The company blames outsourced human customer service agents for the dip in quality.
Swedish fintech giant Klarna has been heavily reliant on artificial intelligence for years now, prioritising the technology to deal with a growing volume of its workload over human employees.
Recently, however, the company’s co-founder and CEO Sebastian Siemiatkowski recently admitted that a “lower quality” of service is leading them back to hiring more human workers.
Since 2022, the ‘buy now, pay later’ start-up has slashed its staff count by 2,000. Klarna also hadn’t hired a single human in 2024.
Last year, the company announced that its AI assistant (powered by OpenAI), handled 700 agents worth of workload – or 2.3m conversations with customers. That number has now gone up to the equivalent of 800 full-time agents.
At the time, it said that customer satisfaction with AI agents was on par with humans – something it still maintains.
“I am of the opinion that AI can already do all of the jobs that we, as humans, do,” Siemiatkowski told Bloomberg last year. Later, in an X post, he stood his ground and said that AI could also do his own job.
However, in a more recent conversation with Bloomberg, Siemiatkowski suggested that the job cuts resulted in a “lower quality” of customer service.
“As cost unfortunately seems to have been a too predominant evaluation factor when organising this, what you end up having is lower quality,” he said.
A Klarna spokesperson elaborated on what Siemiatkowski meant in a comment to SiliconRepublic.com.
According to the spokesperson, the CEO’s statement to Bloomberg referred to the cost of outsourced human agents and not AI as to what caused the service quality to falter.
“Some customers get an amazing agent, some a less engaged agent, prompting repeated contacts and higher costs.”
To address this, Klarna is hiring freelancers on a part time basis to handle complex issues “more consistently” than its current outsourced set-up.
‘Overzealous’ mistake
Forrester principal analyst Christina McAllister believes that Klarna went wrong in the same way many companies do, by “underestimating the complexity of their customer service operations,” combined with an “overzealous pursuit of cost reduction”.
Customers have been conditioned to expect low quality when it comes to automated services, McAllister says, especially considering the barrage of mediocre chatbots they dealt with for nearly a decade. It takes time to change that perspective.
Aggressive AI adoption is not what resulted in the low-quality service experience, the analyst explains. Rather, it was the company’s “overeager” headcount reduction.
AI models are becoming much more useful as time passes and indeed they will at some point be able to provide a similar level of customer satisfaction when compared to humans. But they are yet to reach that.
“They will struggle with novel scenarios, nuanced exceptions, high emotionality – all areas where humans excel,” McAllister says.
So, companies that adopt AI while maintaining human experts as ‘backup’ will see far more success than those who “move too quickly” like Klarna.
Klarna’s strategy and its result should be a lesson for companies to leverage AI in customer service with more nuance. According to the analyst, a “heavy-handed” cost reduction via self service can erode customer trust and satisfaction.
The company is shifting course by hiring part-time customer service workers. This new pilot programme aims to improve the “human part” of Klarna’s customer service, the company spokesperson said.
The rare recruitment drive, the first in more than a year, has hired six remotely working freelance agents as of now, and could expand to around 100 freelancers if proven successful.
For the fresh infusion, Klarna intends to tap into its own user base.
“We also know there are tons of Klarna users that are very passionate about our company and would enjoy working for us,” Siemiatkowski told Bloomberg.
The Klarna spokesperson maintains that the company is still “committed to being AI-first”.
“This pilot isn’t a reversal of our AI strategy.”
According to the company, its AI agent resolves queries in about two minutes, when compared to the 12 minutes humans take – saving it an estimated $39m in 2024.
“However, we believe in a world of AI nothing will be as valuable as humans. We’ll always offer a clear path to a live agent, and we’re investing in higher paid part-time human roles so agents can handle nuanced issues quickly and efficiently for our customers.”
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
