Is decentralized AI identity the foundation for a $1.5 trillion agentic economy?

AI chatbots are growing in popularity, with $57 billion in expected global revenue until 2027, UK-based Juniper Research says. This type of conversational AI has become an alternative to search engines.

AI developers have understood this trend and are focusing on capitalizing on this new opportunity. A similar development happened before: when search engines entered mainstream, we saw new services arise such as Search Engine Optimization (SEO), advertising on search result pages, and content marketing. Similarly, using AI for exploring the Internet is expected to generate new commercial opportunities.

Streamlining the online shopping experience is one of these opportunities, where application programming interfaces (APIs) allow AI agents to search, browse, and buy items across multiple stores without clicking a button to navigate there. This process is called agentic commerce, “involving AI agents in making transactions,” Juniper Research points out, and as the enabler of verified agentic commerce we are aware of it.

Our CEO Evin McMullen shared recently:

In an AI-enabled world, we need to move from highly effortful manual processes toward experiences that are shaped around our identity and our intent.We shouldn’t need to manually click a bunch of buttons and physically hunt through pages of Google search results to arrive at an outcome or a series of actions that we’re trying to accomplish. We can enable our armies of digital butlers and multi-agent workflows to prove they’re acting on behalf of us, like me, Evin, an American citizen over the age of 18 who’s passed KYC (Know Your Customer).That means my agent can interact on my behalf in compliant spaces or spaces that are regulated by being able to prove that I’m a safe actor to engage with. I should be able to express my intent, my goals, my objectives, and my army of digital butlers would be able to interpret and understand my objective, and then help me execute it with the greatest efficiency and the least amount of effort and time.Instead of manually surfing the web and making decisions one set of data at a time, we can instead scale ourselves, our search capacity, the amount of data that can be interpreted to be able to execute against our intent and our objectives.

Juniper Research expects agentic commerce transaction value to grow from $8 billion in 2026 to $1.5 trillion by 2030, noting that pilots were only deployed in 2025 and 2026. Despite strong predicted growth, their report Agentic Commerce Market Report 2026-2031 says, “trust will remain the number one barrier to agentic commerce deployment.” Sam Smith from Juniper Research emphasizes that “as AI agent use and trust grow, agentic commerce will develop into an important access channel.”

Trust is the #1 barrier for mass adoption of agentic commerce (Juniper Research)

Despite the potential for agentic commerce to introduce enhanced authentication and verification systems, “trust remains a significant barrier to mass adoption” because trust in AI technology can mean many things:

• you give AI decision-making authority
• you grant AI access to sensitive personal and financial information
• you trust AI to keep shared data and financial details secure
• you allow AI to transact directly from your bank account
• you trust AI to handle privacy properly and legally safe
• you trust AI as if it was your friend, or do you?

You don’t wake up in the morning and think, “I want to learn more about trust in the age of AI.” Obviously, you don’t. You trust AI because you know what it can do for you:

• buy new tech gadgets (Canon camera, iPhone, MacBook, Blue Yeti microphone)
• remind you to dress professionally in a Google meeting
• recommendations for a new shirt and pants
• analysis of what you spend your money on
• special deals and discounts on Amazon
• services you need when you move to a new city or change apartments
• saving your personal preferences (ethical lifestyle, less alcohol, regular workouts, etc.)

This assumes you can trust AI agents and those involved providing the agents. It also assumes that AI is fact-checking everything and not inventing information, as it happened with the Air Canada chatbot misinformation case.

Our CEO remembers:

Air Canada incorporated a ChatGPT wrapper, basically a little customer service pop-up window enabled by AI. And as you know, AI is wanting to do it and it began hallucinating plane tickets that did not exist and plane fares that were not sanctioned by the company. This altercation between travelers who thought they were purchasing legitimate tickets because they did so on the airline’s own website, and then the airline itself pointing the finger downstream at the provider of that AI resource ended up in court. The outcome of that case was that Air Canada was in fact responsible for the actions of that agentically, AI-enabled experience. What has come out of that is the requirement of accountability for agents - that the last party or the last brand to touch and present agentic information is likely responsible for that output.

Earlier this year, Evin asked SEC Commissioner Hester Peirce how she thinks about AI agents acting on behalf of individuals or entities. In her personal capacity not the explicit view of the SEC, she said that agents can be considered “tantamount to subcontractors.” They are acting on behalf of another entity that is then accountable for their actions.

But of course, accountability requires identity. “There is no neck to choke if the neck doesn’t belong to a person or an organization with a name, a mailing address, and an ability to contact them,” Evin says. The concept of headless, unaccountable AI begs the question, “What happens when bots do bad things?” Whether it is topping up your wallet and payments, diverting from policy, or transgressing legal boundaries... accountability in the agentic space is vital.

Identity is the missing layer in agentic payments

Trust has to be attached to someone. While much of the discussion around agents has focused on which models win or how fast they can shop, the underlying question comes first: who is this agent, who is the human behind it, and who are they accountable to.

Evin makes this point in a chat with Brad Keoun and January Jones from DeAI News Journal, in the context of stablecoin-based agent payments, where the choice of currency is already becoming a question of compliance rather than convenience:

The only way to lead to a coherent outcome, a coherent experience with many stablecoins in the market, is if we have a clear interoperable identity solution to determine who’s an agent, who’s a human, and who they’re accountable to.

The current ecosystem is fragmented across issuers, infrastructures, and messaging protocols, and lacking agentic accountability, which is why we need an interoperable, accountability-powered identity infrastructure.

Stablecoins are “machine money,” and machine money needs accountability

Agentic commerce will not run on card rails (the networks and systems that facilitate card payments, such as Visa, Mastercard, and American Express.) It will run on stablecoins, which Evin describes as “machine money” because they are uniquely well-suited to the capabilities and requirements of agentic interaction.

But that money is already concentrated in ways that complicate accountability. Evin points to Coinbase, which holds positions across multiple layers of the stack at once: the stablecoin USDC, the payment protocol x402, and the ERC-8004 identity protocol:

It doesn’t surprise me that the house that built the agent stack, at least a popular one in the crypto space, is also responsible for the currency of choice of those agents.

Different stablecoins also already behave differently when something goes wrong. Evin notes that Tether, the issuer of USDT, and Circle, the issuer of USDC, have quite an early lead but diverge “in the way they handle things like freezing tokens in instances of illicit activity.” One stablecoin’s compliance policy becomes a question of who can act, when, and on whose authority - and none of those answers exist without identity.

Evin’s conclusion: with stablecoins issued on different infrastructure, different backends, and different messaging protocols, “the only way this all works together is to take care of identity such that we can identify agents, who they’re accountable to, the human beings and organizations they’re associated with, and then make that interoperable across different stacks.”

Accountability needs identity, and identity needs a human neck

Both the Air Canada case and the SEC Commissioner Pierce point in the same direction: you cannot hold a hallucinating chatbot responsible, you hold whoever deployed it responsible. The outcome of Air Canada, as Evin explains, was that “the last party or the last brand to touch and present agentic information is likely responsible for that output.” And Hester Peirce’s agrees: an agent is only ever acting on behalf of someone, and that someone is accountable for its actions.

The technology to transact is here. The legal frameworks for accountability are emerging. What is missing is the connective layer between them: a way to prove that a given agent acts on behalf of a specific, verified human or organization.

What happens next

Your AI agent can already buy groceries, data top-ups, pizza, sneakers, flights, tickets, and gifts... larger sums and real legal accountability are next.

When agents act, someone has to be able to answer: who is this agent, who is the human behind it, and who answers when it goes wrong.

That is the commerce layer we are building. One where your agent shops and buys for you - within the limits you set, carrying your verified identity behind every purchase, so checkout goes through while anonymous bots get blocked.

Prove you are one of the humans behind the wheel. Tap your passport to your phone, bring your own agent, and see what verified agentic commerce feels like.

Join the waitlist at billions.network/commerce