Read time 5 minutes
Agents that let you act while others are still analyzing!
Why it matters:
This week's stories explore what happens to product advice once a personal AI assistant sits between a shopper and the store. One reports Shopify opening a catalog of more than a billion listings to any AI agent a shopper connects, with checkout handled inside the conversation. The other follows a journalist who asked Meta's Muse assistant to pick a robot vacuum and then traced who was included in the answer, who was missing, and who got paid. Read together, they suggest an assistant's recommendation is shaped by more than the quality of the product or the research behind it. It also reflects which stores the assistant can reach and complete a purchase in.
Key takeaways:
Mia Sato, The Verge, 10/13/2026
Shopify is widening its agentic commerce push beyond the Meta partnership it announced in September, which let Muse browse and buy on a shopper's behalf. Now shoppers can use any agent without a partnership integration: they connect it to their Shop account, and it can reach more than a billion listings, including items sold outside Shopify storefronts, with one-click Shop Pay checkout inside the conversation. Shopify markets the feature as a personal shopper that follows the shopper's rules, such as a budget cap or fast shipping only, and the buyer still has to approve every purchase. Orders flow back into the Shop app.
The Verge places the move in a broader contest. Many agent demos end in a transaction, and Google and Amazon are both investing in AI-driven shopping, while Amazon is in an ongoing dispute with Perplexity and, more recently, Muse over whether agents should shop for customers at all. Shopify's approach is to let any agent connect and keep the purchase running through its own catalog and checkout. For merchants on the platform, that means reach into new channels without separate integrations. For everyone else in retail, it makes catalog data and checkout access a form of distribution, because an agent can only buy what it can read and pay for.
Pete Pachal, Fast Company, 10/13/2026
Pachal asked his Muse assistant which robot vacuum to buy and got just a couple of models with purchase links. When he pressed it on how it chose, Muse said it had leaned on the testing site Vacuum Wars, alongside a thin repackaging of that site's data and an AI-generated roundup on an aromatherapy blog. Wirecutter, PCMag, and Consumer Reports were absent, which he attributes to likely publisher blocking but cannot confirm, and Amazon was absent because it blocked Muse from its site. That matters because Vacuum Wars earns its revenue from affiliate links that mostly point to Amazon, so the research Muse relied on was not paid for it. Meta, meanwhile, has said it will take a small fee on transactions in the app.
The pattern that stands out is how access can steer the answer. The models Muse recommended differed from the ones Vacuum Wars itself ranked first, and every one of them had an active checkout page. Pachal's point is that commercial considerations can enter a recommendation without anyone buying an ad. For reviewers, the old model needed one link and the new one may need a deal for every pairing of assistant and store. He is careful to note this was one test in one app in a category where standards are still being written.
Shopify's announcement shows the supply side opening up, with any agent able to reach a very large catalog and pay through one checkout. Pachal's test shows what that looks like from the receiving end: a short list of products that reflects which retailers an agent can transact with, built on research that may not be compensated.
For marketing teams, two questions follow. Is the product data an agent reads accurate and complete enough to be purchasable inside the assistants where shoppers are starting to ask? And are the third-party sources that make the brand credible still being funded in a world where the click no longer reaches them? Both will influence what an assistant says, long before a campaign gets a chance to.
Check back next week for another roundup of Media That Matters.
For more insights on Optimove and our Positionless Marketing Platform, contact us.
AI built for marketers.
Predict, create, and optimize — without waiting on data, dev, or design.
Writers in the Optimove Team include marketing, R&D, product, data science, customer success, and technology experts who were instrumental in the creation of Positionless Marketing, a movement enabling marketers to do anything, and be everything.
Optimove’s leaders’ diverse expertise and real-world experience provide expert commentary and insight into proven and leading-edge marketing practices and trends.
