Klarna on the fight for ‘top of wallet’ in an AI agentic commerce world

Sebastian Siemiatkowski:
If you go back, you’ll find a lot of funny clips from the ’80s where people are arguing, “Why would you use your credit card at Burger King?” People think it’s insane in the ’80s and today’s it’s standard, right? We don’t think about it as a problem. So I just think it’s the novelty of this and the difference of it.

Liz Hoffman:
Welcome back to Compound Interest from Semafor Business. I’m Liz Hoffman, Semafor’s business and finance editor, joined as always by my colleague, Rohan Goswami. Hey, Rohn.

Rohan Goswami:
Hi, Liz. How are you doing?

Liz Hoffman:
Question for you.

Rohan Goswami:
Okay.

Liz Hoffman:
What is the last big ticket purchase you made and how did you pay for it?

Rohan Goswami:
Big ticket purchase. I don’t know when this is going to air, but probably my anniversary gift, which was last week.

Liz Hoffman:
What did you get?

Rohan Goswami:
I don’t want to spoil the surprise she does shockingly listen to the show. It was a (beep) and I paid for it using my credit card.

Liz Hoffman:
Okay. My last big one was a shocking number of dollars worth of wood flooring doing some renovations.

Rohan Goswami:
I remember this saga.

Liz Hoffman:
And I paid for it with a buy now, pay later loan. I asked because our guest this week is Sebastian Siemiatkowski, who’s the CEO of Klarna. Klarna launched 20 years ago, an eternity in startup land, as a buy now, pay later company offering interest-free loans to online shoppers and going after the last great racket in finance, which is credit cards.

Rohan Goswami:
Indeed.

Liz Hoffman:
And it’s been broadening its business away from that, trying to turn into more of a bank doing things like debit cards and loans that charge interest. And that is weird to me because I came up covering banks and the generation of FinTech companies that Klarna belongs to, which launched as just proudly anti-bank. Regulation was for suckers and being a bank was lame. And one by one, they have all caved PayPal, SoFi, LendingClub, Revolut. They just one by one either bought banks or became banks. And part because it turns out that the stuff they wanted to do just you needed a banking license and really you needed a balance sheet, the ability to extend money without having to rely on very fickle investors in the background.

Rohan Goswami:
And of course though, the downside to that model and the reason why they at least very loudly said they didn’t want to be banks is because regulation is hard and compliance is even harder. And when you are a FinTech, you can kind of skirt some of that and take the very Uber-esque, move fast, break things, figure it out later approach, but you can’t when you’re a regulated entity like a bank.

Liz Hoffman:
I mean, regulators really don’t like to see banks growing very fast. And in fact, we talked to Vlad Tenev at Robinhood a year ago about their banking ambitions. He said that’s what had held them back before was that they were just posting this double-digit annual growth that would’ve had regulators crawling all over their office.

Rohan Goswami:
[inaudible 00:02:42].

Liz Hoffman:
But also the other reason is that tech stocks trade much higher than bank stocks [inaudible 00:02:46] nowadays.

Rohan Goswami:
It’s all about the multiple. It’s all about the multiple.

Liz Hoffman:
And as you’ve seen Klarna get more and more into banking, I mean, their stock price is down I think about 50% over the last six months.

Rohan Goswami:
It has not been great, but that’s also hard for them because they are, and I think Sebastian has been really front-footed about this maybe to his chagrin, maybe too much. They are sort of at the forefront and the center of this debate over what role does AI, what role do agents play in finance, in our financial lives? How does that look? What does that feel like?

Liz Hoffman:
They also were early to stop being a Salesforce customer in something that looking back was kind of the first shot fired in the enterprise software war that has turned into the SaaSpocalypse. I know he said he had some regrets about how that all was kind of communicated.

Rohan Goswami:
We can get into that with him.

Liz Hoffman:
And remains a fan of Salesforce and thinks they’ll be just fine.

Rohan Goswami:
But Sebastian is obviously the person best place to answer these questions. He’s waiting for us. Why don’t we bring them in?

Liz Hoffman:
Sebastian, welcome to the show.

Sebastian Siemiatkowski:
Thanks for having me.

Liz Hoffman:
Let’s start here. A whole generation of FinTech companies launched by saying, “We are not a bank.” One by one SoFi, LendingClub, Block, Revolut, they’ve all kind of moved in that direction. Obviously, so have you and so are you. Why?

Sebastian Siemiatkowski:
Oh, good question. I mean, I can’t answer obviously for the others, but I can answer for us. Many people told us that it’s so difficult to become a bank, it’s so hard being regulated and all these things. So for a while we were cautious about it as well, but we got our first banking-like license after about five years of existence back in 2010. And to us, we only saw advantages. It created clarity on how to operate and run our product in compliant ways. It gave us access to things like deposits that allows us to fund the business more and kind of removed some of the structural competitive barriers versus the banks themselves. To me, it’s just been advantageous rather than the opposite.

Liz Hoffman:
Is there something cultural that kind of got lost in the process? The idea that there was a way to serve these customers without the regulatory overhang, or frankly, just the lameness of being a bank?

Sebastian Siemiatkowski:
I haven’t seen that. I mean, I’ve seen obviously that potentially there’s more compliance work, there’s more second-line work, there’s some more process and some more legal requirements that needs to be met. All of these things that are required of one are actually good stuff if implemented well. What I hear when I talk to my incumbent CEO is that in many organizations seems that those are treated as side things that are bureaucracy and only there to tick the boxes and then they don’t necessarily serve their purpose. But if you go and read the law carefully and you understand the prerequisites and the ideas that the regulators had with these, you genuinely see a lot of them are actually quite smart.
Take as an example, we have something called a new product approval process. So you need every new major product approval needs to be combined with risk assessments, financial implications of those changes, product and consumer implications and so forth. That’s a pretty good thing to do and I would argue most good tech companies would apply that anyways, but what ends up happening many times is things gets implemented poorly and then instead it just becomes bureaucracy.

Rohan Goswami:
You have this license in Sweden, you have it in Europe, but not in the US. And you’ve done what a lot of FinTechs have done, which is to partner with smaller banks to make loans behind the scenes. Do you plan to apply for a de novo license here?

Sebastian Siemiatkowski:
We have not publicly commented on that. We haven’t given any [inaudible 00:06:28]. But I think it’s like what I try to say at least is directionally speaking, it would make sense for us to do so over time in the US as well, because we have seen in Europe a lot of advantages of it, but there’s no official timeline or decision made that we would have shared publicly today.

Liz Hoffman:
The administration here is pretty friendly to that. They seem to want some new banks.

Sebastian Siemiatkowski:
I’m very happy to hear that and see that because it was actually a big surprise to me coming from what is usually treated as less competitive and less capitalistic Europe versus US where when it comes to getting a bank license, what you had to do in Europe was just go and apply. And historically in the US it was like, it is not clear whether you will get one and nobody can really tell you what is required to get one because right now we don’t want more, or the regulators don’t want more banks. That was kind of the message we used to get back in the days. So I think it’s a great improvement.

Rohan Goswami:
The competition for FinTechs moving into the banking space, it’s pretty fierce. What’s your edge in the banking space over the other neobanks that have already gotten started here?

Sebastian Siemiatkowski:
I’ve given that a lot of though over the years. I used to have Nigel Morris on my board, co-founder of Capital One. I remember back in the days we were talking about country expansion and how to establish one. And at that point of time, one of the most coolest popular things were, “Oh, we’re going to use Facebook data to do social underwriting and people are going to use that to improve ...” That was one of the big things in FinTech back then, not something Klarna pursued but others were. And I remember speaking to Nigel about it and he said, “Well, look, for Capital One, even if that would turn out to be the thing, it would be like improving their margins a few BIPs. It’s not going to disrupt banking. It’s not what’s necessary if you truly want to go after the incumbents.”
And what we arrived at Klarna is that the key is customer acquisition cost. The fact that in the US it may cost 100, $200 to acquire a customer in the more subprime Capital One category and 4 or $500 if you’re Amex. So to me, the key was that a great app is nice. There used to be simple.com, someone may remember it back in the US, all of these things are nice, but you have to find a way to acquire customers at a lower cost. In Klarna’s case, the fact that we have this huge merchant network of over a million merchants offering Klarna where consumers can come and say, “Hey, I can either use my traditional bank card or I can just with a single almost click experience try this new thing called Klarna,” has become a very efficient customer acquisition channel that has allowed us to achieve 130 million users as basically zero CAC. So that is our entry point, but other neobanks team have found others and we are happy to see even in the US we’re close to 30 million through that mechanics.

Liz Hoffman:
Is the goal to be a better PayPal or are you really trying to take on JPMorganChase ?

Sebastian Siemiatkowski:
What I find attractive is the position that Amex has. And I think there’s something unique about Amex and we are very deeply inspired by what they’re doing. We call it spend centric rather than lend centric, which is the similar terminology that Amex uses because different than your kind of lending, there are a couple of players in lending, there’s Affirm, there’s Synchrony, et cetera. They turn around their books maybe two times a year. So they are predominantly a balance sheet lender. While Klarna is predominantly spend centric, we turn around our balance sheet more than 10 times a year and that changes the characteristics of the business, but most importantly, it grows a different type of relationship with the consumer. Something that you use very frequently, you have a different relationship with than something you just took a loan with at some point of time when you did a big ticket purchase.
So for us, we think that’s the position to be in. Now, what people tend to forget is Amex is the 10th largest bank in the US, but people still don’t talk about them as a bank. And I think that’s the route that Klarna is at. We’re trying to create a differentiated service, a different brand, a different value for our customers. And some of the services we provide are bank-like just like Amex does, but the key is the lifestyle, the brand, the customer service of what we do.

Rohan Goswami:
Is the goal then for people to say, “Oh, I’ll Klarna it,” to become a verb?

Sebastian Siemiatkowski:
It’s fun obviously when people love the product to that extent where they use your name as a description for what they’re doing. But I think the key for us is that we want to be relevant for every type of purchase that the customer do, both the big ticket spend, which is a category we just entered into the US more recently or have scaled more recently. The mid-ticket charge card equivalent type, which is the $100 type of purchase, fashion, et cetera, and the everyday spend, the groceries and so forth. So we need one that is relevant for all of these, but obviously we’re mostly associated with this 0% interest pay-in-four product, which has been extremely popular and well received in the US.

Liz Hoffman:
You’ve moved into these longer term interest bearing loans that do make money not just from the merchant but from me as a borrower paying interest over the life of a loan. I mean, I think you’ve talked about that as a substitute for credit cards. How do you think about a consumer that takes out a 24-month Klarna loan that perhaps they can’t afford versus a consumer that takes out a credit card that they can’t afford?

Sebastian Siemiatkowski:
First of foremost, we do not offer revolving. We think revolving is a bad concept. It’s confusing. It’s difficult for people to understand how long will it take and people tend to pay off for a much longer period of time. So we only do fixed term installments. The other one is we try to make sure that our interest is fair, that it’s well priced compared to any other alternatives that consumers may have for financing that particular purchase. And many times we try to work with merchants to make sure that they are funding the interest so it’s a 0% offering, which also is more attractive than using your credit card and borrowing at a higher rate. Then there’s also a third thing we’re trying to be mindful of not pushing your limit into your face and we’re trying to find a good balance there because we know that sometimes the limit can be misunderstood as I have more money than I believe I have.
And then obviously that the primary product that we are promoting is the charge card equivalent of buy now, pay later, which is 0% and pay-in-four installments. That is basically a charge card product, an equivalent of a charge card product. And that one is also, I would argue, much healthier, just like Amex has been pushing charge cards more than revolving for many, many decades. So I think it’s all of these things where we try to make a difference, but obviously as anyone issuing credit, you will still have some consumers that unfortunately overextend themselves or use the product in a way it’s not intended to be used and so forth. It doesn’t matter if it’s DoorDash or groceries or everything is on your credit card and then you are more likely to revolve. And this is why the average outstanding balance on a credit card is like 5, $6,000, while the average outstanding balance with Klarna is a hundred.

Liz Hoffman:
All right, let’s get into the DoorDash discourse because it’s been something that has just been in the news a lot lately, the sense that particularly younger people who have limited incomes are just spending way too much of it on delivery. You guys had a product last summer partnered with DoorDash to be able to pay for food delivery and installments. It feels like there are more ways for people to spend money that they may not in fact have. And I’m curious how you think about that, what you made of the burrito bond discourse last summer.

Sebastian Siemiatkowski:
Ironically, I saw the same when we launched with Deliveroo in the UK a few years earlier. Because Klarna is so strongly associated with buy now, pay later, that always becomes like a gimmick or an interesting point as you highlighted. But you have to remember we’re competing with PayPal and we offer debit paying in installments like small or medium-sized tickets and big tickets. What we’re trying to do is make sure that we’re relevant in every checkout for every purchase. So going into things like Uber, DoorDash and so forth is part of that PayPal strategy. Nobody is talking about the fact that PayPal is available in these apps such as DoorDash. Nobody talks about the fact that you can use your credit card and revolve on your DoorDash purchase. So one has to separate with Klarna as a merchant network wants to be relevant for everyday spend groceries, gas stations with our debit product predominantly and in the mid-ticket segment with our buy now, pay later and in the big ticket segment with financing.
Now occasionally may that mean that a customer goes to DoorDash and puts a fairly small purchase on four interest-free installment. Yes, it can happen just like people will go there with a credit card and revolve on a purchase they did at DoorDash or a balance that was partially DoorDash. So the same thing may happen, but that’s not the intent. The intent is to give availability to use Klarna at any merchant for all the three different payment products that we offer. It’s interesting that in a lot of countries in Europe where Klarna is widely known to be used for a different type of payment methods, this is not really a debate. You won’t see this in these countries because to them it’s just like, “Yeah, of course they’re there, just like Visa and MasterCard are there. But in some markets where we’re strongly still only associated with buy now, pay later, then this becomes a bigger thing.

Liz Hoffman:
But this was a bunch of US reporters losing their minds??\

Sebastian Siemiatkowski:
No, I wouldn’t say that. You said that, but I would just say that again, I understand why, because we’re so strongly associated with this super successful pay-in-four product and so everyone just inherently assumes that’s what we’re doing. We have a lot of other interesting products. One that really excites me is that another which is like you can take, for example, for somebody like Uber or these kinds, you can aggregate all of their purchases and give you a charge card equivalent product where all of that gets paid at the end of the month, which for a lot of merchants that see a lot of low value transactions where Visa and MasterCard charges a fixed fee in addition to the percentage fee privilege transaction, that’s an amazing opportunity to drive down payments costs. And it’s not different.
I mean, you go back, you’ll find a lot of funny clips from the ’80s where people are arguing like, “Why would you use your credit card at Burger King?” People think it’s insane in the ’80s and today it’s standard. We don’t think about it as a problem. So I just think it’s the novelty of this and the difference of it.

Liz Hoffman:
One thing we talk a lot on the show about is how everything is cashflow now. Where do you see the sort of natural endpoint of everyday transactions being financialized and ending up inside some portfolio somewhere, which I assume is the endpoint of, some of yours, but certainly a lot of firms loans end up getting sold to Wall Street.

Sebastian Siemiatkowski:
It’s interesting because I’ve been in this for 20 years. I remember that after ’07, everyone told me that the idea of packaging and selling off loan portfolios is dead and it’s never going to come back and here we are almost 20 years later and it’s becoming a de facto standard. I was one of the ones who always believed that that model was superior. In the long-term perspective, if you think about it X amount of years in the future, I find it very likely that providers like ourselves will basically to some degree be market makers, meaning we recognize the consumer, we can underwrite them, we understand the risk profile of that consumer based on our internal data as well as combine it with some external data and we are then going to allow people in the market to bet on providing financing at the most affordable rates to that particular cohort of consumers. I think that model makes a lot of sense that you have a separation between that, the kind of underwriting decision and the origination from who provides the balance sheet and who’s willing to invest in that.
Now what we also do know, however, is that over economical cycles, we have seen sometimes the problem that such markets may dry up on very short notice and that’s why from our perspective, we decided to become a bank and fund by our own sourced deposit base, which is still like 90% of what we’re doing today because it just gives us a higher level of stability and predictability in the funding and companies that don’t do that may ... We’ve seen that. I’ve been here in FinTech for 20 years. I’ve seen some of that happen over ... What it was, I think [inaudible 00:18:43].

Liz Hoffman:
LendingClub got itself in a lot of troubles.

Sebastian Siemiatkowski:
Yeah, there was a number of companies that ended up in that situation where suddenly these things dry up and that just becomes a problem. So for us, it’s been important to establish that as only one form of financing, but longer term does it make sense that that is a very popular and dominant form of financing? I would say yes, it makes sense.

Liz Hoffman:
But actually then to make that comparison, you’re talking about the mortgage bond market, which obviously blew up in the 2000s, but that market was at least underwritten on the assumption that people would keep paying their mortgages, that that is the last bill they would possibly skip and that bet in fact turned out actually to be wrong. Are people going to feel the same way about their buy now, pay later loan and how are they going to feel about continuing to pay that down the road?

Sebastian Siemiatkowski:
You have to think about two aspects. In the mortgage world, almost the whole thing is, can this person afford to service this loan over a longer period of time and what’s the risks that unemployment hits them or something else happens under that time? And that is an important factor for us as well. But in our case, also an important factor is just behavioral because our outstanding loans are about $100. The durations are very short. As I said, we turn around our balance sheet more than 10 times a year. So what ends up happening many times is if we see a macroeconomical shift or something happens in the economy, we just change our underwriting in real time and it takes only about 60 days for more than 50% of our balance sheet to be underwritten by the new standards. That’s a level of agility and flexibility that the big banks don’t have.
That’s also a downside. If you’re pushing people into revolving and building large balances like the big credit card companies are, when economical changes happen, you just can’t shift out of it as fast because you have all these customers with all this debt. But with that said, also to your point, there’s also like a preference. Why would consumers pay you versus somebody else? These things play in. And I think that’s why being a beloved brand, having preference, being a service that consumers appreciate and pay actually does matter for repayment rates as well, especially in US, actually a little bit less in Europe where the market works differently.

Liz Hoffman:
You’re saying if the economy turns, the big banks in the US are still going to be writing garbage loans long after you have stopped?

Sebastian Siemiatkowski:
No, their origination will change and their underwriting will change. They can change that pretty quickly, but the problem is they’re stuck with whatever they originated for the last years. I mean, a general bank in general has about, I think 80% of their balance sheet has been originated is two, three years previously. They just don’t have that agility to change it. Even if they change their policies today, it takes a lot of time before the majority of their balance sheet is underwritten by the new policies.

Liz Hoffman:
I do want to make sure that we’re kind of unpacking the risks here a little bit and not making light of them or too much of them. We have seen securitization markets bring down the entire economy before and that was based on a product that people really feel strongly about, which is their house. What is the risk that a lot of this buy now, pay later financing for products that people bought a while ago and forgot about, or again, or vacation that they already took ends up in these complicated bonds spread across Wall Street and that they start to go bad. Are you worried about that?

Sebastian Siemiatkowski:
I’m not worried about that at all. I’ve done this for 20 years and as I said already, I think I’ve answered with the durations so forth. I think for anyone that’s watched The Big short, you also I have to remember why did the mortgage crisis really happen was because the underwriting was really poor and there was a number of big companies that went and guaranteed that that underwriting was not poor, that it was actually great, which was wrong. And so I think in our case, because we’re not securitizing, we’re putting this on our own balance sheet. We have all the incentives in the world to be very mindful and thoughtful about what we underwrite and what the repayments rate. So most people can pay back a hundred bucks. It’s very different, the mortgage, if you’re underwater, like you’re underwater, most people can pay back $100. So if you’re thoughtful and you keep your loans small, you fare well even in economical environment changes. And that I’ve seen with Klarna over the 20 years that we have operated the business.

Rohan Goswami:
Let’s take a quick break there. We’ll be back with more from Sebastian after this. Shifting gears, President Trump last year made some noise about a 10% interest cap on credit cards and that kind of died down, but what would that mean for your business, for Klarna?

Sebastian Siemiatkowski:
In general, I am, as I believe a lot of others as well, very much in favor for free capitalism, for having as less rules as possible. But I also think most people would agree with me that anarchy isn’t great, that in all business there’s value to having some rules of how we interact. And I think that what was wise about that suggestion is it recognizes that in some markets, sometimes some rules may be helpful.
I think about this primarily first and foremost from a US consumer perspective where we are seeing a continuous growth of credit card, of revolving debt, and that it might make sense to add one or two rules to help counteract that. The implications for Klarna, in my opinion, is that it makes us stronger because it makes the offering, this is the offering that we’ve been offering. We have had paying for 0% interest and if the rules that makes the other products more similar to ours, I think it just validates what we have built and the benefits that we provide. So I think in general, it would be a good thing.

Liz Hoffman:
Well, it might validate it, but it also might totally undercut it if JPMorgan is offering me the same product that you are, they have a lot bigger distribution than you do.

Sebastian Siemiatkowski:
You could say that, but on the other hand, what I’ve seen in Europe is that when such caps were introduced that they have been, it meant that the whole loyalty model, all the loyalty points and so forth became less attractive because it was more difficult to execute that. And it levels the playing field as well. The whole idea is that we educate and our consumers see the value that we’re offering and are not lured away by offers that initially may sound very attractive, but in the end are all about you maxing out on revolving $6,000 at 30% interest rate. Right now we’re seeing a big benefit from the fact that people learn themselves that those models are not healthy. We call them the self-aware avoiders. They’ve tried credit cards, they don’t like it, they found themselves in too much debt. They pay it down and they start using Klarna instead.
They’re pros and cons, but generally speaking, I’m always going to speak up for what I think is right for the consumer.

Liz Hoffman:
Let’s talk about AI. Klarna was one of the first companies to really collaborate with OpenAI, cut a bunch of jobs in 2024, replaced them with what at the time we called chatbots. I guess today we’d call them agents. It didn’t go that well and you backtracked. Talk a little bit about that experience and what you learned from it.

Sebastian Siemiatkowski:
That is actually incorrect. And this is one of the kind of myth busting that I think independent of what our PR department tries to do, we’ll probably not be able to get the facts right here because it’s just been resent and reshared so many times that it’s beyond us by this point of time. So let me give you quickly the facts. What we did, we launched an AI customer service bot. It took about the equivalent of 700 agents jobs, but since we’re outsourcing nobody lost their job, they just went and did other jobs for other companies, but it reduced and had a significant impact on our cost base and it helped resolve the most simplistic errands at a higher quality that the consumer appreciated.
Since then, we have increased that and it’s now doing the jobs of about 850, so it’s slightly more than it was back then. What we also tried to communicate and failed in media was that we said, “Look, we have a very interesting learning from being one of the early deployers of this and that is that in a world where AI can do the most simplistic customer service, we believe that human customer service will almost be seen as a VIP thing. It will be one of those things that you’re like, “Oh, wow, I talk to a human”″ And we want to offer the best customer service in the world, but in order to do so, we would also have to change because historically Klarna was not necessarily thinking about customer service as primarily from a quality perspective, but primarily from a cost perspective. So it had to change, we had to recruit a different type of customer service agent. We had to be willing to pay thereafter and we have to give them the tools and the prerequisites to be successful in serving customers in a better way.
Now this then got misunderstood as like we were saying that quality of AI was bad and there’s been tons of misquotes, but that’s what’s actually happened. And then in addition to that, we have also worked using AI internally and it allowed us to become more efficient and through natural attrition, not many layoffs, but through natural attrition, we have shrank. We used to be about 6,000 people, I would say, and then we have become fewer, partially through efficiency and partially through the application of AI. We’re now below 3,000, about 2,700. So those are the actual facts of what’s been going on in there. And we think it’s an amazing tool that accelerates the business and it has tons of opportunity to improve quality of service for our customers.

Liz Hoffman:
But just to be clear, the problem was it that the AI sucked at the job?

Sebastian Siemiatkowski:
No, I mean AI sometimes isn’t correctly hooked up. And mostly it’s about you not providing it the right context. I think most people today would argue if you give AI the right context, it’s actually very capable in a lot of situations. But if your source system is not providing it the right context, it may miss.
But there’s also other aspects For example, if a customer is unhappy or feels that a decision is unfair, most people will say, “I’m not accepting a robot telling me that, give me a human.” And I think that’s a very natural and fair request of a customer to be willing to say, “I don’t accept this robot telling me, no, give me a human to speak to.” And we want to offer that service to our customers. We think they should be entitled to have that service. So there will be situations and there are situations in which humans are the right answer to helping customers in a more thoughtful and empathetic and qualitative way.
Obviously all technologies that you launch, you will always find single instances where that technology hasn’t functioned as intended. That was true for our app before AI came along, that was true for our services. You will always find instances where that’s not the case. And I think the same applies here. I’m a big believer that AI will be helpful to humans and businesses going forward and so forth, but I also wanted to send the message that I think there is a real place for humans in the future state of banking services as well, because I do think that empathy and the availability and speaking to a human is something that people will appreciate.
My wife wisely says two things can be true at the same point of time, but in media, it’s sometimes difficult to communicate two things at the same point of time, which was our learning. In the end, Klarna’s objective is to drive a fantastic product to its customer to help give consumers an alternative to credit cards and the traditional banks and that’s what we’re going focus on whether we apply AI as a technology to do so or not, we will do so when the quality and the reception is positive and it helps us bring better services at higher quality with lower costs.

Liz Hoffman:
Speaking of learnings, you got rid of Salesforce, I think, about a year ago and made some comments about doing something similar with Workday and what now looking back feels like the first shot fired in the enterprise software war, what have you replaced it with? Have you gone back to it and how’s it going?

Sebastian Siemiatkowski:
I’m not going to go into the details of the SaaS story anymore, but what I can say is that there was an important thing that we believed is that for Klarna as a company, we came back to whether it’s AI or just software in general, you want to make sure that you have the right context for whatever decision you’re going to make, if you’re going to help a customer, you want to understand all the different data points that are important in helping that customer with whatever errand that may be. And one of the conclusions we did back then was that the problem is if you have your data spread out over too many systems, it just becomes harder to integrate and extract all of that information that is necessary to support the customer in the best way. And that was the primary objective of why we decided to reduce the number of software vendors because we saw that it didn’t help us in that regard.

Liz Hoffman:
But you’re saying that this was less about vibe coding a cheaper alternative than taking a look at a bunch of systems that weren’t totally working?

Sebastian Siemiatkowski:
Vibe coding is a term that I would argue doesn’t exist anymore. Vibe coding was a term that was created when AI was producing code that was subpar to traditional engineering code. And I think most people that are in the industry and close to AI and coding would argue that since November last year, the code that AI today produces is on par with what human engineers does. Now, that doesn’t mean you don’t need to code review it or verify it and it may make mistakes just like humans does, but today anyone can code. Not everyone can verify that the code is of quality, but the code that AI generates is of high quality today.

Rohan Goswami:
You guys are a tech company, but then also commerce is poised to look really different, agentic commerce in the next few years. What does that actually do to your business? You’ve talked to us a lot about loyalty and about customers choosing to go with Klarna, but what happens when I delegate buying a mattress to an agent and say, “Look, I don’t want to spend more than this. I’m a sidesleeper,” all these attributes. And then finally, “I don’t really care how I pay for it. Find me the lowest cost of capital.” And the agent is suddenly the one making the decisions. How do you advertise to that agent and how do you win against an Affirm or conventional lending product or just cash?

Sebastian Siemiatkowski:
Well, I think again, it comes back to what we talked earlier about Amex. Would you argue that the situation described is a threat to Amex and the traditional credit card companies?

Rohan Goswami:
I use my American Express when I’m doing big ticket items because well, sometimes they give me pay-in-four promos, but also because there’s purchase protection. I know that if there’s a defect with my phone or something else that I buy that Amex will, if the merchant won’t cover it, cover it for me and that’s a huge appeal for me.

Sebastian Siemiatkowski:
So it’s actually funny you say that because buyer’s protection that Klarna offers as well is actually one of the most appreciated features. So I think it comes back exactly to what you said is that in my world, what we concluded over doing this for 20 years is that the key thing is consumer preference. If you have preference, it doesn’t matter that much how many buttons are in the checkout or what happens in the world with new things because you have that brand affinity and that’s what Klarna has been focusing at. When, for example, Klarna is side by side with some of the other buy now, pay later providers, I look at what’s the share of checkouts that we get? How many people click our button versus others? And what we see there clearly is we win. We win on preference. We get the higher share of checkout in those situations.

Rohan Goswami:
I guess I’m talking about a world where the consumer doesn’t see buttons where all they do is they put into their chatbot or whatever wrapper they’re using, “This is what I want to buy.” And the consumer never even touches a checkout. The agent does all of it.

Sebastian Siemiatkowski:
I think it comes back to what I said, brand preference. And I think at some point of time, you need some funds somewhere that you need to settle somehow and you need to have some relationship with that company. If you look at it very long term, and I’ve spoken about this on some opportunities, that you can basically argue that financial services are a combination of a digital assistant that helps you in those every day and the balance sheet. Those are the two parts that customers offer. So we just have to be the preferred way in which people fund those transactions, see value in doing that with us, both from a cost perspective and quality perspective.

Liz Hoffman:
Is, Klarna, airport lounges the answer to all of this? Because that’s a lot of the value of Amex to people.

Sebastian Siemiatkowski:
I mean, we actually do offer a lounge. I mean, we have multiple tiers on our card. Here’s my very nice metallic Klarna card and it does give me airport lounge access. I don’t have a Klarna lounge just yet, but we do have access to other lounges. I do think that yeah, that could be the case. I think it’s consistently changing, why do people prefer an airline loyalty program versus a non-airline or loyalty program? All of these things continuously as the industry evolves, things change and you just have to stay at the forefront.
I think what’s interesting and what to us was the strong belief is that we believe that over time you need to have a very low cost customer acquisition channel, which we have established. We bring in consumers at a very low cost. We have a big consumer base of almost 130 million users now globally. And then it’s about offering those attractive and relevant payments options and financial services that are interesting for them. And that was always the key objective of Klarna. And I think it’s interesting to see today that among our competitors, we’re the largest network in number of users in geographical span and we cover the broadest set of payment methods, both everyday spend, midsize ticket spend and big ticket spends. And we believe it puts us very well-equipped for the future, whatever the future is going to bring.

Liz Hoffman:
If the future of commerce is, again, agents autonomously buying stuff for me in the background, does that make it more important for you to have these exclusive relationships like the one you have with Walmart whereas long as I end up buying the thing at Walmart, you’re going to be in the loop there?

Sebastian Siemiatkowski:
I think that the, generally speaking, sometimes we end up having an exclusive relationship, which is great and so forth. But if you would listen to internal conversations at Klarna, you would hear many times I told my sales guys, “Don’t worry that much about exclusivity,” because the DNA of Klarna is preference, not exclusivity. So what I look at mostly is merchants that offer us multiple options and then I look at what’s the share of checkout that Klarna gains because that to me is like, it really truly reflects what is the true preference of us versus some of our competitors. But then sometimes occasionally it’s also nice to be able to work together with a fantastic partner like Walmart because if you’re in exclusivity, it` also gives you some opportunity to launch new services, test new concepts, drive additional value for both parties.

Liz Hoffman:
That feels like a great place to leave it. Sebastian, thank you so much for coming. This was a lot of fun.

Sebastian Siemiatkowski:
Thank you guys. I appreciate it.

Rohan Goswami:
Let’s just start with the part of the conversation I was looking forward to the most and was, I’ll admit, a little surprised at his answer, which is agentic commerce. I feel like we’ve heard a ton about this world where somehow we’re going to bridge this gap and you and I are going to trust agents with our bank accounts and they’re going to be house managers for us buying things. And I was super excited to hear what their edge would be. And it kind of seems like he just thinks that maybe consumers will say, “I want a Klarna this purchase, go figure it out for me.” I was really confused by that.

Liz Hoffman:
He’s putting obviously a lot of stock in brand loyalty. I was surprised by it too. I think that this is moving really fast and there’s a world where there’s a lot of shopping happening behind a curtain where there’s really no preference expressed and I think brands are going to sort themselves into buckets of, “I really want this brand product for this thing.” Versus, “Just go get this thing.” And how you pay for it is going to fall into one of those buckets too. At some point, you’re just going to hand over your wallet in addition to your shopping list. And I think where Klarna fits on that wallet and how it gets preferenced and surfaced is a question that clearly he thinks a lot about, though he seems like pretty sanguine about it.

Rohan Goswami:
And taking that a level further, if Klarna and Affirm and JPMorgan offer this same cost to the consumer, an agent may of their own discretion say, “Well, Liz has a relationship with Chase. She has a relationship with Amex. She really likes these brands and they offer more protection. Let’s go with them.” So I actually kind of take the opposite attack. I think brands are not even going to be able to compete and the same is going to be true of Klarna where they won’t have a say unless they pay OpenAI or they pay Anthropic into what ends up actually getting used.

Liz Hoffman:
I was also struck by his comments about capital markets versus bank funding. On the one hand, he’s sort of a purest believer in capital market’s ability to match the dollar of funding that I need with the dollar of funding that someone out there has. That said, they are making a bet on having a bank balance sheet. And I remember covering this generation of FinTechs that launched in the 2000s and 2010s and they were so proudly anti-bank, they thought banks were losers and kind of watching them all come around to the idea that it is in fact helpful to have a balance sheet to do things that you want to do has been kind of a fun long arc of my own career.

Rohan Goswami:
The other delightful thing is that this is the second time I think back to back that a guest has brought up The Big Short, but do you think that’s a fair comparison? Is that the risk with going to the market the way Affirm does in selling these things or is it just completely different?

Liz Hoffman:
Klarna is a little bit of the origin story of this show, which is that last summer they came out with this partnership with DoorDash. And look, there are some differences. They are not, in fact, to my knowledge, selling burrito bonds, but I joked, what did I call them? Every good or terrible Wall Street product has an acronym, I called them NACHOS. Was it Non-Amortizing Culinary Holding Obligations, I think is what it was.

Rohan Goswami:
Yes, that’s what it was. That’s what it was, yeah, yeah.

Liz Hoffman:
The assumption of the subprime housing market was that people would never stop paying their mortgage and boy did that turn out to be wrong. And I think that the speed with which people will stop paying their DoorDash installment loans if things go south will be breathtaking. What Sebastian was saying was that they’re nimble and they can change the underwriting and sort of turn that spigot on and off quickly. But a lot of this buy now, pay later lending is ending up inside Wall Street firms that for their own reasons, having sort of nothing to do with a fundamental view on whether this is a good piece of paper, they need a lot of pieces of paper right now and that is a lot of money. They need assets.
And so I worry a little bit when you end up with any kind of financial product that is being driven by the off-takers, which is what happened in 2007. You had all of these big CLO managers and all of these big loan funds saying, “We need more product. We need more product.” And so the banks start making more loans. That is a very different proposition than, this particular financial product sucks, let’s make a better one, which is obviously where Klarna and Affirm have started, credit cards are extremely problematic. I worry that some of what’s driving it is shifting a little bit. I think less actually with Klarna than some others.

Rohan Goswami:
We didn’t talk to Sebastian about this, but I do want your thoughts on this. The K-shaped economy discourse has only gotten stronger and stronger and stronger. And you heard me bitterly complaining about how expensive US Open tickets are and how even the Amex presale couldn’t get me in. But then at the same time, we saw all the banks when the 10% cap was being floated talk about how this would actually slow down the economy, that the poorest people in our country are truly teetering on the brink.

Liz Hoffman:
There’s no doubt that for most purchases in any given moment, a pay-in-four interest-free loan from Klarna, you make four payments, you don’t pay any interest, it’s over relatively quickly is way better than a credit card for the vast majority of people. I think the question is you have to wonder what sort of marginal spending is it encouraging that perhaps we shouldn’t be encouraging, which is how we ended up in the great DoorDash discourse of 2026, which is-

Rohan Goswami:
The [inaudible 00:42:44].

Liz Hoffman:
... that your generation, Rohan, is spending way too much of their money ordering food. They’ve forgotten the virtues of microwave ramen.

Rohan Goswami:
Well, I completely agree, but then I also think we, I will actually soapbox here. It is social media’s fault. People look at a ... It really is. I mean, I don’t know if you see, your algorithm-

Liz Hoffman:
Your brains are just broken?

Rohan Goswami:
No, it’s very different than mine, but I think people see an unsustainable lifestyle very loudly trumpeted by people who are going out to dinner every night, and I realize I go out to dinner a lot of nights, but who are going out to dinner every night, who are going to shows who are sitting courtside and they say, “Well, why can’t I have it?” And Affirm and Klarna kind of say to them, “Well, you can buddy. You can. You can pay-in-four. You can do that. You can have the nice things that you want.” DoorDash is the most shallow, simple version of that, which is just nobody wants to cook and they see cooking as hard and they see cooking as painful and they can get five-star food, but it is really social media’s fault, I think.

Liz Hoffman:
Okay. Well, I think we will leave it there.

Rohan Goswami:
Indeed.
Well, that’s our show for this week. Compound Interest is produced by the wonderful Josh Billinson.

Liz Hoffman:
A special thanks to Anna Pizzino, Katherine Bilgore, Claire Einstein, Rachel Oppenheim, Tori Kuhr, Vilanna Wang, Garett Wiley, Stephanie Chang, and Daniel Hoeft. Our engineer is Bob Mallory, and our theme music is by Steve Bone.

Rohan Goswami:
If you like Compound Interest, please feel free to review us wherever you get with your podcasts.

Liz Hoffman:
And if you want more, you can always sign up for our email newsletters at semafor.com.