AI agents need structured, current product data and direct transaction APIs instead of web experiences built for human browsing.
2
Ionic Commerce combines merchant product feeds with reviews and other context, then exposes the data in agent-friendly formats through an API.
3
Merchants can pay for agent-driven discovery because they gain insight into what models saw, how consumers made decisions, and where transactions came from.
Summary
Justin introduces Ionic Commerce's plan to help AI agents participate in the economy, starting with ecommerce. He argues that the web was built around advertising and human clicks, while agents need to make decisions and complete actions in one place. Static product ads do not reliably contain stock, shipping, or the detailed attributes an agent may need. Ionic works with merchants to enrich product feeds with reviews and other information, then makes that data available through agent-friendly databases and an API. Its transaction layer lets an agent complete checkout with a payment token after helping a consumer choose a product. Justin also describes a business model in which merchants pay for agent-driven customer acquisition and receive information about the path from product data to purchase. He says Ionic already has hundreds of merchant relationships and a catalog with millions of SKUs.
The web was built for ads and human clicks, while agents need to take actions
Justin starts with tasks people may want agents to handle, such as ordering an Uber to the airport, finding a Mother's Day gift, or booking a last-minute hotel. Building these workflows exposed a basic mismatch. Product pages and ads assume that a human will click through several sites, learn more, and eventually transact. An agent should be able to act on the user's behalf without bouncing between sites. The system therefore needs to be designed around decisions and actions rather than around showing an ad to a person.
Agents need product attributes that ordinary feeds do not contain
Justin gives a clothing example to show the difference between human-friendly and agent-useful information. An agent may need to know that the user wears a shirt measured 21 inches from top to bottom, rather than simply an extra-large size. It may also need a precise fabric mix, such as 7% modal and a particular percentage of cotton. Existing ad feeds may not even say whether an item is in stock or how it ships. They are static, so they cannot support the current, detailed decisions an agent has to make.
Ionic enriches merchant feeds and stores the result in formats agents can use
Ionic works with hundreds of merchants that provide access to their product feeds. It uses agentic workflows to add editorial product reviews and other information that could help an agent decide. The resulting data can live in a relational database, vector store, or graph database, depending on how agents should access it. Justin says Ionic will continue working on those access patterns with the wider ecosystem. The API is already live and available for people in the audience to use.
The decision and the transaction should happen in the same layer
After an agent has enough information to choose a product, Ionic wants the purchase to happen without sending the user to another site. Justin describes an example in which an agent helps find headphones for a flight later that day. Once the decision is made, the user can say they want to check out, and a payment token can pass through to complete the transaction immediately. In his view, an agent should be able to carry out the whole task in one simple step.
Merchant payments can support consumer AI products better than subscriptions alone
Justin questions whether consumer AI products can charge every user a $20 monthly subscription forever. He says that model may work for early adopters but will not reach the mass market. His proposed source of revenue is the merchant. Merchants want help finding the right consumer and matching that person with the right product. Ionic is starting in ecommerce, with hotel rooms, flights, and other categories as later possibilities. The payment comes from helping create a useful customer connection, rather than only from charging the consumer.
Agent commerce should give merchants visibility into the path to a purchase
Justin contrasts an agent buying through Amazon with a transaction that keeps the merchant involved. In the Amazon example, Amazon receives the payment and takes the credit. Ionic wants merchants to know what product data went into the world, what models saw it, how they used it, and how the transaction happened. Merchants can then understand the continuing relationship with the customer. Justin says they would prefer to pay the party that provides this connection, and says Amazon may also pay, though less.
Ionic already offers a merchant network and an AI-native catalog
To begin this work, Ionic has formed relationships with hundreds of merchants. Justin says people can work with those merchants immediately. The company also has a catalog containing millions of SKUs, which he describes as AI native and still growing. He directs interested developers to docs.ioniccommerce.com and gives Justin or Owen at Ionic Commerce as contacts for people who want to work with the company.
"The secret is that merchants will pay everyone in this room to help them find the right consumer and match them to the best products."03:38
Who should watch
You are building an AI agent that needs to search products and complete purchases without relying on human-oriented web pages.
Your commerce product has merchant feeds but lacks current inventory, shipping details, reviews, or structured product attributes.
You want to understand how agent-driven customer acquisition could create a merchant-funded business model.
