Meta Muse and Instinct Are Showing Us What Comes After Ecommerce
Consumers are getting AI working on their side. Retailers need AI working on theirs.
For 25 years, e-commerce has been built around one assumption. The shopper is a person. That assumption is starting to break.
Meta’s Muse can browse the web and take actions on a user’s behalf, including negotiating and making purchases with their approval. Instinct approaches the same idea from another direction. It’s a personal agent designed to use apps and devices to get things done on your behalf.
I think we’re starting to see the ChatGPT moment for personal agents. AI isn’t just something consumers can talk to anymore. It’s starting to do things for them.
Shopping may be one of the most interesting places this plays out.
Michael Cantrell, our fearless sales leader at Spangle, recently used ChatGPT while buying a car. It helped him negotiate the purchase and get what he thought was a great deal. He came back incredibly excited about it.
I kept thinking about the person sitting on the other side of that transaction.
The dealership wasn’t just negotiating with Mike anymore. It was negotiating with Mike plus intelligence working entirely in Mike’s interest.
Today Mike had to use ChatGPT himself. Imagine when more of that becomes automatic and available to everyone.
Retailers have spent 25 years learning how to win the customer. Now they have to learn how to win the AI working for the customer.
The asymmetry
Imagine where this goes. A customer’s agent may eventually know their preferences, budget, purchase history, constraints, the alternatives they’ve considered, the offers available across the market, and what a good outcome looks like for that individual.
What’s sitting on the other side?
On the retailer’s side, we still have catalog feeds and collections of systems that often require a person to notice the opportunity and decide what to do about it.
The customer arrives with an agent that has reasoned through a decision. The retailer meets them with an assortment, a feed, and a team.
That’s the asymmetry.
Much of the conversation so far has focused on making retailers more legible to these agents. Better product data. Better enrichment. Better feeds. Better visibility. More pages.
All necessary. But being understood by the customer’s intelligence isn’t the same as having intelligence of your own. And visibility doesn’t mean you’ll be chosen.
Showing up is not winning. If customers have AI working on their side, retailers are going to need AI working on theirs.
What merchandising even means
Merchandising has traditionally meant arranging products, categories, and experiences for a human visiting your digital property.
But an agent doesn’t experience your site the way a person does. A hero image designed to inspire a human shopper may mean very little to an agent trying to satisfy a specific request.
Who is merchandising your business to their agents?
That’s a different problem than search or SEO or “AI visibility.”
It’s the retailer having a point of view about which customer demand it should be pursuing, which of its products belong in that pursuit, and what should change to make that possible.
That work exists today. It’s mostly done by people.
Commerce is becoming intelligent on both sides
I think the next phase of agentic commerce won’t just be about buyer agents. We’ll start to see intelligence emerge on both sides.
The customer’s side is arriving quickly. Muse and Instinct are two visible examples. There will be many more.
The retailer’s side is less developed.
If the customer has intelligence working to make the best decision for them, the retailer needs intelligence working to make the best decisions for the business.
That means understanding the business well enough to decide which opportunities matter, how to respond, and then take on more of the work.
I’ve spent much of my career building commerce systems, and one thing I’ve learned is that new technology often creates more work before we figure out how to make it remove work. I don’t want that to happen with AI.
AI should make retailers money, not make them more work.
At Spangle, that’s what we’re building toward. Intelligence on the retailer’s side that understands which customer opportunities the business can actually win, and that can take on more of the work required to win them, within the retailer’s rules and approvals.
It has to sell. That’s the only measurement that matters.
I have no idea whether Muse or Instinct will be the consumer agent that matters five years from now. That’s not really the point.
We’ve spent years asking what happens when consumers get powerful AI. It’s time to ask the other side of the question.
Consumers are getting AI working on their side. Retailers need AI working on theirs.
Originally published on LinkedIn: Maju Kuruvilla on what Meta’s Muse and Instinct are showing us.