that can bolster revenues. AI is an inflection point that could lead to a new kind of utility, says Abhay Gupta, co-founder and CEO of software company Bidgely, which works with utilities around the world to harness the power of AI.

‘A lot of utilities are looking to transition from selling kilowatt-hours to selling products and services. In as much as a utility can get more visibility about their consumer’s lifestyle and the profile of their usage and what’s important to them and motivates them, the more they can ensure that what they offer is valuable and useful to them.’

Gupta points to one enormous advantage most utilities have over other corporates – access to data.

‘They have a mountain of customer data. For machine learning to operate or to be successful, it has to have a cache of data to learn against. This is an unnatural advantage they have that they are currently not really exploiting or leveraging,’ he says.

Of course, that mountain of data relies on smart meters to capture it – and it’s not cheap. In 2023, Eskom announced its ZAR15 billion plan to replace 6.9 million electricity meters with prepaid smart meters, which caused some ‘Big Brother is watching’ consternation that Eskom would be able to switch off users’ individual appliances, such as geysers. That’s not true, but Eskom is able to limit loads at times of constrained energy availability.

Writing in a November 2024 blog, Tamuka Chekero and Noëleen Murray, of the Operationalising a Just Transition in Africa research project, funded by the International Development Research Centre, look on the brighter side of Eskom’s use of AI and smart meters.

AI analysis of data collected by Eskom’s smart meters will allow the utility to identify vulnerable households at risk of payment shortfalls, they say. But this is not necessarily a bad thing. ‘This enables targeted interventions and support mechanisms, preventing disconnections and ensuring continued energy access for these communities.

‘Advanced data analysis using AI can identify communities most in need of energy access, enabling targeted interventions and ensuring no one is left behind,’ they write.

In addition, ‘AI-powered micro-grids can manage distributed renewable energy sources in remote locations, bringing clean energy to off-grid communities’.

Chekero and Murray envisage AI also having a broader socio-economic impact. ‘AI-powered financial products can also facilitate access to financing for renewable energy projects and micro-grids, even in underserved communities,’ they write.

The other benefits of AI to utilities are well documented. For example, by analysing real-time data from weather patterns, energy consumption and grid conditions, AI ‘allows for intelligent forecasting, optimising renewable energy generation and enabling a smoother integration into existing grids’.

AI can also help integrate distributed renewable energy sources into the national grid, such as solar or wind energy produced by so-called prosumers – residents or businesses that generate electricity.

Not only does AI assist utilities on the supply side, it can also help customers regulate their use of energy or other utilities.

Eskom’s chief information security officer, Sithembile Songo, gives another example of how AI assists Eskom in its operations.

 ‘We’ve been leveraging a lot of AI and machine learning in terms of real-time response and detection, because we realise cyberattacks are quite advanced nowadays.’

AI helps the utility fend off up to 1 billion attempted cyber incidents every month, she says.

Meanwhile, Rudie Opperman, regional manager of engineering and training at ITC manufacturer Axis Communications, underlines the role AI can play in securing water and sanitation infrastructure against unauthorised access and potential vandalism. ‘AI-powered intrusion systems can use thermal sensors, radar sensors and even conventional optical cameras.’

He adds that in AI-controlled facilities, utilities can also keep track of maintenance schedules, unauthorised access and system malfunctions and anomalies.

‘Water treatment and distribution, infrastructure maintenance and process management are essential. Neglect and mismanagement should be measured and visible, and comparisons should be easily possible between facilities over time. Unwanted security breaches or vandalism should be quickly detected and deterred.’

Chekero and Murray argue that the true value of AI goes beyond the initial win.

‘For long-term success, we need to build local expertise. Think of it like this: would it not be awesome if the next generation of AI energy whizzes came from these very communities? We need training programmes and partnerships with universities and research institutions to cultivate this local talent,’ they say.

Wrenelle Stander, CEO of Wesgro, the Western Cape’s trade and investment agency, agrees with that sentiment. Writing in a recent Business Day opinion piece, she stresses that ‘collaborating closely with educational institutions is key. Fortunately, we have strong foundations to build upon’.

‘Prioritising investments in AI education and training programmes will empower individuals with the necessary expertise. Tailored AI-centric courses at universities and technical institutions will cultivate a new generation of professionals capable of driving the country’s energy transition,’ she writes.

Chekero and Murray emphasise the importance of building local AI skills, because ‘solutions designed by people who understand the specific challenges of these communities will be far more effective’. In addition, ‘by including local voices, we can ensure AI solutions are culturally sensitive and truly benefit the communities they serve’.

While the benefits of AI for utilities are undeniable, implementing AI is not without its challenges

Besides the cost of the undertaking, there is – Eskom’s roll-out of smart meters notwithstanding – also a shortage of available high-quality data on energy consumption patterns, infrastructure conditions and weather data, which in turn affects the training of effective AI models.

‘Regulatory barriers also present a notable challenge in the adoption of AI technologies within the energy sector across many African countries,’ say Chekero and Murray.

According to Stander, South Africa needs clear policies and guidelines ‘that will accelerate AI development and adoption in the country’.

While South Africa has no specific law regulating AI, its AI Planning Discussion Document is considering legislation to deal with risks related to autonomous devices that use AI to decide on actions, and the aggressive loss of employment leading to social risks.

In addition, the Protection of Personal Information Act could well have an impact as it extends to any automatic processing of personal data.

However, as National Grid Partners’ Madabushi alludes to in Fortune, utilities must start adapting to the future now if they don’t want to risk being made irrelevant by technology.

 ‘Embracing AI is not just an option,’ says Stander. ‘It is a necessity to propel our nation towards a future where energy crises are mere relics of the past.’

Images: Gallo/Getty Images