KENNY WOLF: Hello, everyone. Thanks for joining us this afternoon. I’m Kenny Wolf, and I lead core payments here at Stripe. We shared a lot of product updates at the keynote this morning and in this session we’re going to dive deeper into what’s coming across Payments and Radar. A lot is changing across the economy right now. Companies are launching and monetizing faster than ever. We see 20% of Stripe Atlas startups launch their first paying customer within less than 30 days of incorporation. That’s about twice the pace we saw in 2020. And these companies are global by default. We see AI companies selling into twice as many companies in their first year relative to SaaS companies four years ago.

And how customers find you is changing too. AI is guiding more customers to your websites. And last holiday season, we saw the traffic driven by AI referrals up almost 700% year over year. So that means more markets, more channels, and more opportunities. But every piece of this expansion is increasing your surface area for fraud. We’ve seen new vectors arising for fraud too, such as free trial abuse, which is up more than twofold in the past six months. Through our daily conversations with users, we’ve emerged three themes on how Stripe can help you. First, you want to sell everywhere, in every market and every channel without constraints. Second, you want to maximize revenue at every single one of those touch points, from the checkout that your customer sees to the invisible authorization logic that is working behind the scenes. And third, you want to prevent fraud across the entire customer lifecycle, and you want us to adapt as fast as the threats do.

In this session, we’re going to dive deeper into what we’re building across these three areas. And so we’re going to start with selling everywhere, and I’d like to welcome Susan Lee to the stage. Susan, take it away.

SUSAN LEE: Thanks, Kenny. Expanding internationally used to be a major milestone for a business. You’d start in your home market, you’d prove the model, and maybe you’d carefully consider one to two countries to expand in every year. Today, that playbook is pretty much obsolete. For most AI, SaaS, and digital businesses, the internet is the default market. And the moment a product launches, it needs to be accessible everywhere. Take Gamma, an AI platform for creating presentations based here in California. Within days of going live, they noticed they had thousands of customers in India, Brazil, Kazakhstan, and more. And in fact, more than 85% of their customers were coming from outside the US. But even with global demand, buying preferences still vary dramatically by country. And when Gamma launched, you could only pay by credit card and in USD. That meant they were leaving a lot of money on the table.

So we built the Optimized Checkout Suite to solve exactly this problem, helping you create local buying experiences wherever your customers are to grow your global revenue. And this year, we’re increasing the number of payment methods you can access through the Optimized Checkout Suite from 125 to more than 140, including Scalapay in Italy, MoMo in Vietnam, and GCash in the Philippines. But those numbers aren’t enough if every payment method comes with caveats. And so we’re rebuilding our platform to make every payment method work across borders, across transaction types, and across currencies. So one of the first payment methods that we’re rolling out this way is UPI, which is the dominant payment method in India with over 500 million users. Now until recently, across all major PSPs, and Stripe is included in this set, you needed a local entity to offer it. Today, you can turn on UPI from right here in the US with a click in the Dashboard.

And when Gamma did this, they saw their UPI transactions surpass card transactions that very same day. And within a month, they saw their revenue in India grow 22%. And that’s because local payment methods drive conversion, and so does local pricing. So that’s where Adaptive Pricing comes in. It localizes prices in more than 150 countries and drives more than a 17% uplift in average cross-border revenue. And we’re rolling out several new improvements. So first, Adaptive Pricing will now support recurring transactions so that subscription businesses can display prices in local currencies, which increases conversion by an average of 4.7%. Companies like Figma and Cursor are already using it to expand in their global markets. Second, Adaptive Pricing is now powered by a new AI model that better predicts the right currency for each customer. So previously, we used a buyer’s IP address to detect the right currency, but that doesn’t really catch all the edge cases.

For example, think about an American buyer on vacation in Mexico. They’d still see prices in pesos even though they want to pay in dollars. And that extra cognitive load can lose a buyer entirely. Now, our AI model analyzes dozens of session signals, things like device type and the past currency you’ve used to dynamically predict the right currency.

And third, we’re bringing market-specific price recommendations to Adaptive Pricing so that you can not only convert the right prices, but set the best price to begin with. These will take into account cost of living, local competition, even willingness to pay, and they’re all powered by the data that we see across the Stripe network. They’ll even take into account cultural preferences like rounding to familiar numbers. So I want to do a quick audience poll. Can anyone tell me, and feel free to shout it out, what are the most common last two digits for a US price? Yes, that is right. It’s 99 cents. So some of you may have watched a lot of late night info commercials like I did back in the day to know 99 cents means you’re getting a deal. But in places like Japan, 99-cent pricing can actually feel really gimmicky, and Japanese buyers strongly prefer round numbers. So Adaptive Pricing will bake into account these local preferences into its recommendations, so you don’t have to know every market’s quirks.

Now, localizing payments is just the beginning. Operating a global payments business brings a whole new set of challenges. Things like keeping up with standard VAT rate changes to be compliant in every country, or maybe you’re finding yourself using Google Translate to troubleshoot support for a Polish customer at 4:00 a.m. So if you don’t want to have to do all that, that’s where Managed Payments comes in. It’s our merchant of record solution that handles local checkout, tax compliance, fraud, disputes, and customer support. And so what used to take months to launch in each market can now be done in minutes. And today, managed payments is now generally available for digital businesses.

And one of the most powerful things about Managed Payments is its flexibility. So where you want full control, like in your core markets where you may have already invested in your tax setup, you can keep doing so. And when you want to go live quicker or test new products, you can use Managed Payments at the country or transaction level. And soon, Managed Payments will automatically route transactions through Stripe’s local entities in several markets. So that means you get the benefits of higher auth rates and more relevant payment methods in addition to all the benefits of a merchant of record.

And then finally, selling everywhere, it doesn’t just mean new markets. It also means entirely new channels like agents. So we mentioned in the keynote this morning that we are building the infrastructure for the agentic economy with our Agentic Commerce Suite, which now supports new agents such as Gemini and Meta for in-ad shopping experiences. So we’re going to go much deeper on that in our agentic commerce session later today, so encourage you all to join to learn more. But for now, I’m going to hand it back over to Kenny, who’s going to share more about what we’re building to help you capture more value from every transaction.

KENNY WOLF: All right. Thanks, Susan. We work hand in hand with our users every day working on ways to help them grow their businesses, whether that’s testing new products, debugging errors, or solving their thorniest payments or business challenges. And we think that guidance should be available to you from the moment you start building your checkout. And that’s why we created a Checkout studio. It’s our new home in the Dashboard for building a highly performant checkout. It’s a single place to configure, analyze, and optimize your experience. No more need to iterate through multiple code changes and test environments. You can do it all in Checkout studio. Its embedded AI assistant helps you preview your integrations and find the best configuration for your brand, and then you can easily pass off the implementation to your favorite coding agent. After that, you can monitor your performance and slice the detailed data that’s available by location, by product, and more.

You can even debug issues at the individual transaction level. So let’s say you have a customer who reached out because they were charged an unexpected tax rate. You can go into Checkout studio and open the session for that individual transaction and verify what the customer’s country was at the time of the payment, therefore ensuring they were charged the correct tax amount. We’re also expanding the UI types that we support within Checkout. This morning, you saw us demo the new embedded form. We’re making Checkout even more flexible with this out-of-the-box UI that works for a variety of use cases beyond just standard payment pages. Think modals, sidebars, chat interfaces, or mobile web. And businesses like Mindvalley will be using it to create a seamless, integrated user experience. Across all of these checkout surfaces, you’ll have more control over the look and feel of your checkout. You can add custom banners or specific fields to ensure that you match your brand.

We’re also investing and giving you more control over your in-person payments with Stripe Terminal. First, we’re growing our hardware portfolio to serve more business use cases. Earlier today, we announced the Stripe Reader T600. It has an eight-inch screen, and that means more real estate for you to upsell and provide loyalty opportunities. And later this year, we’ll be providing weatherized readers that are certified for unattended payments, so you can install these at parking lots, EV charging stations, and more. We’re also making it easier for you to take in-person payments around the globe. By the end of this year, Terminal will be in 40 markets, and that includes 15 new markets, as well as Mexico and Hong Kong.

Lastly, we’re adding support for standalone mode. So when you get your Stripe Terminal device, you can quickly start accepting payments for your in-person payments on these readers without a code or point of sale required. So all of that helps you get more customers to buy, but after that, there are still many reasons that your payment may fail. Maybe the user mistyped their ZIP code or used outdated credit card information. And if a customer is falsely declined when trying to buy from you, 40% of them will never return. And that’s where Authorization Boost comes in. It’s our AI-powered optimization product that weighs a variety of different factors for every single card payment like, how does that issuer want their ISO message formatted? What gateway should we route through? And more. Auth Boost increases auth rates by close to 4% on average while simultaneously lowering your network costs by 3%. All of this with no effort needed by you. Businesses like Notion and Squarespace are using Auth Boost every day.

And we’re continuously improving Auth Boost behind the scenes with ML-driven optimizations. We’re using cutting-edge payment developments like data only globally and smart PINless debit routing, as well as intelligent decline prevention. We do all of this to limit excessive network fees while maximizing your authorization rates. And later this year, we’ll increase your savings even further with L3 data, leveraging Visa’s latest CEDP requirements, and we’re going to provide Smart Retries to optimize your recurring payments experience. 

Finally, we’ll be enhancing the A/B testing capabilities that we announced this morning, giving users more control on how you can analyze the results of your A/B tests. You can slice by market, by card type, where these recurring transactions are affected, and more. Okay, so I’ve shown you how you can maximize every dollar of revenue for your business, but what happens when bad actors get involved? The fraud landscape has fundamentally changed over the last year, and Nicole is going to tell you what we’re doing about it. Nicole, over to you.

NICOLE PAGLIA: Thank you. AI is transforming fraud. The playbook that worked even just last year doesn’t cover what’s happening today. So you of course need protection in the places you’re probably already thinking about, like accepting a payment. But now, with bad actors racking up usage, free trials, and more, you also need it in places you simply haven’t had to worry about until now. So many of you know Radar as our payments fraud tool. Today, it’s much bigger. Radar is now our comprehensive fraud prevention product that helps to protect your entire business beyond just the transaction. It’s powered by the Stripe network, which is made up of more than $1.9 trillion in annual payments volume from across millions of global businesses, which means we can detect fraud that many others simply can’t because of our scale.

So let me walk you through the three main ways we’re improving Radar this year: we’re expanding its coverage. We’re making it more customized to work for your business, and we’re expanding its reach far beyond Stripe. So let’s start with coverage. Radar now protects you at the very beginning of your customer lifecycle. You heard this morning in the keynote, we announced Radar’s new abuse prevention features, which are meant to target some of the fastest-growing fraud types out there, like multi-account abuse. So this is when a new user signs up again and again to exploit those new user perks, and it’s a lot more prevalent than you may think. At AI companies we’re seeing on Stripe, more than 14% of the signups are suspected of this. So now to help, Radar helps protect you at signup so before a payment is ever made, detect those customers, whether they’re a fraudulent account that’s linked to other fraudulent accounts, so that way you can separate your real customers from the bad actors.

And then there’s free trial abuse. So for the AI companies out there like OpenAI or ElevenLabs, this is an existential threat. At scale, you can be talking about millions of dollars in wasted compute costs, all on users who are never going to convert to your business. So now, Radar blocks these fraudulent free trials with 95% precision and all with one click. That’s right. One click. And then there’s pay-as-you-go abuse. So this is when someone racks up usage only for the payments that ultimately fail in the end, leaving you with the loss. So to prevent that, Radar now helps predict the likelihood of that happening. So you can decide if you ultimately want to limit service or request a top-off or take another action based on your risk tolerance. So the bottom line, you shouldn’t have to throttle your growth to keep abuse in check.

And now with Radar, you don’t have to. In the coming months, we’ll continue to roll out these capabilities more broadly, both in our Dashboard as well as via API. And we’re also expanding Radar to cover new fraud types like bot abuse, account sharing, and refund abuse. Okay, so that’s coverage across the full customer journey, but you also need coverage every way that your customers may want to pay. So historically, Radar has only blocked cards. As of today, Radar now blocks all high-risk transactions for all Stripe payment methods. So that includes bank debits, wallets, BNPLs, stablecoins, you name it, Radar has got you covered.

Later this year, we’ll also be expanding to agentic transactions and crypto to help keep you ahead of what’s coming next. And you’ll be able to use custom rules so you can really pinpoint the specific fraud patterns for each payment method or that are unique to your business. 

Okay. So everything I just talked about is about preventing fraud from your end buyers, but for many of you out there who are running a platform or maybe a marketplace, you also need coverage from potentially fraudulent businesses that you work with, not just the end buyers. So these launches are especially for you.

Radar now helps you to spot high-risk businesses at onboarding and gives you tools to prevent account fraud. You can get things like AI fraud scores, manual reviews, or rules writing capabilities. FreshBooks, for example, has blocked over 300 additional fraudulent accounts with Radar in just the first 3 months. But here’s the part I really love. They’re also approving more accounts overall. Because they’re able to pinpoint that fraud more accurately at onboarding, it gives them the confidence to cut down manual reviews and automate onboarding for their good customers. So with Radar, lower fraud and higher growth.

We’re bringing in more tools to manage account risks later this year. So, for example, you’ll be able to identify if a business is at likelihood of going under, or maybe they’re seeing a rapid spike in disputes, and then you can automatically request reserves or pause payouts to prevent that loss. On the flip side, you can also take these fraud signals to drive growth for your platform. So for example, you might be able to identify your best customers and give them access to Instant Payouts to further grow your revenue or differentiate your platform. So whether you’re a single business protecting your checkout or a global platform managing thousands of accounts, Radar’s got you covered.

But even with broad protection, some patterns are just unique to your business, and that’s where customization is key. Now, Stripe can work with your fraud team hand in hand to develop custom models. We marry your proprietary data with Radar’s global intelligence to create a fully bespoke solution just for you. So let’s imagine you’re a retailer—send us your product data, and we’ll help you to catch more disputes. Or if you’re a ticketing platform, you can securely pass your resale data so that we can help you to pinpoint the scalpers. We’ve partnered with businesses on custom models and they’re already seeing a 15% reduction in fraud, and we’re excited to expand availability later this year.

Finally, we’re extending Radar’s reach far beyond Stripe. We believe that this intelligence shouldn’t just be limited to Stripe transactions. You should be able to use it everywhere that your business operates. Today already, users with a multiprocessor setup are able to use Radar signals across their entire fraud stack. Later this year, we’re giving you access to more signals to use across all of your payments or fraud stack. For example, you can get the likelihood that a payment is fraudulent or an early fraud warning, or the likelihood that a business is suspicious based on just their website URL. So to summarize, Radar is going everywhere that fraud is, from signup to payment, to that final invoice, and we’re giving you more tools to customize it to your business or to use it across your entire fraud stack. Fraud will keep evolving; that’s a given. But with Radar, you’ll always be ahead of it.

Let me welcome Kenny and Susan back to the stage, and let’s get into some Q&A.

SUSAN LEE: Alright. Let’s see. So there’s obviously been a lot of excitement around agentic commerce today, but the question is around consumer adoption, which still feels pretty early. What needs to happen for consumers to actually trust and pay through agents? I can open that up to either of you.

NICOLE PAGLIA: Happy to start. I mean, I think the honest answer is we don’t know exactly how and when the consumer behavior is going to shift. That said, I think you probably got the hint from the keynote that we’re very bullish about what’s happening here and want to help keep you ahead of it. And I think no matter when and how it happens, the checkouts that we see today and the manual entering payment information already does and will probably feel very outdated sooner than we all think. So I think at Stripe, how we want to keep you prepared is no matter where we land on the spectrum, whether it’s agents are mostly for discovery and then the human is in the loop for payment or on the other end, fully agent-to-agent transactions, with the Agentic Commerce Suite, you basically don’t need to bet on one of those outcomes, and you have the flexibility as consumer behavior shifts.

SUSAN LEE: Yeah. I think these behavior shifts always feel far away until suddenly they’re inevitable, and that’s what we are betting on with agentic. Okay. Next question is about the, we talked about a lot of additional payment methods that we’re adding to Stripe, so is there any correlation between offering those additional payment methods and an increase in fraud?

KENNY WOLF: I can take that one if you want. So I think, good news is, as we just announced here, you will have Radar coverage, even if it’s a noncard payment method. So there’s a certain level of safety that you’ll get across the board. And it really depends on how you’ve integrated with payment methods. If you have a bunch of card-specific logic in how you’re doing fraud detection, that may need to evolve. But oftentimes we actually see less fraud when you add these noncard payment methods, because if you think about it, a lot of the bank payment methods have stronger authentication, right? So the buyer is actually well validated. So generally, we see if you offer more payment methods, there’s more diversity involved, which makes it harder for the fraudsters, and you have Radar coverage across the board. So generally a good thing.

SUSAN LEE: Okay. Related question. Nicole, you talked about new types of fraud that we’re seeing with AI companies being pretty wild. Is this something that we’re seeing across the board as well, or is this more of an AI company–specific problem?

NICOLE PAGLIA: So we’re seeing a broader rise in fraud. I think I saw a report recently with something like 60% of businesses are seeing a rise in first-party fraud. And some of these fraud types, honestly, they’re not all new. So free trial abuse we’ve talked about. I mean, SaaS companies have had that in some form for years, but I think what’s changed is for AI companies, just one, the scale that they’re operating, and two, how costly it is. So for a SaaS company, the variable cost of an extra free trial is just often not that high. Whereas for an AI company, you’re paying for the compute services for every free trial. So I think the sophistication, the scale, and the cost is fairly unprecedented. So we’re working very closely to develop the solutions for AI companies, knowing that as more industries—SaaS, retail, financial services, gaming, and others—begin to embed more and more AI products or services, even think of many of you might have a chatbot, then you might be susceptible to this sort of fraud as well. So we’re developing closely with AI companies knowing that the solution will be more broadly applicable.

SUSAN LEE: Alright. I’m going to close it out with one more question. I’ll direct it at you, Kenny. So it’s obviously been an unprecedented pace of change that we’re seeing in technology. I believe Patrick called it the first quarter of the singularity. So how does that change the way Stripe approaches its roadmap, and how do we think about staying nimble through those changes?

KENNY WOLF: Yeah, it’s definitely changed a lot of how we work across the board. So at first, it has fundamentally changed how we go through and think about and operate our engineering teams. And that’s something if you think about how it might transform your company, we’re happy to talk about what’s worked for us at Stripe. Second, we have a whole new user, right? We have these agents that operate very different and have different friction points and we’ll find and discover in different ways than our human users have. Third, we’re exposing and using AI within our products. It’s helping for certain Radar scenarios. You saw it with AI assistant for Checkout studio. And then lastly, we now need to facilitate—we’re a developer company historically and APIs are one of our premier offerings. So we are providing a new shape of API to help crypto or an AI or rapidly changing landscape evolve, and these are what we’re calling frontier APIs. And so when you use a frontier API, that’s an indication that this is a quicker changing API. This may be something where we’ll be evolving it rapidly. We may be deprecating it as we learn more. And so you can just be part of that rapid journey with us, with that technology. So hopefully that will be helpful.

SUSAN LEE: Awesome. Well, that is all the time that we have for today. So thanks, Nicole and Kenny. Thanks, all of you, for joining us.