Is AI taking over Fintech jobs? Inside the Klarna story

Last week, I was in IIT Madras talking about all the fun stuff about building Fintech products. And one of the smart kids there asked me, How do I feel about AI taking over the role of product managers in Fintech? And If I personally used any AI in my day to day work.

And you know what, AI has been taking over jobs in the financial services place the last few years. While I don't mean to alarm anyone or stuff, I thought it would be very interesting to see how it is that AI is being used in one of the biggest Fintech unicorns in the world. Klarna!

Because not only has Klarna been at the forefront of building exciting new fintech products, they  have also been on the fast track of adopting new technologies and implement it into their products and teams. 

But before jump in a little bit about Klarna and what they do?

Klarna : The origin story of Buy now pay later

Klarna, the Swedish fintech sensation that's shaking up the online financial services game. They handle payment processing for e-commerce businesses like pros, making sure store claims and customer payments are all sorted. Back in 2005 – Sebastian, Niklas, and Victor – decided to start up Klarna after soaring through at an entrepreneurship competition. With the help of angel investor Jane Walerud, they got hooked up with a killer team of programmers and the rest is history. 

In 2007, Investment AB Öresund decided to invest in Klarna. Fast forward three years, Klarna had spread its wings to Norway, Finland, Denmark, Germany, and the Netherlands, and received a hefty investment from Sequoia Capital. Their revenue skyrocketed by 80% to a cool $54 million, earning them a spot as one of Europe's most promising tech companies according to The Telegraph in 2011. By 2011, General Atlantic led a funding round pumping $155 million into Klarna, with DST Global also throwing in some cash. Anton Levy from General Atlantic even hopped on board Klarna's ship. In a strategic move, Klarna acquired Analyzd, an Israeli company, for their top-notch risk management and fraud prevention services. In 2013, Klarna capped off their journey so far by joining forces with German powerhouse SOFORT AG, forming the mighty Klarna Group. 

In September 2015, Klarna set foot in the United States with grand ambitions, making it their top priority for expansion. They locked down a special deal with the posh department store Macy's, cementing their position in the market. Sweden's Minister of Enterprise and Innovation, Mikael Damberg, even dubbed them one of the country's "five unicorns", alongside Spotify, Mojang, Skype, and King - all successful startups that made it big on the international scene. Fast forward to 2019, and Klarna hit the jackpot with a massive $460 million funding round to conquer the US. Heavyweights like Dragoneer Investment Group, Commonwealth Bank of Australia, and HMI Capital hopped on the bandwagon, pushing Klarna's valuation up to a jaw-dropping $5.5 billion. They've truly become the créme de la créme of European fintech startups. And just when you thought they couldn't get any more boujee, in 2020 Klarna swooped in and acquired Nuji, a trendy platform for all things fashion and lifestyle. 

But wait, there's more! In 2020, the payment legend Ant Financial (linked with Chinese e-commerce giant Alibaba) decided to throw some funds Klarna's way as part of a strategic alliance between the two industry heavyweights. Fast forward to June 2021, and Klarna made headlines once again by bagging a jaw-dropping $639 million in a fundraising bonanza led by SoftBank Group's Vision Fund 2. This golden ticket catapulted the company's valuation to a staggering $45.6 billion. You can almost hear the cash registers ringing!

In November 2021, Klarna introduced a snazzy new physical card that allowed users to split their payments into interest-free chunks. Fast forward to January 2022, and voila! Klarna brought the party to the UK with their sought-after card, boasting a wait list of 400,000 eager beavers. Come March 2022, Klarna made moves by acquiring PriceRunner, a company known for helping folks compare prices of products. Then in May 2022, Klarna made some tough decisions by giving the pink slip to about 10% of their team, in a rather very tone deaf manner. In June 2022, Klarna bounced back with a bang, announcing a sweet partnership with card powerhouse Marqeta to launch physical Visa cards in the US. Exciting stuff, right?

Then in July 2022, Klarna hit the jackpot with a whopping $800 million in funding, all while being valued at a cool $6.7 billion. Unfortunately, the thrill didn't last long as their valuation took a nosedive by 85% in just a year. 

They also waved goodbye to $580 million in losses between January and July 2022. And to top it all off, in September of the same year, Klarna decided it was time to shed some more weight by announcing plans to lay off an extra 100 employees, all thanks to their drop in valuation and those pesky losses. Tough times in the fashionably flexible payment world, huh?

Come October 2022, the company came out with a fresh new tool called "Klarna Creator" for all the cool kids in the retail and influencer world to team up on some killer brand campaigns and keep a close eye on their cash flow.

What is BNPL?

So what was the problem that Klarna was trying to solve for really ?

They were the pioneer in what is now known as "buy now pay later". So here is the crux of this fintech product, in a form I know you all love. The 5 levels of difficulties. 

1. Child Level: Imagine you’re in a candy store and you see a big chocolate bar that you really want, but you only have enough money for a small one. BNPL is like a friendly shopkeeper who lets you take the big chocolate bar today, and you can pay him back when you get your allowance next week. 
2. Teenager Level:  You know how in video games, you can sometimes get an item or power-up now and pay for it with game points later? BNPL works the same way, but in real life. You get the item you want now, and pay for it in installments later. 
3. College Student Level:  Think of BNPL as a short-term loan. It’s like when you split the bill for a pizza with friends, and one person pays upfront, but everyone else pays their share back over time. Except in this case, the company selling the product is the one fronting the cost, and you pay them back over time. 
4. University Grad Level:  BNPL is a financial service that allows consumers to purchase goods and services immediately but defer the payment to a later date. It’s similar to credit cards, but typically involves 0% interest and the cost is broken down into specific installments. It’s a popular e-commerce tool that can help boost sales by reducing the immediate financial burden on customers. 
5. Expert Level:  BNPL is a burgeoning sector in the fintech space, offering point-of-sale lending options for consumers. It leverages advanced risk assessment algorithms to offer instant, short-term credit for purchases, typically splitting the cost into equal installments over a set period. While it shares similarities with traditional credit models, BNPL services often differentiate themselves through 0% interest rates, no/low late fees, and seamless integration into online checkout processes. However, like all financial tools, it requires responsible use to avoid potential pitfalls such as debt accumulation.

How does Klarna make money: The BNPL Business model

In order to make the bucks at the bank, Klarna heavily depends on hitting up merchants for that instead of swindling it out of customers. In fact in an interview, their CEO even mentioned that they don't beleive credit cards are good businesses because "They are criminally incentivized with bad consumer behavior". Essentially, credit card companies make money when you miss a payment.  Time to dive into the juicy details of how Klarna rakes in the dough.

Payment fees

Klarna makes a pretty penny mostly by cashing in on merchant transactions and slapping on those sneaky variable percentage fees. These bad boys can change depending on how someone is paying and where they're from. For instance, over in the good ol' US of A, merchants have to cough up 30 cents per transaction on top of a variable fee that lands somewhere between 3.29-5.99%. Ka-ching!

Payment fees are also generated when customers:

• Feeling lazy? Just a few clicks and you're out of there. Klarna's Instant Shopping feature will set you back $30 a month plus a $0.30 transaction fee for merchants. Watch out for the 3.29% fee for onsite transactions and a 3.79% fee for offsite transactions.
• Need to split your payment into four parts? Klarna has got you covered with the same $0.30 transaction fee and fees reaching as high as 5.99%.
• Prefer to pay monthly? Klarna will charge you a $0.30 fixed transaction fee and 3.29% in variable fees. Oh, and don't forget the interest rates up to 29.99% throughout the loan term.
• Missed the due date? That's gonna cost you. Klarna will slap you with a fee of up to $8 for late or missed payments that are more than 10 days overdue.
• Don't worry about small purchases though, ones below $25 won't incur a late fee. But anything above that and you could be hit with a maximum late fee of $24. 

Interchange fees

Klarna also came out with this cool new bank account thing where you can get a free debit card and everything, thanks to theri partnership with Visa. Basically, when you buy something, the store pays a small fee to the bank that issued your card. Klarna gets a cut of that fee for helping out Visa and hooking you up with their services.

Interest on loans to merchants

Basis the transactional data of  merchants that Klarna has accumulated, they are well positioned to understand the cashflows of these merchants. And basis that they offer up working capital loans to these merchants, in the process earning some good revenue. 

How Klarna uses AI in Fintech

It's March 2023. Your favourite niece's birthday ( assuming you are as old as I am) is coming up, but you are absolutely at a loss for what to buy for a fantasy loving toddler. Are unicorns the new hot toy? Or would she appreciate a fairy more? 

To do away with such conflicts, Klarna, in collaboration with OpenAI , launched a highly personalized and intelligent shopping experience by providing curated product recommendations to users who ask the platform for shopping advice and inspiration, along with links to shop those products via Klarna’s search and compare tool.  

And Klarna is also preparing to take to the high street of fashion with their new AI leverage! Now you can be a savvy shopper by scanning a barcode in-store and instantly getting all the details on a product (I know I used to do this using the google scanner). Check out reviews, see if there are more colors or versions online, or scope out a better deal. And I don't need to be the one going through the different links anymore, Because their AI shopping assistant does all that for me in an instant. So fancy. I know now that malls are gonna become like my personal trial room with something like this in my pocket.

But how do I access this super shopping assistant?

Customers can download the Klarna plugin from the ChatGPT store. With just a few voice commands or text inputs, consumers receive personalized shopping suggestions tailored to their preferences. Consumers have the choice to offer supplementary cues or demand additional product suggestions tailored to their preferences. With a single click, customers can conveniently access product pages on Klarna's compared and contrast tool to evaluate pricing across various manufacturers. 

And just like that, Klarna became the first AI powered Payments company in the world, to have their offerings on OpenAI's GPT store.

AI replaces 700 ex employees

But that isn't all AI can do for Fintech. And Klarna was once at the forefornt on demonstrating that. 

In fact, Klarna CEO even told Sam Altman, CEO of OpenAI, the makers of ChatGPT "I want Klarna to be your favorite guinea pig". And the one thing that Sebastian Siemiatkowski, CEO of Klarna actively montiors is how many of his employees are using AI on a day to day basis. (Currently its at about 50%)

BTW. ChatGPT isn't the only AI integration they have. Klarna has also been working on their own internal AI model, especially for video shoppers. 

And just last week, CEO Siemiatkowski, was heard on a news channel saying that "AI does the work of 700 people at Klarna". 

FYI: Klarna did layoff close to 700 people in 2022, right before their OpenAI partnership.

Klarna's fancy  chatbot, powered by OpenAI, is basically running the show when it comes to customer service. It's handling a whopping two-thirds of all customer service chats - that's 2.3 million chats in just four weeks, in 35 different languages. That's like having 700 human agents working round the clock, except this bot never needs a coffee break. And get this - customers are actually loving the bot just as much as they love talking to a real person. It's more accurate, cuts down on repeat questions, and can solve problems in a flash - all in less than 2 minutes. In comparison, it used to take humans a not-so-efficient 11 minutes to do the same thing. Klarna is even predicting some serious cash flow, estimating a whopping $40 million in operational expenses being cut down by 2024 thanks to it. 

Will thsi lead to further layoffs at Klarna ?

Well, the customer support function was any way outsourced to a 3rd party agency, but yes, as a collective we will see demand for such jobs going down with time, as AI models become more and more human like.

BTW. Klarna isn't the only company int he news for replacing human workers with AI. DuoLingo laid off 10% of their translator staffs, in favour of AI recently.

Even Klarna has taken a decision to not hire more, in most divsions except engineering, because AI is doing a good job already. 

Obviously such news helps feed into the word on the street. Which btw has been ripe with the gossip of Klarna eyeing the US markets for an IPO listing. And after that 85% slash of their valuation they need a lot more good news for that to come true. 

AI is revolutionizing Fintech

Enough about Klarna. What about other Fintech? Is AI being actively used to do what previously humans would? 

Automation of routine jobs

One of the things Fintech need to do is actively monitor several of their metrics, whether its success rates, costs, acquisition funnels. And take action in real time.

I remember when I was leading the Analytics tea at PayU, we had built a bunch of such dashbaords and triggers, and built an internal tool, which users could use to pull out a combination of any metrics and attributes of their choice, to do a deep dive. 

AI is now doing all of that (Thank god we proved our salt before the fact). 

Decision making using internal and external data point

Imagine you are a trading platform. And on Twitter someone just posted ghastly pictures of an oil nation being bombed (please don't let that happen in reality God). 

But you only see that tweet a day later by when the oil prices have already shot through the roofs.

AI, with their round the clock availability, and ability to parse throguh social media posts, would have noticed this in real time, had the ability to cross check with news channels to check for the veracity of the same, and take quick decisions to hedge against the potential rising price of crude. 

Fraud Detection Using AI

Frauds are a huge expense for Fintech.

On one hand if their algorithms are too strict, then they'd even deny legitimate customers entry because that would be judged as a fraudulent transaction. (Happened to us once when we were testing out AWS's security measures, when it kept detecting one of us a bot, and refused us entry to our app). 

But if its too flexbile then the fraudsters get away with cheating you out of your money. or worse your users money. And you lose trust forever. 

Now AI has been doing that job, and doing it pretty well. In fact one csae study was from the banking giant HSBC who had partnered with Google to use their AI engines to ward off money launderers and fraudsters. 

How I use AI in my day job?

And finally, lets come to the question of how I use AI in my life. 

As a content creator, most of my creatives are AI generated. And some of the charts that you see, yep, I used Gemini to build them with the raw data I fed it. 

I also use it to rephrase my sometime very lengthy posts and articles, and hopefully make it a more interesting read, without compromising on the depth of material shared here. 

And because I am a creatively challenged person, all of the art work, is done by AI for me. Including that not so fancy banner for this article. 

In my day job when building fintech products, the one thing I have to constantly stay abreast of is what the regulators are saying. And those circulars from time to time cane be a big read (yes I find them long that is saying much isn't it ?). So one thing I love doing is using AI to share the crux of every heading and sub heading, with respect to the context that I am building. And then For each section where I see a potential impact, I drill down asking it more questions, and personally reading up those section to ensure it didn't hallucinate the answers. 

Another area of my job that I love doing is brainstorming. And of late I have been turning to a few (paid) AI apps to help me refine that process. One of them is Boardmix, which I use to sketch out mindmaps and wireframes. the other is when I am writing my PRDs and need a slightly more technical explanation, where I use piktoChart to generate flowcharts and UML diagrams. now I'll admit the first draft of it isn't directly useable, but they do a great job of saving at least a couple hours, and then I can work on just refining it and adding other edge cases (which I use Gemini for to critique my model and find loop holes) and voila. I have my PRD ready in a matter of a couple of hours. 

And because I also love handling data, Gemeni pro helps me debugging my code when I get stuck as well, especially in those times, when even stackoverflow couldn't help. 

My next step? Helping my data crunch his experiments data better and faster with AI. I'll keep you posted on how that goes over the months !

That's it for now and I'll see you again next week, when we discuss the technical and fundamental value of Bitcoin ahead of the halving, and its impact on other cryptocurrencies. BTW, you may wanna catch up with the earlier version of bitcoin technicalities before that