How generative AI is changing Klarna's work culture

Hollie Adams/Bloomberg

The Swedish payment company Klarna is providing an early view of generative AI's impact on an organization, where the buzzy technology is changing fundamental tasks and potentially impacting job growth. 

More than half of Klarna's 5,000 global employees are using gen AI for customer service and to produce enhanced shopping recommendations for consumers. Three months in, the company is already recognizing changes in workflows and cost reduction. 

"Klarna's integration of AI across certain areas of the company is automating mundane and uninteresting tasks that are now done more efficiently and with higher quality with AI," said Martin Elwin, Klarna's engineering director. "That has led to millions of dollars in efficiency savings." 

Klarna, which offers payments, financial services and buy now/pay later lending, uses gen AI from OpenAI, the Microsoft-backed lab that has become well known in the past year as the developer of ChatGPT. While Klarna is one of the first payment companies to use gen AI on a large scale, other firms such as Stripe, the major card networks and Block are also using gen AI — a form of artificial intelligence that uses machine learning and Large Language Models to produce original content. The technology has the potential to revolutionize IT work, customer service and even product development, but these changes will also impact jobs. 

"We're also encouraging wider adoption of these tools among our team," Elwin said. "Greater engagement not only streamlines our workflows, but also promotes skill development and innovation."

Whose job is it anyway?

Seventy-nine percent of professional technology workers across industries have had some exposure to gen AI in 2023, and 22% say they are using it regularly in their own work, according to McKinsey. AI could replace up to 300 million full-time jobs by 2023, according to Nexford University. Customer service representatives, receptionists and accountants were the three top jobs likely to be replaced by gen AI, according to Nexford.  

Nexford also cited research from Goldman Sachs that found gen AI could add 7% to global GDP during that time, a trend that would create technology jobs. 

Klarna CEO Sebastian Siemiatkowski told The Telegraph, a British newspaper,  last week, that AI will perform more job functions, and the Swedish financial institution would no longer hire staff outside of its engineering department. Siemiatkowski said certain tasks outside of engineering can be done faster and more efficiently via AI. Klarna is not planning layoffs; work will be transferred to AI via attrition of human staff. Siemiatkowski also told the British paper that the hiring strategy is based on the potential of AI and not on business performance. 

Klarna returned to profitability in its more recent quarter after four years of losses. The company laid off about 10% of its workforce about a year ago, as the larger technology market suffered through a slump that followed fast growth during the pandemic. More recently, Klarna settled a labor dispute with Engineers of Sweden and Unionen. Before the settlement, the union told American Banker that AI's impact on work roles was not part of the Klarna dispute. The union did not provide comment for this story. 

Klarna's public relations office said the company had "nothing more to add" beyond what was in the Telegraph story. 

In a recent interview with National Mortgage News, a sister publication of American Banker,  Katherine Campbell, former chief marketing officer at AnnieMac Home Mortgage, said AI will reduce jobs in the mortgage industry. "We know there's many people working on capabilities, including tokenization of data from borrowers. Once data marries data, we won't really need people in the middle to get hard copies of data," Campbell told NMN, noting that AI will also reduce pressure to reach productivity goals in specific time windows. 

"This move by Klarna is a signal that other fintechs are going to be looking in the same direction, particularly in the customer service area, where there continues to be pressure to reduce costs by pushing consumers to chatbots," said Aaron McPherson, a principal at AFM Consulting.  

McPherson said he has seen numerous examples of AI-driven chatbots, and their capabilities are becoming more impressive as they are integrated into underlying applications.

Payments

9 new AI moves in the payments industry

November 3, 2023 10:10 AM

"Klarna is actually being cautious here," McPherson said. "While it has frozen hiring, it is not planning additional layoffs, which suggests that it wants to wait and see how successful its gen AI investments are."

And hiring engineers goes against the conventional wisdom that gen AI will put programmers out of work by automating coding, according to McPherson. "Siemiatkowski seems to see AI as more of a growth opportunity than a cost-cutting one, but it is both. That makes it more likely that other companies will go further."

Gen AI at Klarna

One of the early benefits of gen AI is streamlining customer service. 

The most common reason customers contact customer service is to ask why they have been denied credit, Elwin said. Traditionally, this required time-consuming investigations by human agents, sometimes involving expert analysts. 

Klarna recently started using Large Language Models to analyze plain-text questions from customer service agents on why a customer's purchase was denied. It then generates custom structured query language requests (SQL refers to a programming language that makes it easier to work with data).

Klarna uses these SQL requests to gather information from various databases to be summarized by an LLM in a draft response, which the agent reviews before it's sent to the customer. "This helps us get an answer to the customer's question in minutes rather than days," Elwin said. 

Another recently-launched gen-AI product for customer service has made solving merchant disputes more efficient, saving at a rate of more than 60,000 hours annually, according to Elwin. 

"We see a massive opportunity to leverage AI in ways which will both improve our products for consumers, and also make work more interesting and engaging for Klarna employees," Elwin said.  "By automating mundane tasks, Klarna employees can spend their time on more interesting work."

At other financial institutions, recent uses of gen AI include automating corporate travel at Brex, fintech investors funding companies that use gen AI to mitigate fraud and Commonwealth Bank of Australia using the technology to combat financial abuse. CBA is also considering how gen-AI-powered bots could test new products, though the bank insists the bots will not replace human testing. Few, if any of these uses, directly impact jobs at this point. 

AI is already having an impact across financial services, and the payments industry is no exception, said Aaron Press, research director for worldwide payment strategies at IDC Insights. "Risk and fraud management have been leveraging AI for several years," Press said. 

With the advent of gen AI and LLMs, the biggest impacts will be on customer service, where there is already gen AI use in limited production, Press said, adding that gen AI will also be present in exception handling and dispute management. 

"People will still be important, but AI will make them more efficient, which may mean fewer people will be needed in those roles. My hope is that AI-driven automation will mean re-deployment rather than layoffs," Press said.