ByCoinGape
Agentic Commerce: If there is one trend quietly reshaping the crypto and payments industries, it is agentic commerce — the rise of autonomous software agents that can transact, pay, and interact on behalf of users.
What began as experimentation is quickly turning into real activity. New standards like x402 are seeing rapid adoption, with transaction volumes scaling from millions to tens of millions in just a few months. The trajectory is clear: agents are no longer theoretical. They are participating in live economic systems.
But the most important shift is not just scale. It’s interface.
This momentum is only accelerating, according to Paolo D’Amico, Senior Staff Engineer at Tools for Humanity – the technology company building for humans in the age of AI, currently focused on World – which recently worked with with Coinbase to launch AgentKit. Through its integration with payment infrastructure like Coinbase’s x402, World’s toolkit enables platforms to cryptographically verify whether an AI agent is acting on behalf of a unique human, helping distinguish legitimate activity from coordinated bot activity.
Paolo explains, “For World, we saw it as a natural next step of what the project was already doing with proof of human. With agents now participating in e-commerce and overall online interactions, the internet needs a way to reason about uniqueness, fairness and access at the human level.”
Also Read: VARA Counsel on The Next Phase of Digital Finance
From Mastercard and Visa to Coinbase, Stripe, and Polygon, there is hardly a major company today that isn’t working on agentic payments infrastructure.
The scale is already visible: 40,000 on-chain AI agents are executing transactions. Protocols like x402 have processed more than 100 million payments in just a few months.
Today, agentic commerce is largely concentrated around digital services. This includes API access, compute, data, and SaaS tools rather than physical goods.
“I believe in the near-term, dinner reservations, concert tickets, travel bookings and other online shopping activities will get increasingly executed through agents,” D’Amico says. “Agents will become a general purpose interface for everyday computing tasks.”
At the same time, he is clear that World’s AgentKit is just the beginning.
“Much more than a particular use case, the overall shift is extremely interesting to me. It’s not only about supporting agentic interactions, but about anchoring those interactions to human intent. If I am a restaurant owner, I want to make it as easy as possible for guests to book reservations, but I want to prevent hoarding, re-selling and abuse overall.
Beyond AgentKit, we are looking into broader improvements focused on agents at the protocol level, in the next generation of World ID. For example, I am passionate about discovering the right way to empower agents to demonstrate human intent responsibly and securely.
All of this being said, I am personally very excited about my agent booking my travel reservations. I could spend an inordinate amount of time optimizing travel.
Also read: Coinbase x402 Developers on Agentic Commerce
Interestingly, many industry leaders, including Jeremy Allaire of Circle, are starting to ask what happens to human transactions as agents take over. Some believe agent-driven activity could soon rival, or even surpass, human volumes in certain categories.
That shift is already visible. AI agents have executed over 140 million transactions, with an average value of just $0.31 per payment, showing how deeply microtransactions currently dominate this space.
Paolo D’Amico agrees this shift is real:
“Yes, I think that’s possible, particularly in some categories… Where a human before could only perform a number of these tasks sequentially, they’ll now be able to deploy multiple agents to perform them.”
But he also flags what could slow things down:
“The limiting factors will be primarily around the infrastructure… We will need infrastructure that supports payments, permissions, policy enforcement… If the infrastructure doesn’t mature fast enough, platforms will start blocking agents.”
Also Read: Crypto Market Report Q1
As AI agents become more widely used, one clear pattern is emerging: most agentic payments today are small, low-risk transactions. This includes micropayments for APIs, compute, and digital services — an early but important phase in adoption.
AI agents are starting to tap into a $46 trillion stablecoin payment flow, though they’re still a small part of it.
But that’s not a limitation. It’s a phase, according to Paolo.
Asked whether agentic payments would ever handle high-value transactions, he says:
“I think it will take some time, but I do see that happening. Micro-payments are a natural first step for payments – they’re a great way to drive adoption of standards, and to overall grow support across the internet from platforms, merchants and agent tools.
As adoption increases, and particularly other parts of the agentic stack develop, this will become feasible. For example, it’s reasonable to expect a higher degree of trust to perform high-value transactions. This is something where factors like identity come into play. It’s analogous to how high-value transactions occur in the offline world today, there are different requirements and different risk modeling for high-value transactions.”
To close, the trajectory from here is becoming clearer. Paolo D’Amico sees three shifts defining the rest of the year:
“First, I think we’ll break out of the demo and proof-of-concept stage to real workflows… and move into production use.”
In other words, this stops being experimental and starts becoming infrastructure.
“Second, I think we’ll see significant developments in the infrastructure realm: new standards, adaptations of existing systems – and broader adoption across internet tooling.”
The rails are still being built, but they’re catching up quickly. And finally, the shift that may matter the most:
“The conversation will shift from capability… to whether an agent should do it, or under which conditions.”
By the end of the year, agentic commerce may begin to look less like a novelty and more like a foundational layer of how users interact with digital systems.
Sneha Agrawal
|
Newsletter
Subscribe to our newsletter for new updates
By signing-up you agree to our Terms & Conditions and Privacy Policy.
Share