Some people like shopping and others don't. Some people are browsers and some people are buyers. And most people are both depending on what they are shopping for.
Browsers like scrolling down clothing store websites almost like people scroll down their social media. Buyers say "ugh I gotta buy a shirt" and buy one of the first things they see. The same person can be a browser for clothes and a buyer for batteries.
So what does all of this mean for agentic commerce for individuals? Well it means it might meaningfully diverge. Buyers are likely going to adopt agentic commerce and let agents shop for them first because they don't want to shop in the first place. The interface can largely stay as it is today as long as it gets the job done. From the finding of items to buy to checkout, less improvement is needed.
However, browsers like to shop and scroll down websites to find things to buy. Maybe even when they have nothing in mind, or they end up buying more than they went to the site for. That is a big part of why Amazon has fought outside agents shopping on its site. It sued Perplexity over its shopping agent and has blocked ChatGPT from shopping on Amazon. Amazon made $68.6 billion from advertising in 2025, and agents don't get sold to as easily as people do. Amazon isn't against agents though. It's against other companies' agents. It has its own shopping assistant, Rufus.
I think the near future of agentic commerce for browsers will look like a traditional website but generated by the agent for that shopping session. The user will describe what they are interested in looking at and the agent will pull together different items from different websites and it will look like one website. Each item they click will have pictures, the description, the sizes, and all of what people can see when they currently click on an item on the store's website. Then, if they want to buy something, they will add it to the cart and let the agent finish the rest.
So I think we will see a divergence between browsers and buyers in their uses of agentic commerce products, and I think AI agents will get to understand what each person wants and show it the way they would like.
However, this is very different from how businesses buy things. Right now, for smaller purchases, items get put on the company card (whether it's an employee's company card or a card for specific purchases). Even for small purchases like buying pens for the office, the business might have a preferred vendor and a spending policy, and someone still has to find the item, check out, and make sure it gets expensed correctly. Tools like Ramp and Brex have made the expense report part much easier, but the buying itself is still done by a person.
Buying for businesses is a chore for everyone. Nobody browses for printer ink. Business buying doesn't diverge the way that consumer buying does. It's almost entirely buyer mode. So that is why I think the real use cases for agentic commerce are going to be for businesses making purchases. An agent does not mind looking through a list of preferred vendors, shopping on the website for them, adding the shipping and billing info, staying within the company's spending policy, and logging the purchase.
So that is why I believe agentic commerce for businesses will likely be the vast majority of agentic commerce in the future. Not that agentic commerce for individuals won't happen or be big, but the design considerations for browsers and the difference between browsers and buyers will make it harder. Plus, online stores like Amazon that make a significant amount of money on advertising products on their website and the add-on buys people make are an additional headwind.