Klarna and AI: Conscious Decoupling

I’ve never been a fan of Klarna or its ilk (Affirm is the first that comes to mind): my unimportant gut feel is that people are deciding to make a purchase first, and then Klarna comes along and entices them to extend the payment terms of the purchase, and Klarna pockets some above-market interest on the transaction. I see these offers on my credit card statements, when I buy things online, and in advertisements. Did we as a society actually not consume things before Klarna came along? I don’t think they’re solving an actual problem.

So I was interested in this interview with their CEO, because I’ll read anything that indicates that Klarna is having a bad day, and their recent Q1 performance, with a 17% increase in credit write-downs, was just that. The interview actually focuses on their climb-down on the use of generative AI in customer service. Like most tech companies, they had really amped up how AI was going to replace x% of their customer service staff, with no deterioration in service levels. As most of us have experienced AI-driven customer service can attest, it hasn’t worked out so well. And now the enthusiasm and peppy language swings the other way: “Really investing in the quality of the human support is the way of the future for us.” says the CEO.

I’m certainly happy that he has seen the light, or been forced to see the light, but the frequently observed blind faith that qualitative support and human interaction can be replaced with GenAI rings false. As much as we drive each other crazy, there are few better customer experiences than dealing with someone who a. seems genuinely pleased to meet you; and b. actually knows what they are talking about. A purchase or investment is an act of putting your neck out, and getting some assurance that you’re making a good decision makes the difference between transacting or walking away. It is nice to see some vendors ‘get it’, but Klarna didn’t have to waste a ton of money and goodwill to figure it out.