Artificial Intelligence, or AI for short, is nothing new; it goes way back to the 1950s. But things are different now; the vast volumes of data and the computing capabilities we have today mean we can do things better. So, what pain points are utilities seeing today that AI can help with? Let me share some examples.
My first is about how AI can help optimize aging production capabilities while, at the same time, minimizing maintenance costs. Looking at the history of this industry, much of the equipment – except for renewables – is getting on for 30 years old. AI opens up a means to predict how that equipment will run so companies can ensure it will run safely and optimally – both in terms of the technical process and of the financial impact aging equipment has on profitability.
My second example is around how AI can make the interactions call centers have with customers more efficient and more engaging. When a customer calls the center to subscribe to something or maybe complain, the operator has to follow an internal process that involves opening a bunch of different applications with various pieces of information and various fields to fill in. It’s not very interactive, and it’s also very slow; they lose their grasp of the conversation. For utilities outsourcing to a call center provider that’s billing by the minute, it can get very costly.
This is where AI, particularly when combined with natural language processing (NLP), can support the call center operator. By listening to the customer, an AI-enabled bot could automatically open the right applications and fill in the fields. The operator is then free to have a more interactive discussion with the customer. They are also quicker and sharper in providing information, making the call more efficient, more effective, and less costly – which is better for customer satisfaction and better for profitability.
Supporting essential transformations
But AI isn’t just helping to address pain points; it’s also helping with some essential transformations as we enter the new energy era. Let’s look at some examples here too.
My first shows AI helping utilities optimize their investments in renewable energy. Capex is driving the renewable energy market; utilities have to finance their wind farm or solar farm. That finance will most likely come from funding and third parties, and utilities need to show these stakeholders that they are optimizing investments. That’s something AI can help with. AI can forecast what production a specific farm in a specific location might achieve.
The second example is on the retail where AI is reducing churn and increasing sales by helping call center operators understand customer requirements. The retail market is becoming more and more competitive with the energy and utility sectors discovering customer churn 36 months ago, or 18 months ago for the most regulated markets. Players need to react quickly. And AI is helping them understand what they need to offer the customer next to encourage them to stay. The telecoms industry is already using AI to reduce churn, and the energy and utility sectors could simply copy. For a utility company, it might look something like: I have a customer on the call. Should I offer a package for gas and electricity? Should I offer to set up a solar panel on their roof? AI can support this.
My third example is where AI is not only driving the grid but also modeling the grid. Today’s TSOs [transmission system operators] are looking for an AI application to help them manage their investments for the next 20 or 30 years. They want to be able to study how best to set up their grid, taking into account various factors that would impact the grid. Examples might include analyzing how to set it up if a factory (which would have a high level of consumption) or a wind farm (which the grid would need to ingest power from) were built nearby in two, maybe three, years.
Another example could be a power company looking to install a transmission line between two cities. They are looking for an AI modeling approach that can grasp information from diverse sources, bring it all together and analyze it, so they understand what they need to build first, then second and so on. Whereas today, their analysis can take around five factors into account, but with AI and deep learning, they could take 10 or maybe even 20 into account. Together these factors may drive as many as 15,000 different scenarios, showing what may happen in two months, six years, ten years, 20 years or even 30 years – and whether the line is even worth building.
AI essential in today’s complex world
The utility business is becoming increasingly complex, so complex that players can no longer handle all the information they need to manage with traditional technology. They should now use AI, with all its capabilities – including natural language processing, deep learning and reinforcement learning – to minimize the impact of the new world.
We’ve seen how AI can address pain points and help with essential transformation as we enter the new era. But while AI has much to offer, it also brings with it new ‘digital dilemmas’ the likes of which utilities have never seen before. Read our latest Journey 2022 ‘Resolving Digital Dilemmas’ report, researched and written by the Atos Scientific Community, to learn more on emerging digital dilemmas and explore strategies to resolve them.
ABOUT THE AUTHOR:
Franck Freycenon, E&U Business Development Director, Worldgrid, Big data and new energy services
Franck Freycenon joined Atos early 2015. Today, its mission is to enhance the digital transformation in the energy sector and Atos value proposition on this topic. As such, he is responsible for business development of the platform for new services in the energy, based on Atos global platform initiative in Big Data.
Source: smart-energy
