An AI-generated illustration of a humanoid robot customer service assistant sitting at a desk and typing on a computer.
Klarna used Generative AI to answer 2.3M customer questions and drive a USD$40M profit.

While many companies are exploring Generative AI, most haven’t been using it long enough yet to show clear, measurable results. Klarna, a Swedish shopping services company, was an early adopter of generative AI, and they now have the numbers to show it’s benefits: 90% employee adoption, AI doing the work of 700 customer service reps, and $40M profit improvement using Generative AI. Their story offers powerful insight for other companies exploring generative AI for their business. Let’s learn more!

Why is Generative AI a Fit for Klarna?

Klarna is a fast-growing financial services companies based out of Stockholm, Sweden. In 2023, They saw 22% YoY revenue growth, earning SEK$23.5B on SEK$981B in gross merchandise volume. They offer two connected operations:

  1. A buy-now-pay-later offering that lets customers split large expenses into multiple smaller ones
  2. An e-commerce platform that provides payment processing and claims management for companies, as well as coupons and order tracking for end-users.

With nearly 5500 employees, their business activities are very knowledge-focused, with an emphasis on software, infrastructure, and customer service. This means their business is well-positioned to benefit from a range of Generative AI use cases.

Achieving 90% Generative AI Adoption by Employees

Shortly after OpenAI released ChatGPT in late 2022, Klarna executives quickly reached out to see how they could integrate OpenAI’s generative AI tech into their business.

Rather than driving integration from the top down, leadership instead put an emphasis on employee independence by strongly encouraging employees to experiment with the new tools. In their press release, CEO Sebastian Siemiatkowski said “We push everyone to test, test, test and explore.” Key teams integrating Generative AI in their business:

  • The communications team use ChatGPT to provide an objective analysis of press articles about the company. Head of Communications Filippa Bolz says this “helps us ensure we communicate clearly, that it is understood and resonates with each one of our target audiences – rather than bundling them together.”
  • The legal team use ChatGPT to create bespoke contract drafts for further refinement. Selma Bogren, Senior Managing Legal Counsel, says that this approach can reduce an hour’s worth of work down to just 10 minutes.

Another central driver behind AI adoption has been Kiki, a version of OpenAI’s ChatGPT fine-tuned as an internal AI assistant for Klarna employees. Kiki’s focus is on “management and distribution of internal knowledge“, answering 2000 employee questions daily, and over 250k questions total to date. The press release notes: “Kiki generates answers to a wide range of topics within 1 to 5 seconds, and answers are context dependent. So the answer to ‘what is my team working on’ will be personalised to the person asking the question.”

Generative AI Does the Work of 700 Customer Service Agents

Along with using Generative AI extensively internally, Klarna has also incorporated AI into their customer support processes, with stunning results.

As of February 2024, their customer service assistant had been used in 2.3M conversations, representing 2/3 of Klarna’s customer service chats. This is equivalent to the work of 700 customer service agents! (And, for a company experiencing as aggressive growth as Klarna, a huge deferral for hiring and staffing!) Critically, the AI assistant performed on par or better to human agents, with a similar satisfaction score, 25% drop in repeat inquiries, and an 85% improvement in resolution speed. Klarna highlights this as a critical driver of a USD$40M in profit improvement last year.

An underrated benefit of this assistant is it’s ability to serve Klarna’s uniquely diverse customer base. Not only does it support operations in 23 different markets 24 hours a day, 7 days a week, but it also serves customers in 35 different languages. AI translation technology has been a major beneficiary of advances in generative AI, and will be a significant aid to international expansion for businesses going forward.

The Numbers Don’t Lie

A major challenge with Generative AI to date has been a dearth of companies sharing hard metrics on their results. Much of the business world is still in an exploratory phase with the technology; employees are still finding the best use cases for Gen AI, with impact measurement being secondary for now.

That’s why this information from Klarna is so valuable. Not only have they demonstrated Generative AI’s value in multiple business contexts, but they’ve demonstrated to themselves and the world that it can deliver real, substantial value to both the customer and the bottom line. Many thanks to both Klarna and OpenAI for this information.

Stay informed on the most important AI news in under 30 minutes a month by signing up for the Executive Summary newsletter by AI For Business.

If you liked this post, have future posts delivered straight to your inbox by subscribing to the AI For Business newsletter. Thank you!