Klarna intends to grow its business by deploying generative artificial intelligence instead of hiring new staff.The “buy now, pay later” credit business believes that it will continue to expand its operations and revenue despite a hiring freeze that was announced in December, because AI is making work more efficient.

“Outside of engineering, we’ve slowed down hiring across the board, because AI allows us to create efficiencies,” Martin Elwin, the financial technology group’s head of AI, said. “We still have high growth ambitions and previously it would have meant growing the organisation a lot faster as well. But now we don’t need to do it.”

This year Klarna said that its virtual assistant tool had replaced 700 of 3,000 customer service jobs, which would save the business $40 million a year. Government departments in Westminster and a leading airline have been among those approaching Klarna to find out how it is deploying the technology.

Elwin said Klarna was not anticipating more job cuts, but added that it was changing its attitude to hiring: “As we are putting these AI solutions in place, what is actually needed might shift a little bit. And if someone leaves the company, we need to think is it really necessary to replace them immediately? We can be a little bit more careful and mindful there. Many things just can be automated, can be simplified, without having to invest a lot of engineering hours.”

At Klarna, the average customer support waiting time has dropped from 3min to 7sec and the average time to resolve a problem has gone from 11min to 3min, while customer satisfaction has remained the same.

Klarna was an early partner of OpenAI, the creator of ChatGPT, and as a result tests the latter’s latest models and how the technology works in the sensitive world of finance. Although it is looking at other generative AI businesses, such as Anthropic and Mistral, OpenAI’s models were said by Elwin to be “hard to beat”.

The business does not yet use generative AI as part of its credit checks, because of fears about the technology’s reliability. There have been warnings over AI reinforcing stereotypes or “hallucinating” and seeing misleading patterns.

Elwin said: “It’s obviously something that we’re exploring. Can we use these technologies for things like making a credit decision more fair? Maybe we can consider new information sources with these models that we haven’t been able to consider previously. But, as it is now, there is a need for more predictability, more determinism in that, in that process.”

In an interview with Rishi Sunak last year, Elon Musk, the Tesla boss who co-founded OpenAI in 2015, said that the rapid advance of AI technology would “eliminate” the need for jobs entirely, with people working only for “personal satisfaction”.