JOHN AFFAKI: Thank you, thank you, thank you, and good morning everybody. I hope you’re having a great Sessions week. It’s an incredible pleasure to host you all. So, in my role at Stripe, I actually have the pleasure of working with an entire team that builds the roadmap for all of our payments products. In order to build that roadmap, we have to think about the trends of what’s going to shape the industry over the next five to ten years. For example, how’s AI going to help us build products, both as developers but also the capabilities that we’re able to provide? How is it going to actually help fraudsters try to attack all of us in this industry? And how are we going to use AI to fight that? How are new payment methods, like stablecoins or RTPs, going to shrink borders and enable you to grow businesses even faster as we go through that? So all of those things factor into those roadmaps. But as we look at how we build it, the biggest factor that shapes it is your feedback. All of you.

So, to level set, we know that managing payments is really hard. We can see it, but you are living it. So that feedback is very important, whether it’s an online comment, a form, meetings that we’re having, or if you flag us down on the street to basically yell, “Hey, John, we need this.” It’s very important for us to do that because this is what helps us build the products that we bring to you. So, as we’re going to go through today, we’re going to talk about the top three things that we hear from you the most often. First, better payment performance. You actually want to maximize your auth rates and you want to minimize fraud and costs. All of you are asking, how can AI help? Two, you want more flexibility. You want Stripe to work with your existing payment setup, third-party tools, and even other processors that you might have already integrated with or want to integrate with. Third, you want ease in expanding globally. You want to be able to start as a global-enabled company so that you can expand to new markets faster, create localized buyer experiences, and serve the world in one shot.

So you heard us talk about a lot of our new payment launches during the keynote, but this session is really aimed at going a layer deeper and to give you a little bit more detail on things that we haven’t talked about and things that are launching later in the year. Now we’re going to talk about other things like Revenue, Connect, and Money Management, but that’s going to happen in other keynotes later today if you go across the street. So look forward to those, also. All right, now let me bring up Nicole, who’s going to talk about our first theme, Payments Performance. Welcome Nicole.

NICOLE PAGLIA: Roughly 1 out of every 10 payments fails. That’s a terrible success rate. Imagine if your car only started 9 out of 10 times. You’d probably get it fixed. While many of those payments are fraudulent, so you’re happy to block them, a lot are real sales from customers that should never have been declined. You all know these false declines are particularly painful. You don’t just lose out on that one-time sale, but it’s all of your future revenue. Plus you’re eroding your customer trust. So many of you are asking, is this something AI can help with? Let’s actually see a show of hands. How many of all of you plan to add AI to your payment setup, let’s say within the next year?

All right. I have to say I expected a little more in San Francisco. We’re actually seeing over 42% of payment leaders are already using AI in their payment setup. Another third plan to add it within the next two years. So definitely something you should be paying attention to. That’s why we built the Payment Intelligence Suite to help. It uses AI to make hundreds of real-time decisions to automatically optimize across auth, fraud, and cost simultaneously, all with no work from you. So today I’m going to highlight some of the products where we’re already putting AI to work for you and preview some of our upcoming brand-new ships.

So let’s start with authorization. Earlier this week, you may have heard in the keynote, we announced Authorization Boost. As the name suggests, it boosts your auth rates by 2.2% on average. So how does it work? It uses AI to automatically make sure that your payment information is always up-to-date with no work from your customers, using things like network tokens or real-time card account updater. But it also has Stripe proprietary optimizations that fine-tune the ISO messages, figure out how to route a transaction, and decide how, when, or if we should retry a transaction—all in real time. These proprietary optimizations are always getting better. In fact, we recently deployed a deep neural net model that helped to figure out when we should optimize a failed transaction. The results? We recovered a record number of sales for all of you while also lowering network costs by 22%, on average. So higher revenue and lower costs.

So, next up, you can expect to see a lot more cost optimizations like this. We’re also thinking about how you can use 3DS more dynamically to recover more sales. Users like Clear, GitHub, and Slack already rely on Auth Boost every day, and we’d love for you all to try it, too. We’re going to automatically be rolling out later this year, but you can actually go online and request access today. Okay, so that’s Auth Boost, and it’s getting more legitimate transactions to succeed. 

But, of course, there are some that you want to block because they’re fraudulent, and that’s where Radar comes in. It helps to stop those fraudulent charges before they hit your business. But fraud detection is becoming a harder and harder problem to solve. You’ve got advancements in AI, and you’ve got bad actors with increasingly creative schemes. Radar takes care of that for you. It reduces fraud by 38%, on average, and it automatically adapts to those new fraud trends, again, all in real time with no work needed from you. So over the last year, we’ve actually seen a 40% increase in noncard payment methods on Stripe. But more payment methods means more fraud vectors that you all have to think about and manage. So to help, we’ve built new AI models specialized for ACH and SEPA fraud. We’re going to be expanding this to more payment methods later this year for all of those other payment methods your customers know and love, like PayPal or Klarna, and more bank debits.

Okay, but what about the fraud charges that actually do get through? All of you say that disputes are a huge hassle, so we’re investing to make them less of a hassle and less costly than they are today. First up—this is partly coming from users, like John said, an online tweet, like Levels.io here—we’re introducing a new feature called Smart Refunds. Smart Refunds uses AI to proactively recommend payments to refund based on the signals that Radar gets after the payment’s submitted, but before the dispute’s actually filed. That’s right. You actually get Radar’s AI not once, but twice with this. The early results have been very encouraging. Honestly, a lot better than even we expected. We’re catching 26% more fraud in our pilot with just this extra check. So we’re gonna be adding this to Radar automatically later this year.

But you also tell us it’s not just the fraudulent disputes that cause the headache. It’s so-called friendly fraud, although not sure how friendly it is. This is when customers actually dispute legitimate charges. So maybe they forgot their purchase or their kid used their card. Whatever the reason is, it’s real lost revenue for all of you. So to help, we’re rolling out support for Visa’s Verified products and MasterCard’s Ethoca products to help prevent friendly fraud and keep you out of those chargeback monitoring programs. So these tools do things like remind your customers of the order details. We’ve all been there. You log on, you see your statement, but maybe don’t recognize the charge. Well, no more. They also do more advanced things, like you can set up rules to automatically refund transactions before the dispute’s actually filed. Finally, when the dispute actually does occur, we want to help you reduce cost and save time there, too.

You all tell us this is actually one of your least favorite parts of managing payments, and it’s no wonder why. Chargebacks cost businesses $50 billion every year, and the dispute rates keep going up and up year after year. So, again, many of you are asking, how can I spend less time and money on disputes? Earlier in the keynote, we announced Smart Disputes. So rather than spending hours figuring out which dispute to fight, what’s the evidence, how do I submit it all, it does all that heavy lifting for you. So it’s an AI-powered tool that will automatically compile and submit a tailored evidence packet all on your behalf. Again, best of all, we’re putting skin in the game. If you counter a dispute and lose, we’re going to cover the cost to fight it. Some of our early adopters, like Vimeo and Squarespace, are already recovering 13% more chargebacks in our pilot. We’re looking forward to rolling it out to all of you and getting it into your hands later this year. We’re also going to be adding integrations with third-party tools so you can give us information like a tracking ID to help improve your submission even more.

So taking a step back across all of these products, from authorization all the way through disputes, what really underpins the AI is the data from the Stripe network. We process over $1.4 trillion annually across millions of global businesses, touching almost every geo and industry. We want to put that scale to work for all of you. So some of the things that we’re working on to improve AI and we’re really excited about—we’ve got multivariate optimizations, new models like our Payments Foundation Model that you heard in the keynote, and then incorporating more non-Stripe data into our models to improve the performance. With that, next, I want to welcome Roshan, who’s going to share more about how Stripe is becoming flexible to meet your needs.

ROSHAN SADANANI: All right. Hello everyone. Thank you, Nicole. Now, you’ve told us that you want more flexibility in how you use Stripe. You want it to be easier to use the parts of Stripe that you want the most with your existing tech stack in ways that kind of mesh together. For example, many of you don’t only process payments online, especially given that over 78% of payments still happen in person. Your customers expect seamless experiences, though, online and in person, across those channels when they shop. Stripe Terminal is our in-person payments product, and it helps you deliver those frictionless experiences. Terminal is an end-to-end solution. It includes hardware, fleet management, tap-to-pay, global coverage, and much, much more.

Working with users like Hertz and Shopify, who’ve deployed Terminal at scale, we’ve learned that many of you want to pair those Terminal capabilities with your existing providers and that existing tech stack that I mentioned. For example, you might want to use Stripe hardware to try out a new use case, like on-site subscription sign-ups, but you need it to work with your existing point-of-sale and subscription management systems. Or maybe you acquire a new brand and they’re using a different in-person payments processor. You need to unify your online and in-person data through a common operational dashboard. Whatever the reason, we’ve made Terminal more flexible. You can pick and choose the components you need to best grow your revenue. We now have direct support for third-party hardware, starting with Verifone. You can use Verifone hardware with Terminal software, including automating ordering and fulfillment processes, building custom on-reader experiences, and using Stripe’s fleet management software stack for both Stripe and Verifone devices.

This complements the integrations that we’ve already built. We offer Stripe connectors directly to point-of-sale providers like Oracle, Merco, Retail Realm, and many more. Through our partnership with FreedomPay, a leading independent payments gateway, you can also easily pair Stripe with over 1,000 point-of-sale systems and multiple hardware vendors, too. Now, in-person payments isn’t the only place we’ve made all of this possible. Even for payments processed outside of Stripe, you can personalize your checkout with the Optimized Checkout Suite, you can manage subscriptions with Stripe Billing, and you can help prevent fraud with Stripe Radar. You’ll be able to use Stripe’s authentication optimizations across all payments later this year, too. We’re excited about that.

There’s a piece of this puzzle, though, that we haven’t talked about yet. So Terminal, the products you see on the screen. We know that there are lots of reasons why you might want to use multiple processors, multiple payment service providers, whether it’s for redundancy, to support local acquiring, because of a new acquisition, and the list really goes on. Working and reconciling payments across multiple processors is hard. You don’t have to take my word for it. Wix, Wayfair, and others have told us that they need tools that can help them work across multiple PSPs and seamlessly interoperate across the rest of their tech stack. For those that undertake building their own in-house orchestration, it can be particularly painful. So this is a real quote about how optimization and intelligent routing is often such an important part of those solutions, but can often be one of the hardest things to build yourself.

So this week we announced Stripe Orchestration to solve all of these problems for you. As you heard in the keynote, it’s a new product to help manage multiple providers and optimize performance all through Stripe. With Orchestration, you can use Stripe to route payments to your processor of choice, get full visibility into performance across PSPs, streamline refunds and operational processes through a single Stripe Dashboard, and simplify the operational burden of managing multiple processors. For the developers in the room, you’ll be excited to know that this includes just a one-line change to your Stripe integration to go live. We are just getting started with Orchestration. Soon, we’ll add support for A/B testing across PSPs, bringing your own processor integrations to Stripe, managing disputes and other post-transaction flows across processors, and much, much more.

We’re really excited to be building Orchestration as part of the next iteration of Stripe’s financial infrastructure and for what this can do to power all of your businesses. We’re working with companies like FlyCode, URBN, and Miro as design partners for this work. You can request access to Orchestration today, and I’m really excited to learn what you’d like to build. Okay, so to shift gears now: it’s not just flexibility for online, and offline, and processors, but we often hear from you that you want help unifying operations across your entire business, even if that’s already running on Stripe. With Organizations, you can now bring all of your Stripe accounts into one centralized view. Intelsat and over 20,000 other businesses have already created Organizations on Stripe.

Today, we’re adding something many of you have long asked for: the ability to share your customers and payment methods across all of your Stripe accounts. Best of all, there are no lengthy migrations here. It’s just one click in the dashboard and the sharing happens automatically without making any changes to your integration. With customer and payment method sharing, you can seamlessly use your payment methods that you’ve already collected across Stripe accounts, and as I mentioned, with no integration changes. All right, so that’s a whirlwind tour through how we’re investing in flexibility. But, of course, that’s only one part of the puzzle. With that, I’d love to welcome Susan up to the stage to chat about global growth.

SUSAN LEE: All right. Let’s do another show of hands, a quick one. How many of you here serve customers in more than one country? Okay, seeing a lot of hands. What about more than 20 countries? Okay. How about 100? All right. Got some of you there. The point is that more and more Stripe businesses are telling us we need to be a global company from day one. We’re especially hearing this in super-fast growing industries like AI, SaaS, and cloud computing. These are obviously products that can easily scale beyond local borders. The problem is that even with global demand, selling your product in a new country is still complex.

Take tax compliance, for example. Every country has their own tax laws and compliance requirements, and they frequently change. So just this year, Slovakia increased their standard VAT rate from 20% to 23%, and the Philippines introduced a 12% VAT on digital services provided by remote sellers. So to stay compliant, you’d have to constantly monitor these regulations for every country you operate in, and then add to that country-specific payment regulations, fraud risks, and more. That obviously takes critical resources away from building your core product. We’ve read dozens of really painful posts, just like these, underscoring the value of using a merchant of record platform to offload the burden of scaling globally. But until now, you couldn’t do this on Stripe. Last year, we acquired Lemon Squeezy to change that with Stripe Managed Payments. We like to think of it as an easy button for launching and scaling a digital goods business anywhere in the world.

With Stripe Managed Payments, we take care of the hard parts, like tax compliance, fraud protection, dispute handling, fulfillment, and even customer support, so you can focus on building products and growing revenue. We’ll start out supporting digital goods businesses on Stripe Checkout. Over the course of the year, we’ll be rolling out to more payment UIs, more business types, and more countries. Now, having customers around the world means it’s important to provide local buying experiences. These are checkouts that are native and familiar to every customer, wherever they are. That’s been a big focus of our roadmap for our Optimized Checkout Suite, which is a set of prebuilt UIs and tools designed to maximize conversion. A big part of that local buying experience is showing prices in local currencies, because, in fact, 90% of buyers choose to make a purchase in their local currency when given the option.

That’s where Adaptive Pricing comes in. It localizes prices in more than 150 countries, and businesses on Checkout and Payment Links today are seeing an average of 17% uplift in cross-border revenue when they turn on Adaptive Pricing. So, today, we’re expanding the support of Adaptive Pricing to more services, so businesses using Elements and Hosted Invoicing will be able to start seeing those benefits, too. We’re also now offering access to more than 125 payment methods on the Optimized Checkout Suite, more than 25 of which we added in the past year. We’re still working on adding dozens more, including Bizum in Spain, PayPay in Japan, and Wero in Germany. As of today, we’re also now offering custom payment methods, which will allow you to display any payment method with any payment processor right in the Payment Element.

But, you know, with all these payment methods that we’re adding, we’re actually kind of introducing a new problem, which is which payment methods to display to each customer to maximize conversion. These are all really highly local, highly personal preferences. They’re almost impossible to predict without dynamic, real-time responsiveness. So that’s why we’ve been using AI to personalize payment method presentment, so that instead of relying on rigid rules, our AI models dynamically determine what payment methods to display, in what order, for every checkout session. So a buyer in Singapore, for example, may prefer PayNow for most purchases, but Klarna for high-value items, and then a travel rewards card when they’re out of the country. The Optimized Checkout Suite recognizes the difference and uses AI to adapt the checkout accordingly.

We’ve seen that this personalized approach actually translates into measurable business results. When at least one additional relevant payment method beyond cards has dynamically surfaced, Stripe businesses see an average 12% increase in revenue. This year, we’re getting into even more experiments about how we can personalize even more parts of the checkout experience using AI. One place that we’re really excited about is layout personalization because not every checkout feature is equally effective in every placement. For example, if you think about a buyer in Oman, they read from right to left, and so they might be more likely to add an upsell when it’s presented on the right side of the screen. What we’ve seen is that personalizing these layout decisions—you know, where to put an upsell, or a cart, or a promotion—can actually significantly boost your conversion globally. So you can expect a lot more from us to come here.

Lastly, we’ve heard that one of the most valuable insights to have is benchmarks and recommendations based on your peer set. So coming later this year, we’ll be launching a global command center within the Stripe Dashboard. It’s a comprehensive view of how your business is performing in every market, where your customers are buying from, and where your business is best positioned to expand. All right. Now we’ve just covered a whole bunch of payments, products, and features coming this year. I’m sure you have questions, and so I’d like to bring our speakers back to the stage and get into some Q&A.

All right. Welcome back, speakers. So, Roshan, you talked a little bit about Orchestration and the opening of the Stripe ecosystem. Could you share a little bit more about the vision behind this? Should users be worried that a PSP now potentially has access to their data?

ROSHAN SADANANI: Yeah, it’s a good question. Actually, we heard that a lot from customers we chatted with as we were building the product and kind of thinking of how to do this. What I’d start by saying is that I think the reason that we went down this path with Orchestration and the open ecosystem for Stripe is that we just heard from so many of you all that this is important, that you were working with multiple processors. I actually spent most of my day yesterday talking with customers who were asking for additional market coverage from Stripe and have had to go to other processors to do that. So really the sort of impetus for this change and product suite came from feedback we heard, just like you saw in some of the slides earlier, as well.

In terms of the data story, the approach we’re taking is that you’re in full control. So you have full auditability of being able to see how payments are routed. You specify the exact routing logic within Stripe. For those of you who attended the keynote, you saw an example of that with some of the UI builder that we have there. So really, we’re taking this approach of, “It’s your data” and we’re just helping route payments for you and kind of providing that sort of single interface for it.

SUSAN LEE: All right. A lot of people have been talking this week about the new Apple ruling and the bad week that Tim has been having, but it’s been a comparatively good week for mobile app developers. So the question—and I’ll throw this to John—is: how does Stripe support payments on mobile, and are our plans changing in light of this new news?

JOHN AFFAKI: So plans aren’t changing to some extent. I mean, it’s never been a technical limitation. We support online payments, mobile payments, and in-person payments in a very similar way. Our goal is to enable a very common interface. So, when we talk about our Optimized Checkout Suite, it actually spans mobile and web. We prioritize features in the same way. We want our mobile developers to have access to those features. I think the big difference this week is what you’re allowed to do. So we have the ability to link out to OCS, and that’s always been there. We’re also launching a new mobile Payment Element this week, which is going to enable you to build even more cool features directly on mobile apps. But while the ruling is going to be great for mobile developers—for all of you—it doesn’t really change our roadmap, to some extent. We’ve already been supporting you all in that path.

SUSAN LEE: All right. I’ll throw this one to Nicole. You talked a little bit about AI and payments, but what else have we not talked about yet when it comes to our roadmap for AI? Also, maybe for the group, how do we think about payments and how we build for payments in a world of agentic commerce, where it’s not just necessarily humans checking out, but it’s also agents?

NICOLE PAGLIA: Yeah. Actually, this is still payments. So one thing that made me think of it with your question that we didn’t cover today was how we can bring some of those AI payment performance optimizations to other processors. So, right now, we’re really focused on Stripe volume, but I think there’s a great opportunity to bring that to Orchestration. But that being said, payments is just the beginning. A couple places that we’re excited about. So one, I think, is actually more use of LLMs and assistants. So you see some of this today with Radar Assistant, which you can just use natural language to more easily write fraud rules. We also have tools in the Dashboard to help you find info or analyze your business. So I think there’ll be more coming there. You can think of performance recommendations all working with your Stripe Assistant.

The other one, agentic commerce. So this is one that, we’re very early but super excited. I think I actually saw a stat. It was something like 700 startups signed up on Stripe last year just focused on agentic commerce. So we really think we’re just kind of at the beginning here.

Agentic commerce—that’s basically where the agent is actually going online and making the purchase instead of the human. So we recently rolled out our agentic toolkit, which allows you to plug in Stripe into your agentic workflows. So for example, you can use Stripe Issuing to actually issue a virtual card that your agent can then go across the internet and spend for you. But I think it’s really going to be a rethinking of what does the checkout page have to look like and how do we build for kind of that future? So more to come there.

JOHN AFFAKI: Yeah, I mean, there’s also an interesting play. We’ve been focused heavily, historically on the merchant experience and how do we integrate and enable payments for you. But this agentic flow is actually going to create some very interesting buyer challenges where you’re essentially giving your payment information to a robot that doesn’t necessarily have context as to what’s going on in other versions of the bot. This is where things like Link can actually start to basically hold consumer preferences or spending limits. This is something that we’ve been building in general with our Issuing product, and how do you drive spend limits, permissions, and whatnot. So I think there’s going to be some very interesting opportunities to build with features for the consumer in this agentic world, too, on Link. So we’re thinking about it in both ways.

SUSAN LEE: Yeah, so if you don’t already have Link turned on in your integration, now is a great time to do so. We’re just coming up on time, so I’ll just ask one more question, which is, given all the problems there are to solve, how does Stripe approach prioritization in our payments roadmap? I’ll just throw it to anyone.

JOHN AFFAKI: We look at it at a few different angles, right? We look at what’s the demand. How are we going to enable the most users to do the most things? How do we drive the most revenue for you all? We like to make money, also, in some cases. But that’s not the only thing, right? We have to look at trends. We have to be able to sometimes blow up a roadmap when something is really changing, right? If you would have asked us two years ago about agentic commerce, we would have been like, “What the heck is that” on that front. But as soon as—we’ve actually been pushing in that direction, and once we saw traction, we’ve really pivoted a lot of our work.

AI hasn’t really been a new thing. We’ve been constantly evolving in it. But as we saw the strength of new LLMs and what we could do, we started leaning harder into it. So I’d say the number one thing is really what our users need, but we need to also make sure that we’re ahead of the trends that are going to start really making differences. We can’t just invest in what you’re asking for today. We need to start looking at that. So it’s a trade-off conversation every few weeks as we get new items coming in.

SUSAN LEE: Great. Well, I will end us there. So thank you, all, for being here. I hope this peek into our payments roadmap has been valuable to you all. Thank you.