Guess Who?

Back in the late 1900s, kids played a board game called "Guess Who?" Two players draw cards with headshots on them. They don't show each other the cards. The players take turns asking "yes" or "no" questions to eliminate possible matches from the board in front of them. The player that eliminates false matches and guesses who’s on their opponent’s card wins.

The modern high-stakes version of this is something like "is this thing I'm looking at real or AI?"

This matters for things like journalism where you can now just ask AI to make up a story and a live breaking report that's convincing. It also matters for things like buying socks. Grifters have a lot more tools to fake documents or steal credentials like payment cards than they used to. Cyber fraud increased 33% in 2024. Figuring out whether a customer or merchant is real or not got much harder.

There's a balance though. Make it too hard to check out and customers go somewhere else. Make it too easy and you make your business a soft target for fraud.

Existential thought

Lamplighter's talked about Adyen's efforts to generate more revenue for their customers as a payments service. Two of the ways it does this are by better fraud detection and higher authorization rates for legitimate transactions. When questions come up, the state-of-the-art is two factor ID. Until now this meant providing two of:

1.      Something you know — like a password,

2.     Something you have — like your phone or a bank statement, and

3.     Something you are — like a fingerprint or face that gets scanned.

AI makes these easier for fraudsters to fake.

At its investor day earlier in November, Adyen pitched its "dynamic identification" capabilities as a foundational piece of the value it delivers to customers and support its future growth. Just as AI has made things easier to do a fraud, it’s given Adyen a tool to use against it. It uses behavior rather than conventional ID to tease out shady activity on its platform. Dynamic ID isn't a separate service. It's downstream of Adyen processing trillions of dollars of transactions and all the clues that come along with that from its core business.

There are easier things to tag. Say you buy a coffee every morning and lunch in midtown Manhattan, maybe you didn't also buy a Jaguar at a Johannesburg used car lot that afternoon. There are tougher problems too. If you're a brand new business and your first customers are customers active at other Adyen merchants, you're probably real. If Adyen's never seen any of the cards or accounts your customers use, you might have to answer a few more questions. Or if you're a new business that also makes purchases using cards and accounts that go through Adyen, you're probably good.

IDing activity this way takes a lot less time than calling up every new merchant and asking for bank statements and utility bills. It also opens up Adyen to serve lots of new businesses faster. For AI to fulfill even a small slice of its potential, it has to expand markets not just cut costs.

One card left

Adyen billed its dynamic identification capabilities as the headline during its investor day. Its co-CEO, Ingo, presented the section. The crowd yawned. Only one analyst asked about it. The audience focused instead on things like how recent customers are growing and what regions have seen success — important things, but ones that Adyen and Lamplighter have covered before. Adyen's dynamic identification abilities offer the company a chance not just to improve fraud detection and authorization rates, but to expand the market it serves. You might expect that to inspire some investor imagination.

Whether dynamic identification delivers for Adyen is to-be-seen. But the idea that identification of what's real or not based on time and behavior in a world where AI gets less and less sloppy at faking it every day is a compelling one. Guess who won't be sleeping on that?

Disclaimer: None of this is investment advice. It's meant to illustrate ways LCM thinks about investing. Things that LCM decides are good investments for LCM and its clients are based on many criteria, not all of which are covered here. Some or all of LCM's ideas may not be suitable for other investors. LCM does not recommend investing either long or short any position mentioned. LCM may own positions in some of the companies mentioned. Some of its ideas will lose money — investing entails risk. See full disclaimer here. 

Previous
Previous

Stone Soup

Next
Next

One Last Job