Artificial Intelligence and the invisible future of online shopping
Artificial Intelligence has already transformed the way we search for products, compare prices, and complete purchases online. That is no longer news to anyone. However, while the entire market talks about AI agents shopping independently, almost nobody stops to think about what needs to exist behind the scenes for all of this to actually work without glitches and without headaches.
Because let’s be real: one thing is a virtual assistant recommending the best pair of headphones based on your preferences and budget. A completely different thing is that same assistant accessing your card, navigating an online store, picking the right product, and completing the payment without you having to lift a finger 🤖💳. That is exactly where the conversation gets more interesting and more complex than it seems at first glance.
A New York-based startup called Circuit & Chisel is betting big on the idea that the future of online commerce depends on an invisible layer of infrastructure. We are talking about a protocol capable of allowing Artificial Intelligence agents to make autonomous payments without friction, without errors, and without relying on makeshift solutions. Founded by two former Stripe executives, the company raised 19.2 million dollars in funding and has already caught the attention of major names like Stripe itself, Polygon Labs, and Samsung Next.
The problem nobody was solving
Today, when you make an online purchase, the entire process was designed for humans. You see buttons, fill out forms, type in your card number, confirm your address, and click checkout. It seems simple, but behind all of that there are layers of verification, authentication, and transaction protocols that were built exclusively with people interacting with screens in mind. The problem is that AI agents do not interact with the world the same way. They do not click buttons, they do not read captchas, and they do not fill out form fields in the conventional way. And that is exactly where the bottleneck that Circuit & Chisel wants to solve lives.
Most attempts to automate purchases with Artificial Intelligence so far worked in a pretty shaky way. The agents basically simulated the behavior of a human user, virtually clicking on page elements, filling in fields, and hoping nothing changed in the site layout between one purchase and the next. Any minimal change in an online store’s design could break the entire flow. It was the digital equivalent of building a house on quicksand — it works for a while, but it is not sustainable long term.
What Circuit & Chisel is proposing is something fundamentally different. Instead of forcing AI to behave like a human inside systems made for humans, the company is creating native infrastructure so machines can communicate directly with payment platforms. That means transaction protocols specifically designed for autonomous agents, with dedicated authentication, configurable permissions, and audit trails that ensure security at every step of the process. It is a technology innovation that tackles the root of the problem instead of just slapping a band-aid on it.
Who is behind Circuit & Chisel
The company was co-founded by Louis Amira and David Noel-Romas, two professionals who built their careers inside Stripe, one of the largest digital payment infrastructure companies on the planet. Amira served as head of crypto and AI partnerships at Stripe, while Noel-Romas led the company’s crypto engineering. Both left Stripe last year to found Circuit & Chisel and launch ATXP, which stands for Agent Transaction Protocol.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
In an interview conducted in February 2025, Amira explained that the departure from Stripe happened on good terms. According to him, Stripe’s founders always showed conviction in bold ideas, even keeping the crypto team running after the collapse of the FTX exchange shook market confidence. However, the vision that Amira and Noel-Romas had for the future of autonomous payments was too ambitious to fit within the structure of a large company at that moment. When Stripe saw what the duo was building on their own, it decided to invest in the initiative. This type of relationship between former employees and their old company is not common, but it reinforces the project’s credibility and shows that those who deeply understand digital payments believe in this path.
Beyond Stripe, the list of investors includes Polygon Labs, a blockchain infrastructure company, and Samsung Next, the investment arm of Samsung Electronics. In a post published in September, Samsung Next stated that it believes payments are one of the most valuable layers of technology and that Circuit & Chisel has the right team, the right product, and the right timing to turn ATXP into the standard rails for autonomous software. When a giant like Samsung puts money and reputation behind a startup, the signal is clear: this is not a passing niche, it is a trend with the potential to become an essential part of global digital infrastructure.
How autonomous payments work in practice
To understand how this system works, imagine the following scenario: you set up an AI agent to monitor the price of a specific laptop you want to buy. You define a maximum budget, choose your trusted stores, and authorize the agent to act when it finds the best deal. With Circuit & Chisel’s infrastructure, that agent does not need to simulate your clicks or access the store as if it were you browsing through a web browser. Instead, it communicates directly with the store’s payment API through standardized transaction protocols, sends the payment credentials securely, and completes the purchase in seconds. All of this happens without friction, without filling errors, and without that awkward moment when the captcha freezes everything up.
Security, naturally, is one of the biggest concerns when we talk about autonomous payments made by machines. Nobody wants to give a robot a blank check to go spending money with zero oversight. That is why the protocol developed by the startup works with a granular permissions system. The user can set spending limits per transaction, restrict product categories, limit authorized stores, and even require human confirmation above a certain amount. It is like giving your Artificial Intelligence agent a controlled allowance — it has the freedom to act, but within well-defined rules that you set yourself.
Amira himself used a similar analogy during his interview. He described the ideal model as something like setting an amount the agent can spend freely — 100, 200, or 400 dollars — and letting it make decisions within that limit. For bigger or more strategic matters, human interaction would continue to exist, but at an increasingly higher and less granular level. In other words, you will not be reviewing every electric bill or every grocery run. You will focus on the decisions that truly matter while the agent handles the operational side of daily life.
Agents as actual customers
One of the most surprising points Amira shared in the interview concerns who exactly is using the ATXP protocol. Initially, as expected, it was humans who signed up on the platform and started testing the tool. However, something unexpected began happening: AI agents are signing up and installing ATXP on their own, without direct human intervention, in order to use the available tools and services.
This might sound like something out of a sci-fi movie, but in practice it is a logical consequence of how autonomous agents operate. When an AI agent encounters an obstacle, like the inability to complete a transaction on a given platform, it searches for available solutions. If ATXP solves that problem, the agent incorporates the protocol into its workflow. Amira admitted that a year ago, mentioning that AI agents would be his main customers in the future got him laughed at in some meetings. Today, that idea already sounds much more plausible to a growing number of people in the industry.
This shift in perspective is significant. We are moving toward a scenario where Artificial Intelligence is not just a tool used by humans, but an active participant in the digital ecosystem, with its own infrastructure, authentication, and service access needs. Circuit & Chisel is positioning itself to serve this emerging demand before it goes mainstream.
A market full of uncertainty and opportunity
Despite all the excitement, Amira makes a point of staying grounded. He openly acknowledged that when he left Stripe just over 12 months ago, he believed he had a clear vision of how the future of autonomous payments would unfold in the years ahead. Since then, that certainty has faded. The AI agent ecosystem is evolving so fast and in such diverse directions that even the major market players hesitate to make massive bets in any single direction.
According to Amira, Circuit & Chisel has already talked with roughly ten other companies occupying similar positions in the market, as well as strategic names like Coinbase, Solana, and several heavyweight investors. The general conclusion is that nobody is willing to go all-in on a single model right now, precisely because the landscape is extremely uncertain. That does not mean the market is standing still — quite the opposite. It means companies are testing, learning, and adjusting their strategies in real time, which is actually pretty healthy for a tech ecosystem in its early stages.
Circuit & Chisel seems comfortable with that uncertainty. The company has conviction about what it needs to execute in the coming months and is building infrastructure that can adapt as the market matures. In an environment where the only certainty is constant change, flexibility and speed of adaptation are worth more than a rigid five-year plan.
What this changes for buyers and sellers
On the consumer side, the promise is a shopping experience that requires less and less effort. Think about how you already use virtual assistants today to create shopping lists, compare prices, or get promo alerts. Now multiply that by ten. With autonomous payments actually working, your AI agent could handle automatic restocking of everyday items, take advantage of flash sales while you sleep, or even negotiate prices on platforms that accept this type of machine-to-machine interaction. Online commerce would stop being something you do and start being something that happens in your favor in the background, continuously and intelligently.
On the retailer and e-commerce platform side, adopting transaction protocols compatible with AI agents could mean a significant boost in conversions. Today, a huge percentage of shopping carts are abandoned because of checkout friction — lengthy forms, typos, authentication failures. When the buyer is a machine operating through a standardized protocol, those barriers simply cease to exist. The transaction happens cleanly and directly, which potentially reduces abandonment rates and speeds up the sales cycle.
On top of that, stores that adopt this type of infrastructure early could position themselves as preferred destinations for AI agents, which will tend to prioritize environments where they can operate more efficiently. It is an interesting dynamic: in the future, optimizing the shopping experience may mean not only thinking about the human user but also about the experience of the autonomous agent representing that user.
The future of interaction between humans and agents
Amira shared a fascinating vision of how the relationship between people and AI agents should evolve over time. According to him, with each passing year, agents will take on more responsibilities without needing to check in with the user. Recurring bills that show up every month, for example, would no longer need manual approval. The agent recognizes the pattern, knows it is an expected expense, and handles it on its own.
In the view of Circuit & Chisel’s CEO, the trend is for each person to interact primarily with a single main agent, perhaps split between personal and professional life. That central agent, in turn, would activate sub-agents or one-off processes as needed. Instead of micromanaging every decision, the human would interact with their agent at a strategic level, setting priorities and general guidelines rather than approving each transaction individually.
It is a paradigm shift that goes far beyond technology. It involves trust, control, and a new way of delegating financial decisions to autonomous systems. And if startups like Circuit & Chisel have their way, the infrastructure needed for this to happen safely and at scale is already being built right now.
Why this movement is worth keeping an eye on
The nearly 20 million dollar investment raised by Circuit & Chisel shows that the market is taking this vision seriously. The participation of Stripe, one of the world’s largest payment infrastructure companies, is not just a footnote. It signals that the financial sector’s own giants recognize that the next big wave of online commerce will be driven by autonomous agents and that the current infrastructure simply is not ready for it.
The ATXP protocol is still in its early stages, and the ecosystem as a whole is far from consolidated. But the signals are consistent. Major companies investing, AI agents already adopting the tool on their own, and a founding team that lived the challenges of digital payments from the inside, on the front lines of one of the biggest fintechs on the planet. We are looking at a technology innovation that is not just incremental but one that could redefine the relationship between consumers, machines, and the digital payments ecosystem as a whole 🚀.
In the coming months, following how Circuit & Chisel and similar companies evolve will be essential to understanding where online commerce powered by Artificial Intelligence is headed. The infrastructure being built today is what will determine whether the AI agents of the future become efficient assistants or just promises that never made it off the drawing board.
