There needs to be more clarity on who takes ownership of AI, says First Utility

The gas and electricity provider says it can be difficult convincing a board to invest in AI-based technology such as chatbots.

chatbots

First Utility claims businesses need more clarity on which department looks after artificial intelligence innovations such as chatbots.

Speaking at Marketing Week’s Supercharged event today (4 July), the gas and electricity provider’s lead software engineer Natalia Konstantinova said there are still big questions around who is best suited to lead internal AI projects.

“Is it marketing or customer support?” she asked. “Who are the people who say ‘we need to add this content into the solution, and make these edits’, and are in the middle of gathering this information all over the company? If you have a new tariff or news, [they] need to be in the middle collecting this.”

The brand created its own chatbot three years ago out of necessity after finding that relying solely on its call centres was “inefficient” as they were dealing with the strain of a 670% surge in new customers over a five-year period. It claims it was the first within the home utilities sector to launch a chatbot to deal with customer queries.

First Utility’s chatbot has different use cases, and can answer general questions around topics such as smart meters, as well as have multi-layered conversations around meter readings, tariff comparisons and a customer’s bills.

READ MORE: First Utility’s CMO on why it wants to become a tech company

However, Konstantinova admitted it was “tricky” persuading board members to invest in artificial intelligence in the first place. That said, she now believes it is easier for others to make a business case due to “all of your competitors doing it.”

Others speaking at the event were more cynical about the benefits of chatbots. Nicola Millard, customer insights and futurology at BT Global Services, said she believes that despite the “hype” around chatbots, they are currently only useful for “basic” tasks and are only as good as a company’s data.

She concluded: “It’s more difficult to use for complaints and complexity. If we get angry , we tend to tell long, rambling stories that are hard to analyse. Sarcasm also tends to throw algorithms, as it has a limited ability to detect the emotional context.”

Recommended