Why AI is a non-negotiable for utilities of the future

Artificial intelligence, or AI, is transforming the way utilties operate, turning these dinosaurs of the power system into dynamic, decarbonised power houses of the future.

“AI seems to have exploded onto the utility scene in a short period of time, fuelled in part by maturity of the technology but also the pressing need of the organisations when they are faced with the transformation of their 125 yea- old businesses to a clean energy future.”

Carol Johnston, global vice president of Energy, Utilities & Resources at enterprise software company IFS explained how the scope and speed of the energy transition is impacting utilities and forcing them into new operational territory and heightened performance.

AI has been slowly evolving and maturing over a long time, said Johnston. The first artificial neuron came about over 80 years ago and Eliza, the first chatbot ever developed, was released in 1966.

“Sadly I’m sure some organisations like my cable company still use Eliza today,” said Johnston, referring to the slow uptake of automation in some organisations.

However, Johnston emphasised that because of the need and complexitity of managing the future grid, distrbuted energy assets and the two way energy flow, AI is a non-negotiable for utilities and must be coupled with a corresponding strategy for digital transformation.

Have you read?
Digitalisation defines industry-leading energy organisations survey finds
Austin Energy deploys AI-driven early wildfire detection

Beyond Eliza – complex use cases for AI

According to Johnston, “AI is shaping up to be a huge game changer in this industry…enabling better insights and decision making support and filling in the gaps in resource availability…”

And there are very specific use cases proving AI’s worth, from business and grid planning to automating and improving customer service to advancing field maintenance and service activities.

Johnston mentions six areas where utilities are experiencing the most disruption:

1. Forecasting and simulation: Referring to a variety of events and assets, as well as developing resources and inventory maps.
2. Optimisation: Referring to dynamic scheduling, route planning to maximise carbon offsets, automatic bundling of work to boost efficiency and planning of restoration activities after an event.
3. Anomoly detection: Allowing for real-time asset health monitoring, predictive and preventative maintenance.
4. Recommendations: Smarter systems can suggestions on risk-based issue detection, providing greater insights on workforce deployment and crew matrix recommendations.
5. Contextual knowledge: Allowing complex software to be navigated more easily, helping people complete accurate and quick data capture and analaysis.
6. Content generation: Such as generative AI, used to automate capacity planning adjustments to meet demand, offer personalised training materials, and automate work and purchase order processes.

“The opportunities are quite literally endless, and we need to focus in on where are we going to get the biggest disruption and bang for our dollar.”

Johnston excitedly elaborates on how predictive maintenance is being revolutionised, especially for utilities that have millions of assets in the field and data coming in from all of these sources. “AI-driven anomoly detection is a real game changer,” said Johnston, adding it can save time, money and improve safety for the community and workers.

Utilties are also embracing AI-driven schedule optimisation like never before, explained Johnston, especially as workforces are reduced and work loads are increased.

And of course, when it comes to building out the grid to manage spiking load curves due to data centre demand, AI helps to guide this development without “blowing the lid on affordability metrics.”

Risks associated with AI deployment

Johnston stated that for utilities, there are risks of unknown and unexpected consequences of adopting AI too quickly.

“This is not an industry where big risks without safeguards are rewarded, it comes with big penalties if you dont get it right.”

AI shows great promise to improve operations, she said, however, it’s important to employ it ethically to avoid unconcious bias or AI hallucinations, which provide bad results.

The technology will only do what we train it to, added Johnston.

And so we need to train it well, not avoid it.

“Turning away from all risk and deciding not to embrace emerging technologies like AI…will see organisations left behind.

“As AI really fuels innovation, operational efficiency, new business models, strategic decision making and unlocks new revenue streams, it’s freeing the human condition to become more innovative and focus on more strategic thinking…”

An effective digitalisation strategy

Johnston reiterated that in terms of digitalisation, “we are all learning as we go.”

However, there are elements of an effective digitalisation strategy that will help to ensure its success.

These include a solid architecture and data foundation, as well as an ability to adjust the models as results are obtained. It’s important for utilities to be flexible, she added.

“We can argue about whether a utility is ready to embrace AI or automation to those levels or not, but the technology and the capability is there when the industry is ready to embrace it.”

Another important element of any successful digitalisation effort, explained Johnston, is to ensure the workforce is ready. Technology is ultimately enabling and supporting people and you need to bring the work force along on this journey, whether “in great fear or great excitement.”

She admitted that some jobs will be eliminated, but more jobs in the industry will be created.

It’s a shift, not an elimination, added Johnston.

For more insights from Carol Johnston about how AI is revolutionising utility operations, listen to this episode of the Energy Transitions podcast.