top of page
  • Vince Cheung

I asked ChatGPT how utilities can better serve their customers. See what AI said, and my reaction.

It’s been a couple of months since OpenAI debuted ChatGPT, the latest among a handful of beta-stage generative artificial intelligence (AI) tools. Already, it has generated spirited debates in the academic, legal, marketing, and many other arenas.

Like many, I’m curious about what generative AI can do, so I put on my “Utility Marketer hat” and asked ChatGPT this somewhat open-ended, term-paper-style question: List five ways utilities can better serve their customers in 2023.

Below, I will share what I received back (verbatim), my initial reaction to each, and my overall take on how marketers can contend with AI-generated content.

What ChatGPT said about five ways utilities can better serve their customers in 2023:

Overall, ChatGPT did a good job summarizing a few key themes that reinforce my own point of view and advice for utilities.

Primarily, those are “undeniable truths” such as a focus on CX, transparency with customers, and greater community engagement. Beyond that, whether deliberately or accidentally, ChatGPT also did a decent job highlighting two “hot topics” in my mind – rising cost of natural gas (and impact on rates) and how the sunsetting of COVID-era billing relief/deferral practices will impact utility billing and assistance.

That said, while I’d consider the five points above good thought-starters, they left me hanging in terms of what concrete actions utilities can take, and how best to communicate their benefits to customers to drive better results.

For example, what are the key stages and triggers in the utility customer lifecycle we can pinpoint to gain more mobile app. adoption, sign-ups for budget or average-monthly billing plans, and consideration for switching to time-of-use plans? (More on Pragmatic’s approach to the energy customer lifecycle here)

Closing this critical gap definitely requires human intelligence – from more fact-checking to committed thinking on actionable strategies, anchored on current industry and customer knowledge.

Additionally, in order for machine learning to generate and assemble better text “deliverables,” its algorithm must first have access to source text “inputs” from content that already exists. In other words, even as the AI algorithm is constantly trained and improved, someone has to have created, and published, at least fragments of an original idea in order for AI to process them into more coherent prose.

Forrester recently characterized the outputs of AI tools like ChatGPT as “coherent nonsense… producing material either deliberately or accidentally inaccurate.” My own take is that no one, utility marketers included, should readily dismiss the role that AI can play in customer communications. We should, instead, seek to harness new capabilities like this under thoughtful human supervision and intelligence, and find ways to “turbo-charge” the utility customer lifecycle to stay relevant with the consumer experience.

And, it’s perfectly okay to enjoy an occasional AI-generated limerick.



bottom of page