Klarna, the fintech company that powers e-commerce for some of the world's most recognizable brands, including Expedia, Macy's, and Nike, is at the forefront of AI adoption. The company is integrating artificial intelligence across the company, most notably its AI chatbots, which it recently announced will do the same job as 700 customer service agents.Klarna, which employs around 4,000 people, recently released statistics These are questions about how effective and effective this tool is amid delicate and high-stakes debates about the role of generative AI in business, how humans interact with it, and its impact on the future of work. It shows how effective it is. CEO Sebastian Siemiatkowski explains why we're so transparent about our AI capabilities and what concerns us most about new technology. This interview has been edited for length and clarity.
AI taking over human jobs is a sensitive topic. Why did Klarna decide to publish data about its use of AI to replace customer service agents?
We are generally concerned about the impact this has on society, so we decided to be upfront about the fact that we had great results for our customers.
Currently, we perform the equivalent of approximately 700 full-time jobs. [customer service] Agent. Klarna does not employ its own customer service agents and uses some of the leading customer service providers. They are outsourced and are not employees.
We announced that as a result of implementing this technology, we will require, on average, 700 fewer full-time agents than usual. On average he needs 3,000 agents, but currently he needs just over 2,000.
We wanted policymakers to recognize that this is happening now, not in the future. We believe it is important for society to start thinking about this major change.
So, does this have anything to do with the job cuts that Klarna implemented in 2022? Could AI have taken those jobs away?
In 2022, as a result of changes in investor sentiment, we had to downsize our company. The number of employees was reduced by approximately 800.None of them [the people we laid off] I was a customer service agent. Now, two years later, we are looking at this interesting technology we developed at ChatGPT. These two are completely different and are confused.
In the short term, Layoffs or impact on employees This is the result of launching this customer service AI chatbot.
What does a chatbot do? How do you measure its success?
Two-thirds of customer service chat inquiries are handled by this service. Satisfaction levels were on par with humans, and repeat inquiries from customers were reduced by 25%.
in front We filmed this live, but we already had a co-pilot to help speed up the work of our customer service agents and other employees. This AI actually communicates with itself and solves customer problems. While we think it's important for people to have the option of talking to a human, we've found that people who choose to interact with an AI chatbot are very satisfied and that it's more focused and effective. We also make sure that we feel that we are on target. Being able to communicate in 35 languages ​​means a much improved experience for many immigrant and expatriate groups.
Larger organizations have less time to spend creating real value for their employees, customers, and shareholders. AI will take away less productive jobs. In the best cases, everyone joining the company will be able to use their creativity to create real value for customers, rather than creating impressive presentations for the CEO. It's not an added value. It's those kinds of jobs in particular that I'd like to downsize.
Are you worried about making a mistake or providing inaccurate information?
Unfortunately, we need to remember that we humans are not perfect. Humans are great, but we also make mistakes. [give a query] They should do their due diligence and training, but it's not necessarily their fault. The point is, especially when you have millions of customer inquiries, it does happen that some inquiries aren't as good as you expected.
A key metric we use is to see if they make fewer mistakes than humans on average. That's the goal and we've been monitoring it closely. We make sure that AI mistakes are not that serious or common.
How, if at all, has AI changed your approach to recruiting?
We have stopped hiring in the past 6 months. We are shrinking as a company, but that is due to natural attrition, not layoffs. At Klarna, we're applying AI to every product, service and job we do. It affects the number of people a company needs. This is one single product improvement that has significantly reduced the need for customer service agents.
How do employees feel about this?
It's not easy not to talk about this, but given the impact on society, we want to be honest and transparent in telling people about this. Regarding our employees, we have been working hard to communicate this internally since we stopped recruiting in October. We believe this allows us to do more with less.
Our goal is to increase our investment per employee and increase compensation for existing employees as we become a highly profitable company.
Internally, many of our employees are excited about and leveraging AI, and Klarna is known as a leading AI company, so we think that makes our people more attractive to other employers. Masu.
How was the estimate that AI would save Klarna $40 million in costs achieved?
We were purchasing customer service from our suppliers to respond to inquiries. We were budgeting how much we would spend on these suppliers this year. Now, thanks to technology, we spend much less. 40 million is the difference between what you expect to spend and what you actually spend.
How big an impact do you think AI will have on society?
We are sharing this information to ask society as a whole how to respond to this major change.
From what I've heard, this seems to be more about general safety than job security. This means we need electronic device ID to know that we are actually talking to one another.
Also, on the other hand new [AI-related] jobs could be created At a societal level, it is not an individual solution [who is replaced by AI].
The more we as a society can support individuals who may be affected, the more positive development we can expect. The goal is not to stop it. I don't think that's correct. The question is: What support can we provide to those affected?