AI Satisfies Customers Better than Human Customer Support, CEO Says

A UK energy supplier found that AI could do 250 humans’ work and score better results

Robot typing AI
Getty Images

Greg Jackson, the CEO of UK energy supplier Octopus Energy, wrote a piece in The Times of London giving insight into his company’s AI experimentation. Specifically, he discussed how AI is now doing 250 employees’ worth of customer support work and netting better average results than those attained by humans.

The artificial intelligence’s output scored 80% customer satisfaction compared to trained workers’ 65% results, hence why Octopus Energy has been leaning more on AI for the task since it started experimenting with the technology in February.

Jackson clarified that artificial intelligence’s stellar report card wasn’t likely to lead to job losses for his employees. One reason for this may be because AI doesn’t have full autonomy at Octopus. In a statement to Insider, the company clarified that the team “supervises the answers AI provides, so, for example, drafting a personalized response that a team member can review and then send on.”

Octopus Energy’s approach to AI is a bit different than, say, Samsung’s, wherein third-party generative AI has been flat-out banned by the company for fear of proprietary data being accidentally fed to the data pools of tools such as ChatGPT.

Still, Octopus’ eagerness to experiment is a prime example of companies utilizing AI in fresh ways, similar to how average people are hopping on the artificial intelligence bandwagon for a wide variety of tasks, including mapping out investment strategies to better optimize their stock market plays.

As detailed in a recent Finder experiment, 8% of surveyed UK adults stated they already get financial advice from ChatGPT. Nineteen percent indicated they’d be game to receive advice from AI.

AI still scares a lot of people, though. Especially in the Hollywood circuit, wherein writers are worried that hungry studio execs will prioritize robot-written scripts since machines don’t go on strike. And apps such as TikTok still aren’t sure how to handle lifelike AI-produced content.