900 Million Weekly Users Is Not the Hard Part

ChatGPT may have 900 million weekly users.

But that does not mean agentic commerce is solved.

The hard part is not just letting an agent recommend or even pay.

It is answering the real question:

who carries responsibility when an agent spends money?

That is where the real battle is happening now.

The six layers no one talks about

We have written about the shift from seller-controlled commerce to buyer-side agents before. But there is a deeper layer most discussions skip.

Agentic commerce is not only about discovery, recommendations, or checkout. It is about who owns the outcome when something goes wrong.

The infrastructure that matters includes at least six layers most product conversations ignore:

• Authorization — who gave the agent permission to act, and what were the boundaries?
• Payment rails — can money move safely between a buyer's agent and a seller?
• Identity — how do you know the agent represents a real buyer with real intent?
• Governance — what rules apply when the agent operates across borders, platforms, and jurisdictions?
• Liability — if the agent makes a bad purchase, who is accountable?
• Recourse — how does the buyer or seller recover when the transaction fails?

These are not edge cases. They are the difference between a demo and a market.

Why services make this harder

Products are relatively simple. A SKU has a price, a description, and a return policy.

Services are different. Scope is variable. Timing matters. Availability changes. Price may depend on context. The right provider depends on skill, location, urgency, and constraints. Exceptions are common.

That means agentic commerce in services is not just about connecting supply and demand. It is about building trust, structure, and accountability into every transaction.

An agent that can book a flight is useful. An agent that can hire a plumber, handle a quote, manage scheduling, and stand behind the outcome is something else entirely.

What this means for Aune

Aune is not building a chatbot layer. We are building the transaction infrastructure that makes agent-driven service commerce actually work.

That means:

• structured provider data that agents can understand and compare
• clear constraints and availability so agents know what is possible
• executable booking paths that turn intent into real appointments
• trust signals and fallback logic so buyers and sellers both feel protected

The future of services is not a better search engine. It is a system where agents can understand, compare, book, and take responsibility for real-world outcomes.

Scale is exciting. Infrastructure is what makes it real.

Read more

We wrote about what businesses need to do now as buyer-side agents become part of discovery and commerce:

[The Buyer's Agent Is Coming. Here's What Businesses Need to Do Now.](/blog/buyers-agent-is-coming-what-businesses-need-to-do-now)