Commerce is entering a major transition. For the last 20 years, product discovery has relied on traditional search: customers visit a brand’s website or Google a product. Search engines show online stores and product cards, and the shopper makes a quick purchase decision.
AI agents completely change this pattern.
Instead of typing keywords into Google, consumers will increasingly talk to an AI agent and describe the product they want. The agent then searches across websites, marketplaces, local retailers, and recommerce channels.
Rather than browsing product pages, the consumer receives curated options.
This changes e-commerce from an active, time-boxed process (“I have 15 minutes to buy jeans”) into a passive one (“Find me jeans in this size and price range whenever they appear”).
AI agents will continuously scan for the right product across multiple channels. Consumers will set conditions such as:
• preferred price range
• preferred distance from home
• preferred sellers (e.g., small shops instead of big retailers)
• condition requirements for recommerce
• willingness to wait days or weeks
This makes long-tail and secondhand items dramatically easier to discover.
Customers will increasingly ask whether recommendations are:
• based on their preferences
• based on the agent’s model incentives
• influenced by paid placement
Different AI agents will emerge — some optimized for transparency and trust, others optimized for convenience and speed.
For recommerce, the shift is even bigger. Unique, one-of-a-kind items benefit enormously from continuous AI-driven discovery because:
• every item has unique attributes
• condition varies
• images matter
• listings appear and disappear quickly
AI agents can scan multiple recommerce platforms, compare individual items, evaluate images using vision models, and present the best matches.
This turns the AI agent into a true personal shopper — not just a search engine.
Tuomo: So how is e-commerce or commerce in general going to look like in a couple of years as AI and MCPs and all of these new technologies are swimming into the world of commerce. It's probably easiest to start thinking from product discovery. How do people actually discover products? Now I think the product discovery route might be that if you need a product you might already go direct to the brand's website or then you just simply Google it. And if you think about Googling stuff, you usually have websites as results, the online stores, but you also see those product cards there on the top.
Tuomo: Now AI-driven discovery is: instead of going to Google, you might go to ChatGPT. So you go to a chat client where you have a discussion around “I want to buy an iPhone, where might I find it?” or you have a more specific need like in recommerce: “I'm looking for an old Jimi Hendrix vinyl from this era in this condition.” What happens there is that instead of you searching for an endpoint like a specific website, you are hiring an AI agent to do the search for you. I think that's a big change in the product discovery process. You will have a discussion with an AI agent that then goes out and tries to find that product for you from different endpoints like websites and so on.
Tuomo: Maybe a thing that relates to this is your purchasing preferences outside of price. It's not necessarily always about a specific price. We've gotten used to instant e-commerce purchasing where you designate 15 minutes to buy your jeans online, find them quickly, and want them shipped immediately. I think what AI agents might open up is more passive search. You hire the agent and say you're looking for these jeans in this price range, and you're willing to allow the agent to scan and crawl for two or three weeks to see if anything pops up in a recommerce marketplace. And you can give soft preferences like “I'm willing to go to a store if it's within 10 miles” or “I prefer a mom-and-pop shop over a big retailer.”
Karri: What about transparency? I think that’s something that worries a lot of people. If the AI agent says this is the best place to buy it, is it the best place for you as a customer, or is it the best place for the AI agent?
Tuomo: Whatever agent you choose probably depends on your purchasing criteria. It sounds like you would like to use a purchasing agent that is extremely transparent, whereas someone else just wants the job done. They’re less sensitive to some aspects. It’s a preference thing. For some people it’s more mass buying, for some it’s specialized buying. That's probably why there will be different providers that focus on different aspects of the purchasing experience.