AWS re:Invent 2025 - The next frontier: Building the agentic future of Financial Services (INV209)

[Music] [Music] Heat.

[Music]

Heat.

[Music]

Heat. Heat. N.

Heat. Heat. N.

[Applause] [Music]

Please welcome to the stage managing director of financial services at AWS, Scott Mullins.

Folks, good afternoon and welcome to the 2025 in innovation talk for the financial services industry here at AWS reinvent 2025. For those of you who

reinvent 2025. For those of you who don't know me, I am Scott Mullins and it's my privilege to lead the financial services business here at AWS.

I've been very fortunate to be a part of many of those watershed moments that you saw highlighted in the in the video that we just shared with you. And that was like a one minute kind of roller coaster ride through the last 13 years of what

we've been doing. And in the very first thing that you saw was uh one of the only and first sessions for financial services at the very first reinvent in 2012. And while some of the things about

2012. And while some of the things about reinvent have changed over the years, from the number of attendees that we have to the advancements in both AWS services and our customers use cases,

what hasn't changed is our shared willingness to continue to push past what is currently thought to be possible. For more than a decade now,

possible. For more than a decade now, we've been moving the industry forward together. At times with deliberate and

together. At times with deliberate and important incremental steps, and at others with leaps and bounds, but always with the conviction that we can make it better. And as a result, today AWS is

better. And as a result, today AWS is the home to the to the missionritical systems that power the global financial services industry. And together we're

services industry. And together we're actively evolving how financial consumers are served. Evolution happens

quickly in our industry because financial institutions have chosen to change instead of just waiting for change to happen to them. One reason

that we now stand on the threshold of a new frontier today where agents won't only chat with us but act on our behalf is that financial institutions have been preparing for this moment for years.

They've moved their data to AWS and set up processes and guardrails that would enable them to securely and confidently take advantage of machine learning and generative AI.

And in doing so, financial institutions were ready to embrace the opportunity to tackle long-standing challenges and transform the way they operate with agenic capabilities. This week at

agenic capabilities. This week at reinvent, we've heard from many financial services customers how agents are helping them rethink their core businesses and become more productive,

secure, and efficient.

Moody shared how the company deployed multi-agent systems to process large data sets and conduct specialized research and analysis.

Ripple shared how it transforms security operations by building a multi- aent system that analyzes massive log volumes and enables guided investigations.

And Commonwealth Bank of Australia shared how it's using agents to automatically analyze legacy code, perform security assessments, and handle network flow implementation. And that's

just a sample of the industry focused breakout sessions our customers are delivering just this week and a fraction of the agentic AI use cases we're tracking across the industry today.

The common thread here is that financial institutions are entrusting agents with increasingly more critical responsibilities and allowing them to interact directly with their customers, handle sensitive data, and execute

complex tasks. But even more powerful

complex tasks. But even more powerful change is on the horizon. Autonomous AI

systems, not just goal-based agents, are poised to transform the way that organizations build financial services and the way that we all consume this.

So, what will this look like?

Well, across the next frontier, Agentic systems will more and more become a part of our daily lives.

Let's see how this looks. So, let's say you've been offered a job that would require you to move and maybe need to buy a new house.

Because my personal financial agent knows me, it can look for properties that align to my taste, needs, and financial resources. You can see I I

financial resources. You can see I I have a particular taste in the style of home. Imagine that it finds a property

home. Imagine that it finds a property that looks really great for me.

But before I move forward, my agent actually takes a deeper dive into the disclosure statement for the house I'm interested in and finds out that the basement flooded a few years ago. That's

actually happened to the home I bought recently. My agent then accesses a

recently. My agent then accesses a weather database on its own to find out that changing weather patterns have put this home in a risky flood zone going forward.

This is the type of action that agents can proactively take on their behalf today to save us time and money, but also to minimize risk.

Agents will also do things like analyze your financial situation in real time and evaluate your current holdings to see if you're properly balanced and shift funds accordingly to make sure

that I can actually afford that home.

Your agent will make it easier for you to make informed decisions when it comes to evaluating which equities I might want to transfer around or sell to minimize capital gains taxes or how much money you should set aside for perhaps

your kid's college fund based on your current earnings and savings. They can

independently perform research to make a recommendation and then present enrollment forms. So, Agentic systems will make real really make life easier for consumers like you and me. But

they'll also make life easier for providers as well.

Let's say your agent thinks you may want to purchase some life insurance based on your recent interactions. Well,

insurance underwriting is a difficult and timeconsuming process because it requires manual data collection and extensive internal and external research and analysis. Agents are already

and analysis. Agents are already streamlining this process by automatically extracting data from submission documents. But in the near

submission documents. But in the near future, autonomous AI systems will go a step further by performing risk assessments and then enabling agents to negotiate prices with other agents.

New capabilities have put this vision of the future within reach. The length of task that agents can accomplish has increased from seconds to hours while costs have decreased from dollars and cents. And we heard from Matt Garmin on

cents. And we heard from Matt Garmin on this on Tuesday. But for this autonomous agent-to-agent future to work at scale, we'll need more than just technical capabilities. We'll need trust.

capabilities. We'll need trust.

Organizations will need to trust that autonomous systems can operate securely, comply with regulations, and protect sensitive customer data. And as an industry, we're going to need to align on standards and protocols that

facilitate and incentivize trust between companies. The path forward to this

companies. The path forward to this trustbased future starts with Amazon Bedrock. And as you heard from Matt

Bedrock. And as you heard from Matt Garmin in his keynote yesterday, we continue to add new foundation models, 18 new foundation models to the broad selection we already provide through Bedrock so that you have the freedom to

choose the right model for the right job. Matt also introduced and announced

job. Matt also introduced and announced two new features within Bedrock agent core called policy and evaluations. Two

new capabilities to help you deploy and scale Agentic AI systems securely without operational headaches. And I'm

really excited about those two features in the financial services industry.

Today we're here to learn how financial institutions are using Amazon Bedrock and other AWS services to build the future of the industry. And we're

fortunate to have two customers joining us to share their experience and guidance. First, we're going to hear

guidance. First, we're going to hear from Axel Shell, the CTO of Alian Technology, who will share how one of the world's largest insurance companies has built a framework for bringing AI powered solutions to production

securely, consistently, and quickly. And

then we're going to hear from Eric Repel who is head of engineering of Coinbase's developer platform. He'll share how

developer platform. He'll share how Coinbase is making the vision of Agentic Commerce a reality through its X42 protocol. So now, please join me in

protocol. So now, please join me in welcoming Axel from Alian to the stage.

Thank you.

So pleasure to be here and what we want to do is we want to talk about moving the frontier to the next level. One of

the questions I get regularly asked is how will business and our technological environment look like in the next three five or eight years and that's a very

difficult question that's the famous $1 million question. And one of the answers

million question. And one of the answers is and that's maybe something that we can say is in short almost everything

will change in the future. We will for sure shift boundaries and frontiers.

Now let's try to understand why moving frontiers and boundaries is first of all very important but also secondly why is this relevant these days? If you look at the current world we're living in, you

see we call it poly crisis world where lots of things are happening. Number of

ma outleted increases more data breaches, more natural disasters, a lot more deep fake related to AI but also

global terrorism and IT attacks increase over time. So you see we live in this

over time. So you see we live in this poly crisis world that makes the environment quite difficult. And when we look into this global perma crisis, you

see what's going to happen. First of

all, we have a very fragile geopolitical situation.

Climate change is coming and approaching and that's relevant for human and also for our economy. And you see lots of demographic change that influences all

of our lives. And last but not least, there's something called AI and digitalization, which of course adds on top of all those elements. And I think

one thing that we can say is that this results in an environment of change and lots of uncertainty.

Now the key question is if we live in this perma crisis poly crisis world

the key question is is it these days a catalyst of change so is it actually something that is the core of disruption and makes the disruption even faster or

is it an instrument of resilience and that means if you're ahead of the curve and you're more or less leading the pack you're early enough then this can be an

instrument to create resilience. So does

it disrupt or does it create resilience?

And what is absolutely clear is that if we want to look at the future and to advance the boundaries and to move the frontiers to the next

level, one thing is very clear. We have

to manage our future. The future of IT and the future of companies. Therefore,

whatever world you live in, we have to actively play a role in managing this future. Now, the question is how do we

future. Now, the question is how do we do this? So, very quick introduction

do this? So, very quick introduction about Aliance that you understand a bit of the background. We're roughly 150,000 employees. We operate in 74 countries

employees. We operate in 74 countries and um we are the leading financial services insurance brand and we have roughly 2,200

billion assets under management. That's

the responsibilities that customers gave us. We are 135 year old company

us. We are 135 year old company and the key question is now if we look at the IT the disruption the resilience what's going to come there. One of the

key question we ask ourselves on a daily basis is how do many people shop around for insurance today and how do will they do that in the future. If we look at how

they do it today, there are lots of aggregators. You go on an aggregator

aggregators. You go on an aggregator page, you get a couple of comparison for insurance product and then you maybe decide for which one you want to go to.

But the question is now is AI maybe the new aggregator and I have an example and let's start from the end customer which is here my

cat. So what we do now is we buy a pet

cat. So what we do now is we buy a pet insurance for my cat in a modern way.

All I do is I upload this animal passport with a couple of key information on which is nothing else than the name of the cat, when was it born, the gender. So, a couple of key

information. That's everything that I

information. That's everything that I use. And what I'll do now is I use a

use. And what I'll do now is I use a single prompt to buy an insurance product. The single prompt is order for

product. The single prompt is order for my cat um an pet insurance product with a best price to performance perspective.

Proceed to the checkout. Do as less as possible interactions. And what you see

possible interactions. And what you see now is here an agent. And then you I think you're very well aware of the agentic browsers that are out there, the

comets of the world, the atlases of the world. We just use here one example. Go

world. We just use here one example. Go

through the process and starts to look at various websites, tries the comparison. The whole thing takes

comparison. The whole thing takes roughly four to five minutes. I will not go through all of the details now, of course. And it ends with buying a cat

course. And it ends with buying a cat insurance or a health pet insurance for my cat.

So I have already three pet insurancees now because once I tried it the first time, the second time I recorded the video and the third time when I did a live demo. So I will not do a live demo

live demo. So I will not do a live demo today. Uh three insurance is already

today. Uh three insurance is already enough. Um by the way from different

enough. Um by the way from different brands which is also an interesting experience.

Now what you saw here is aentic browsing here specifically for financial services insurance products. What you see here is

insurance products. What you see here is that we move away from in the old web days a grammar constraint decoding to now very reliable or at least some more

much more reliable tool calls. We move

away from browser APIs that gencast can do read modify or click or navigate to navigation in real web UIs. They're

completely flexible. The agent doesn't care how the website looked like if they did a change on the design the day before. The navigation now is different.

before. The navigation now is different.

And those agents have lots of capabilities. They plan, act, observe.

capabilities. They plan, act, observe.

Now, correct. They do multiple times.

There's reasoning behind. So, they're

quite good. That gives a lot more capabilities. You have now full web

capabilities. You have now full web workflows, not only for a web page, but also internally for some of your web applications maybe. And we can pass and

applications maybe. And we can pass and combine data from various sources. And

of course, it's also great if you just open a new tab, you can navigate across multiple systems. Now what it means there is a universal

automation layer now existing with aentic browsing capabilities for all kind of web applications including legacy applications even some of them that don't have proper interfaces and

don't have APIs and it replaces also a bit the custom integration with one agentic interface how great your web design is is no

longer relevant for the agent that's not the most important thing um And of course my productivity um increased a lot. I mean the prompt writing the

lot. I mean the prompt writing the prompt took me 20 seconds and that's it.

I just had to check then if I'm happy with the result. But also productivity is completely different.

Now what does it mean and why is this relevant? Now I think we are very well

relevant? Now I think we are very well aware of the world where a user interacts with a website.

Now we live in a world where maybe the user in the future interacts with an agent and ask the agents to do certain activities on the web procuring buying products.

But is this the world that we think that everything will end up with? I don't

think so because we don't want as an alian insurance that thousands of agents crawl our website and click on the buttons. I mean, what you just saw is

buttons. I mean, what you just saw is like large language model processing pictures and then telling the mouse to move 50 pixels up, 50 pixels down, and click on the next button to see what happens. Then take a screenshot, bring

happens. Then take a screenshot, bring it back to the large language model, understanding what's on the screen, giving instructions to the agent. That's

not what you want. A lot of compute, a lot of capacity is needed. So, what will happen is that we will also have our own agent. The Alian agent then shows up and

agent. The Alian agent then shows up and says, "Ah, cool. You're an agent. I'm an

agent as well. Let's interact with each other in a completely different way um with lots of protocols and also payment will be interesting and Eric later will um talk exactly about this kind of

agent-to-ag payments which will be a very interesting journey as well. So agents

communicate with each other in a totally different way.

Now that was for us the starting point to understand how our AI journey looks like in the next couple of months. I

mean nobody knows what happens in two years. We live in the famous cat and dog

years. We live in the famous cat and dog world of AI where uh years of AI where time moves so quickly. But we started to

work with colleagues from AWS to do some kind of code development and to better understand the basics of um agent capabilities. And we selected an area

capabilities. And we selected an area where we started with the team um of our risk colleagues and they spend a lot of time gathering and aggregating

information. That took a long time and

information. That took a long time and um what we actually wanted to do is to make their time there's yay time and nay time. The yay time is stuff they like.

time. The yay time is stuff they like.

The nay time is just data gathering which they don't like. We want to shift the yay and the nay and um move the 20% of their actual work into a much higher

productivity level. So we wrote a kind

productivity level. So we wrote a kind of model agnostic multi- aent framework that we collaborated on. You see here

how the features look like Amazon bedrock we use clot um as large language model and that kind of model is first of all built in a way our framework is

built in a way that it's an open model selection is possible we have of course agents that work and we heavily rely on MCP this is how it looks like from an

architecture perspective this is our multi-agent framework architecture nothing super fancy um I think you would expect how that looks like you have an S3 bucket where we store the data. We

have our VPC endpoints. We have our Genai lab agent. This is actually our home for the agent where the agent lives. And then we use LAN graph and

lives. And then we use LAN graph and also some Fargate capabilities. Um we do network address translation with the internet and then process everything and

also um integrate.

Now what why is this important? Again

I I like this slide a lot because it shows the journey where we come from.

You remember all of the times where we heavily discussing that monolithics um monoliths are not what we want. We have

to move to microservices. So more

developer efficiency very simple simplified operations and tech stack. So

this is what we thought but there were some downsides. Duplication of effort

some downsides. Duplication of effort inconsistent standard increased complexity complex governance. I think

you are very well aware of the challenges of microservices. The second

thing is on cloud. Everybody said we have to move away from on-prem and have to move to cloud and there were many advantages. I just put one here. Cost

advantages. I just put one here. Cost

reduction, improved resilience, many benefits from cloud. But it also comes with downsides. You have a technical

with downsides. You have a technical lock in through maybe lift and shift.

You have maybe increased parallel run cost, complex transition and so on and so forth. Just a couple of downsides.

so forth. Just a couple of downsides.

And what I want to compare the agentic world of the future is exactly with those problems. We will love agents.

They will do unbelievable cool things.

But they will come with a lot of downsides.

And we should be aware of those already now because if we don't think about those, if we don't have a proper

governance um a proper structure how to work with agents, we will run into many many problems in the future. Now it's

still manageable because it's early stage but we have to think now about this. What I foresee is something like

this. What I foresee is something like if you look at what many companies do they start with these siloed PC's

there's a meeting preparing assistant co-pilot supporting assistant legacy modernization HR bot um source code management public cloud as underlying

applications. So many of those agents

applications. So many of those agents there is very low reusability of those agents across the verticals and they are very tightly linked to use cases and

incremental value realization is what happens. So this is the world of the

happens. So this is the world of the many PCs. It's nice. The use case is

many PCs. It's nice. The use case is great. But how do you actually really

great. But how do you actually really leverage the power of bringing things together? Is this something that is

together? Is this something that is scalable? And we should never forget one

scalable? And we should never forget one can key sentence which is you can scale everything but you cannot scale complexity. You cannot scale complexity.

complexity. You cannot scale complexity.

And there's high risk on the left hand side to run into very complex environments. Now what you need is

environments. Now what you need is something maybe even more on the right where we call it genai mesh where you have certain agent discovery

capabilities, registry and orchestration capabilities where you try to reuse um agents across any of those use cases.

you need a kind of a flexible plug-and-play system, no lockins and that will lead to um at scale value realization, not only incremental value

realization but really at scale value realization. So keep that in mind when

realization. So keep that in mind when you see your agents growing you need to find a way how to move into this genai mesh and move away from vertical pilots.

Now how do we do that? This is our current thinking and I'm not saying that this is the best what you can do but this is what how we try to solve the problem and this is why I think this is

relevant. We started to define certain

relevant. We started to define certain levels for agents. We have first of all our system of records like um databases policy assurance case database solution

database whatever and then on top we stack more or less some agentic layers.

We have first of all utility agents on level three that can be highly reused for documents, for video imaging, for processing, for voice, for email. On top

we have so-called business agents on layer two that are more business specific. They deal with notifications,

specific. They deal with notifications, with solutions, with case management, with policy assurance. And then we have level one planner agents. They

orchestrate the workflow. And all of that is then linked into our customer journeys. And with this kind of

journeys. And with this kind of architecture and reference structure, we're able to utilize agents in a way that the reusability is much higher that creates

less complexity that leads to more harmonization and is a lot easier to scale.

The way how that works is we call it um framework an agentic architecture framework with more or less four components. There's first of all the

components. There's first of all the core stack where we have front- end orchestration and also AI workflow tooling. You see here some of the

tooling. You see here some of the technologies that we leverage. We have

our DevOps tools which is mainly around automation but also the continuous deployment processes and then of course some AI models. Flexibility needs to be

built into the architecture and you need some foundational service layers as well where you for example do all of the container orchestration

where you have also the possibility to work with API management. you have the full stack observability and the way how this looks like is

something like this.

What you see here is a bit an agent control plane that we have developed that is kind of a marketplace internally because as I said observability

reusability only works when you can also discover those and what you more or less need is something like an organized marketplace

where a space where you can collect showcase and engage those agents.

You need to administer them properly. Um

and that includes of course config. You

need to um monitor them. The consumption

there will be for sure after the phops world also something like an agentic ops where you just need to reduce cost because of the LM models will at a certain point become quite expensive and

of course you also need some kind of agentic prompt playground where you can experiment with new agentic applications.

And if you do one more level deep dive into this, I have here an example which is a claim and claim handle agent. These

are the kind of agents you would find on the marketplace in the various L1, L2, L3 levels which you can draw from. And

you have here a couple of capabilities.

You can either use MCP to connect or also connect with a REST endpoint. You

see what tools are available that are coming with the claims handler agent.

You see for example also the possibility to run a test case. You enter a test case number. You execute the tool. You

case number. You execute the tool. You

see the results so that you can play around with it and understand what the agent can do. And then you see also the on the right hand side the very important step all of the traces and the traceability what the agent is actually

doing because in a regulated industry you have to make sure that you know what the agent also did. So you see for example the reasoning behind you see all

of the stops docu the stop the steps properly documented you see when the execution is complete you see timestamps you see audit logs all of that on the

right hand side and this is the kind of setup that we use in order to join this new agentic world aentic world

from my point of view will grow extremely fast I Two years ago, we didn't discuss agents. Last year, agents were a big thing. This year, I have a bit the feeling that people now

understand how powerful those are and what is going to happen in the future.

And we have to prepare ourselves to be ready for many many agents that we need to manage, orchestrate and um also as an

IT govern properly.

Now, let's imagine and let's leverage the agentic capabilities for a second.

Let's imagine there's a blackout, no power, and what happens is your fridge also has no power and everything that is in the

fridge is no longer eatable after a certain amount of time.

This is one use case that we have implemented where we use agents to do claims processing and also settlements.

We have a couple of agents. First of

all, we have a planner agent that initiates and coordinates the workflow.

This is, as I said, on level one, the one that does that. We have a cyber agent that um works on the security part. A cover agent that actually

part. A cover agent that actually verifies if the insurance c customer is actually covered by his policy. There's

a weather agent that actually looks, was there really a power outage happening?

Was there, I don't know, a thunderstorm or something like that. There's a fraud agent who looks into all of the fraud capabilities.

There's the payout agent in the end who does then the payout to the customer.

And there is an audit agent. And what we could do is we could reduce our processing time by 80%. Um something

that took more or less around 100 days and now we have seven agents in a workflow and it's a very powerful tool that many of our um customers like and

this is how it looks like.

So what happens is you start your mobile app, you welcome, you get a welcome, you get welcomed by the agent and it says, "Hi, my fridge lost power during the

storm and all my food has gone bad. I'm

really sorry to hear." First of all, it wants to understand who are you. So

basically, you provide your email address. And then the next step is it

address. And then the next step is it wants to understand the full name, date of birth, home address. Understood? The

address has been verified. And then it continues with when did this happen?

Please send me a picture of the fridge.

Um, and then the agent does all of that.

And in the background, this is what you see on the right hand side. What the

customer cannot see is all of the various agents that are starting. Now

the coverage agents, the fraud agent um, is looking into this and then of course the weather agent will check if there is a weather incident. The payment agent

does the stuff and so on and so forth.

And you see then that in the end it ends up with all of the audit trails and the logs why we took a certain decision how much the payout is so that we are

covered from a regulatory perspective.

So I think that's a good example that is one use case how you can work with agents but now imagine we run thousands of those use cases and we have to make

sure that we have all of this under control. Therefore, better plan now for

control. Therefore, better plan now for what's going to come in the next one to two years. Thank you.

two years. Thank you.

[Music] [Applause] So, Axel, thank you for sharing how you've built a platform for systemically

and securely deploying agents to reinvent insurance. Now, a theme we've

reinvent insurance. Now, a theme we've heard a lot this week is trust. We need

to be able to trust agents to do what we need them to do and to act responsibly on our behalf. And as we move further into our future, we need to trust agents to work responsibly with one another.

But that trust isn't going to magically manifest itself. We're going to need to

manifest itself. We're going to need to create and to incentivize it. Which brings me to our next speaker, Eric Reppel from Coinbase. Eric is going to share with us

Coinbase. Eric is going to share with us how Coinb Coinbase is taking pragmatic tangible actions to build the trust that will form the foundation of the agentric future of commerce. So please join me in

welcoming Eric to the stage.

[Music] So thanks so much Scott. I'm super excited to be here. I'm Eric. I lead engineering

for Coinbase developer platform. We

we're kind of like the AWS to what AWS is to Amazon, CDP is to Coinbase. We've

taken the tools and capabilities that Coinbase has built over the years around blockchain infrastructure, custody, payments, and uh we're now externalizing those services and making them available

to other customers who want to build on top of our infrastructure, which we of course build largely on top of AWS.

So I'm going to talk a bit about a vision for the future of payments. But

where I want to start is I want to start with the internet. So the internet was really designed for humans. We as humans use the internet all day every day. And

we kind of use it through these devices that we have. And a lot of the standards that we have to use the internet are really built with this concept of human in the loop where a human is driving a

computer that is then interacting with other computers. And in web 1.0 back in

other computers. And in web 1.0 back in the 80s we got really robust standards for sharing information. The internet

really started in academia with researchers wanting to share papers and content and their work with other people around the globe. And that was great. We

got HTML, we got HTTP, we got the worldwide web client, which we now call a browser, much better name in my opinion. And then we we moved on. We

opinion. And then we we moved on. We

went to web two. In the 90s, we got web two. And we kind of got the explosion of

two. And we kind of got the explosion of utility that we see today. And that was really powered by standards like JavaScript, Ajax, JSON, which let you

write. The web was no longer read. Now

write. The web was no longer read. Now

it's read and write.

And really my point here is the internet evolves through standards. Standards are

conventions that we all agree on.

There's no controversy involved that you're going to write your website in HTML. And the power of that is that any

HTML. And the power of that is that any browser that you use works with any website that you visit without there having to be some kind of bespoke agreement that goes on. Every company on

the planet writes HTML and uses HTTP.

Uh, and that openness creates a positive sum feedback loop where things that were possible, things that weren't aren't possible if

you have to negotiate one-on-one are now possible.

And so all these formats do very specific things and they've largely answered a bunch of these questions like how do I give you information? What should that

information? What should that information look like? But we've never actually managed to solve the problem of how do I give you value? There's

actually no standards, no open standards tied to the internet for payments.

And there's this Mark Andre quote. I

think the original sin was that we couldn't actually build economics, which is to say money at the core of the internet. And we kind of never resolved

internet. And we kind of never resolved this. This is a a quest Mark Andre went

this. This is a a quest Mark Andre went on in in the 80s and 90s to try to make payments native to the internet, not something that's done kind of in the requests that happen, but payments that

can accompany the requests as part of the part of the core backbone of the internet. But eventually we got HTTPS,

internet. But eventually we got HTTPS, we got encryption, we became comfortable sharing credit card or bank details in internet requests and consumer ch

behavior changed. They became

behavior changed. They became comfortable typing in a credit card and hitting send and sending their information to a company to process offline.

And we mostly forgot about this idea of internet native payments.

But you see remnants of it. You see

artifacts in the history of engineering that's happened over the years that indicate to this forgotten past. And one

of them is this 402 payment required status code. For those of you in the

status code. For those of you in the audience who aren't familiar, HTTP status codes exist to convey a programmatic message of what went wrong, what happened or what went right. What

happened when I your client made a request to my server? If I send you a 200, everything went perfect. 200 just

means okay. Probably everyone in this room has seen a 404 not found. 402 was

encapsulated at the same time 404 and 200 were. And this is the entire

200 were. And this is the entire definition of the of the standard 402 payment required status code is reserved for future use. And so we never actually

got around to using the status code.

It's not it's kind of used occasionally in payment workflows. But there's no standard that works alongside 402 to have a consistent open payment experience like you can have where you

visit any website, you download their HTML and you can now view the content of the website. Each API that you need to

the website. Each API that you need to for payments that you need to integrate is slightly different. Have slightly

different conventions. It's not a do one integration. You get payments

integration. You get payments everywhere.

And payments are the hardest part of the internet.

But they're the most important unlock in the last 25 years. Checkout is a very humanentric experience. You see,

humanentric experience. You see, everyone in this room has probably filled out a form that looks similar to this where you put in a credit card number and an expiration and a security

code and an address and you know repeat UX is is quite good. Typically companies

will store the credit card but first time UX usually you have to go through this step. And while visually these may

this step. And while visually these may look very similar between different services or different websites, technically they're very different. They

may have nothing in common between different websites you visit. Visually,

this looks it's similar, but what the machine sees, what a computer sees could be entirely different.

It's also complicated. This form has 14 fields.

It's a, you know, it's we've gotten used to it as humans. It's it's an okay experience. I'm being a little being a

experience. I'm being a little being a little spicy on this slide and saying it's a bad experience for humans, but it's okay. It's worse for agents.

it's okay. It's worse for agents.

Agents are really probabilistic systems which means that they may not do the same thing every single time. I'm sure

all of us have experienced this using chat chat interfaces or LLMs where you know you ask your LLM the same question.

First time it gives you a great answer.

Second time the answer is a little different. Maybe it doesn't quite hit

different. Maybe it doesn't quite hit the same points. And what that means is you if you have to probabilistically accomplish 14 tasks in a row, the probability of completing the entire

task is now dependent on the entire set of tasks being complete. And so if you have a 85% probability of completing the fir each field, you're now tsing 085

time 08.585 time 085 and you can get flaky really quickly.

And agents also have to be really good for you to trust them with your credit card. I think all of us are in this room

card. I think all of us are in this room would be a little hesitant to give unfettered access to their credit to an AI agent at this point in the in the cycle.

And so AI really forces us to re-evaluate internet native payments.

Payments that are adhere to a standard and that works everywhere. Similarly to

how HTML works anywhere the internet works.

And I think at Coinbase we have this broad thesis that stable coins are a really good use case for AI agents and work extremely well for the agentic

commerce scenarios. And you know, look at the

scenarios. And you know, look at the robot. He goes from sad and then you

robot. He goes from sad and then you give him a stable coin, he gets very happy.

And what I'll talk you through today is how we've been developing an open standard for payments, uh, internet native payments, and how it can leverage stable coin and crypto. It will it'll

work with fiat as well over time. But

what how stable coins have really uniquely enabled some things in agentic commerce that I think are very difficult to replicate.

So what is a stablecoin? A stablecoin

for those of you who may not be familiar is is a blockchain token that is backed one to one typically by reserve dollars.

So the the stable coin Coinbase is known for is uh USDC. USDC has audited reserves in treasuries and liquid dollars and bank accounts backing each

digital dollar that exists. Uh, and what that unlocks is because blockchains have gotten really good over the last 5 years. You have 247 availability, you

years. You have 247 availability, you can settle transactions in seconds. Not

T+2 now means not t plus2 days, but t plus2 seconds.

And the fees are incredibly low. The

blockchain level cost to send any amount of money, whether it's a cent or a million dollars, typically on a modern blockchain like base is about a tenth of a cent.

This is true of many blockchains. Base

is of course a great blockchain, but Salana, Sweet, Aptos, many blockchains have this like incredibly quick to transact, incredibly cheap to transact property.

And so, let's get to let's get to the the meat of this. An open standard for internet native payments. This is where I introduce X42. So, X42 is an extension

of the 402 status code. We want to leverage the existing infrastructure of the internet.

What is it? It's an open standard for internet native payments founded at Coinbase. It's soon to be moved into an

Coinbase. It's soon to be moved into an independent foundation. So the

independent foundation. So the governance of the standard is not just done by Coinbase, but it's done by anyone who wants to participate and invest in this idea of an open standard.

But really X42 is a bridge. Payments

kind of exists on one side of the the water. The internet kind of exists on

water. The internet kind of exists on another. X42 aims to build a bridge so

another. X42 aims to build a bridge so that payments on the internet feel native and we can move value in a similar way to how we move data.

And X42 kind of exists in what I'd call the Agentic standards stack where you're starting to see things like MCP that standardize how tools are defined and

how tools are used. X42 handles payment.

ADA standardizes how agents communicate with each other. And you're starting to see more experimental standards like EIP 804 for agentic reputation and discovery. And there'll be more to come.

discovery. And there'll be more to come.

There's since I've made this slide, ACP started to emerge, you're starting to see AP2. All of these standards, I think

see AP2. All of these standards, I think similarly to how HTML and CSS and JavaScript and HTTP are the standards we needed to make the web really sing.

You're going to see the composition of these standards work together to really unlock the agentic age.

So, how's it going? Well, in the last in the last 30 days, X42 has done over 70 million transactions using real stable coins and real dollars. The GitHub is

incredibly popular.

There it's had over 4,000 people star the GitHub. There's been almost 100

the GitHub. There's been almost 100 contributors. Uh, and there's been over

contributors. Uh, and there's been over $30 million of volume. I believe it's actually about 38. This is uh slightly outdated. Uh, and there's been over a

outdated. Uh, and there's been over a quarter million people who have bought a good using X42 or a service using X42.

And commonly that'll be data or that'll be access to information or API use is a common use case. Anytime you have a digital native good, X42 works incredibly well because it natively

integrates into your infrastructure, natively integrates into your web servers, uh, and allows you to include payment directly in the exchange of data that is already happening.

So what does this enable?

So what I've done here is I've given my LLM access to two X42 tools and I've asked it to do some research for me.

I've asked it to research X42. How has

X42 usage and adoption been going? And

you'll notice it it does a web search to try to understand the X42 protocol and it gains information that's available on the public internet and then it searches the first it uses the first tool to

search for paid resources that may have information about trends on social media and then it will actually just go and pay 5 cents to access that data. It I

don't need to go and create an API key or integrate an MCP. That's an API that exists and it can pay directly for that access to content.

And what you see here is I can create an experience where this agent has dedicated siloed funds. It has about 56 cents and I can create rules about what that agent can do and how it can spend.

And these isolated funds, there's no way for the agent to spend more than the funds I've given it. I can say, hey, you're only allowed to spend 25 cents per session or a dollar per day. And it

becomes trivially easy to codify these kinds of rules into the experiences that you build because these dollars are digitally native. It

really is money that is truly digital versus kind of systems that you need to interact with via APIs that are not money moving from account to account. And the great thing

about this is because this is open standards, anyone in this room with a good engineering team can build exactly this product that we have. And any

service that's available that accepts X42 payments can be paid for in this server using this this client using this MCP.

And this creates what I call dynamic agents where right now if you have an agent, you kind of need to integrate one tool at a time. And your agent kind of operates in a box that you define for

it. And the capabilities that it have

it. And the capabilities that it have are really set by you. And if it doesn't have a capability to to accomplish something or to pick to use a tool that

may be useful for the task it's trying to perform, it's just not going to be able to do that. But if now you open the aperture and you have an open standard where anything that accepts these

payments can happen without a pre-existing negotiation, your agent can find anything available that accepts these kind of payments and

just pay for the the services that it uses. And because stable coins are a

uses. And because stable coins are a tenth of a cent to transfer, it means that paying a cent or 5 cents or a dollar becomes truly possible. If you

you might have noticed on the slide a couple slides back, there's been about 70 million transactions, about $38 million in volume. That's less than a dollar on average per transaction, which

is actually quite the feat. And this

works for payments from, you know, a cent to a million dollars. Fees are the same. The cost is the same. There's no

same. The cost is the same. There's no

percentage based fees. In fact, X42 doesn't have any fees in it at all as a standard. There might be costs

standard. There might be costs associated with doing the underlying payment networks transfer, but there's zero fees at all at at the standard

level of X42.

So why now? Why now? Well, AI is the obvious answer, but I think there's some more subtle ones. I think stable coins of product market fit and are here to stay. We're getting more and more

stay. We're getting more and more clarity on the regulatory and compliance side of stable coins and more and more countries are investing and looking into what stable coins will look like and how

the benefits of stable coins can serve their constituents.

The UX for blockchains has gotten really good, really, really good. And it gets better every single day. I think you can now create experiences that leverage blockchain technologies that as a human,

as a user, you don't even realize it uses crypto.

But you get all the benefits 247 lowcost instant settlement.

And now as I alluded to at the beginning, AI agents are really going to force nonhuman loop payments to exist where the human is not always driving

the use of the computer. The agent is now driving the use of the computer and AI breaks the fundamental economic arrangement of the internet in many ways.

So if you think about the internet, roughly the consumer web monetizes in this pattern. I'm handwaving a little

this pattern. I'm handwaving a little bit, but typically as a user, you'll see content and ads served to you by a website. In the background, a merchant

website. In the background, a merchant has done an ad placement and paid that website to show you that ad. And then at some probabilistic uh rate, you will

convert. You'll see the ad and you'll

convert. You'll see the ad and you'll buy a product from the merchant.

Except the issue is with AI, you end up not seeing the ad. You may

have noticed, but when an LLM scrapes a website, it largely focuses on the content, not the advertisements.

And so, if you're not seeing the ad, how does the payment occur? I think this is going to be a big problem. And we've

seen a lot of this in big labs going and scraping data and the publishers not being able to capture any of the value that they're creating.

And what X42 does is you can now directly pay for the content within the same request that happens. It's a much more native to the internet way of moving value. Your AI can go to a

moving value. Your AI can go to a website and say, "Hey," the website can notice it's an AI and say, "Hey, I want 25 cents for you to access this content. This is a really

valuable blog about the best restaurants in Vegas to eat at. If you're going to consume this, I want 25 cents." That may be a fair exchange, right?

So how does this work? So X42

exists in this flow between an agent, a server, what we call a facilitator, and when you're using stable coins, a blockchain. The facilitator really

blockchain. The facilitator really exists to abstract away the complexities of building in crypto and make it feel like just building in the web.

You can integrate X42 into most languages and frameworks in as little as one line of code.

And so given that flow, you can do things like leverage Amazon Bedrock or other uh tools that fill these open standards to build aic discovery and

agentic payment. And uh AWS provides

agentic payment. And uh AWS provides great primitives with via Bedrock and Lambda and EC2 and the full stack, both the host servers that monetize with via

X42 and to build agents that can pay with X42.

And if you're using stable coins, you you likely need to have a wallet. CDP

offers best-in-class wallets that uh are great and very well optimized to work with agents and to perform operations with stable coins. The demo video that I showed you is built on top of our wallet

stack. Uh we like to eat our own dog

stack. Uh we like to eat our own dog food. Uh and our wallet stack is built

food. Uh and our wallet stack is built on top of AWS Nitro uh and the scaling and KMS systems that Amazon offers.

And so what does it look to actually bring these things together? If you're

using Bedrock, you can use Coinbase's complimentary agent construction kit called agent kit, which gives your agent a wallet. You can connect the wallet.

a wallet. You can connect the wallet.

You can tell Bedrock about these tools that exist, the tools for discovery and payment that I mentioned in that demo.

And then in maybe like 10 lines of code, you have an agent that can buy anything on the internet that supports X42.

So one time one time integration costs forever the benefit of open access.

So what's the upside here?

Well, one, we unlock the monetization of content. We unlock paper APIs. We unlock

content. We unlock paper APIs. We unlock

microp payments and we we have money now that is natively digital and exists as a first class citizen of the internet.

Really, I think the the value is we get more useful AI. If your AI has more capabilities, it can do more things for you.

If we don't have these open standards for transferring value, we're going to end up with a walled garden where each platform, each vendor has kind of a competing standard that isn't

interoperable. Or if you're using an

interoperable. Or if you're using an agent from one company, it can't speak to an agent from another company. If

you're using an agent that can only pay with one company's standard, it can't transfer value to another company's standard. And you end up with a

standard. And you end up with a fragmentation of the open internet. And

so my belief is that in order to keep the internet alive and to keep much of the web free, we need to create these open standards that make that actually

possible in the agentic age.

So what's next?

Well, for Coinbase, our mission is to increase economic freedom in the world.

We think if money can move globally 247 with no access, no limit to participation, you'll get a better global financial system.

And so if you want to learn more, you can go to x42.org.

You can check out the open source repo.

All the code uh for x42 as long as the standards are fully open source and licensed under Apache 2.0. Uh, and then if you want to learn learn more about my day job and Coinbase developer platform,

go to cdp.coinbase.com.

Thank you so much.

[Applause] [Music] Thank you, Eric. Thank you so much for sharing with

Eric. Thank you so much for sharing with us what you're doing at Coinbase.

There's a critical point that both Eric and Axel made in their presentations today that I want to re Reiterate as we close to realize their potential, new

technologies require new ways of thinking. We can't build tomorrow's

thinking. We can't build tomorrow's systems and applications on yesterday's technology. Nor can we build them on

technology. Nor can we build them on yesterday's dogma.

We have the opportunity to open our minds to the fact that we are going to build and consume financial services in a dramatically different way. And if

we're willing to lean into change, our financial lives will feature more choice, greater access to capital and opportunity, and better protection from risks to our businesses, our families,

and our prosperity as a society. So,

thank you to Axel and Eric for sharing their perspectives today. And thank you all for joining us on this passage to the next frontier in financial services.

Have a great reinvent.

[Applause]