Muse vs Instinct: what personal shopping agents mean for retailers

Muse and Instinct are personal AI agents that shop on a person's behalf. Muse is Meta's agent, launched in September 2026 in its own app and WhatsApp. Instinct is a startup's agent reached by text message or WhatsApp. Both can search, compare and buy, so your products now have to make the agent's shortlist before a shopper ever sees them.

Two very different companies launched these agents within weeks of each other, and both focused on shopping first. Below, we compare how each one finds products and checks out, and what that means for your product data.

What is a personal shopping agent?

A personal agent does things on your behalf. A chatbot gives you information and links. A personal agent can remember context about you, use accounts you connect, and finish the job: search, compare, then buy once you say yes.

For shopping, that means the agent reads the product page before the shopper does. The agent takes a request like “a winter coat for my sister, warm but not bulky, under $250”, goes looking, and returns a handful of options. The shopper may never see the rest of the catalog.

Muse vs Instinct at a glance

Sources: Meta's Muse launch post (8 September 2026), Connect 2026 recap and Q3 2026 earnings call (21 October 2026); TechCrunch reporting on Instinct (24 August, 28 September and 10 October 2026); posts by Instinct's founder and Shopify on X (28 September 2026).

How does Muse find products?

Meta hasn't said which sources Muse reads. Shopify's Help Center documents one route: Shopify merchants' products reach Meta AI and Muse through Shopify Catalog. From 20 November, Meta's revised Commercial Terms also let it crawl the product pages a business's catalog links to. Meta had planned 30 October and moved the date, a spokesperson said on 23 October.

So you need the same product facts in the Meta catalog, in Shopify Catalog and on the product page. Our guide to how Meta Muse finds and picks products breaks this into three stages.

How does Instinct find products?

Instinct's founder describes the shopping flow as searching Shopify's merchant network for live price, size and stock, then preparing the order for Shop Pay checkout. He also lists price-drop alerts, back-in-stock alerts and order updates. Instinct's own site says the agent uses a phone and computer the way people do, which suggests it can also work outside Shopify.

For a Shopify retailer, that means Instinct is shopping straight from your variant, size and inventory data. If a size shows in stock after it has sold out, the shopper ends up with a failed order in their text thread.

Scale is harder to judge. Instinct's founder said on a podcast that the platform is approaching a billion dollars in annual transactions, more than half of them travel. TechCrunch pointed out that he didn't explain how that number was measured, and the company hasn't shared user numbers.

What do the two agents have in common?

The two agents are different, but they need the same four things from a retailer:

It all comes back to the product record. An agent can only recommend what your data lets it understand, which is why the same enrichment work helps with every agent. Our guide to product data enrichment covers how that record gets built at catalog scale.

Where do they differ for retailers?

Which other shopping agents should retailers watch?

Muse and Instinct are two of several agents that now take a shopper from request to purchase.

• Google AI Mode and Gemini. Google announced checkout in AI Mode and the Gemini app on 11 January 2026, built on the Universal Commerce Protocol it co-developed with Shopify, Etsy, Wayfair, Target and Walmart.
• ChatGPT. OpenAI scaled back Instant Checkout as a stand-alone feature in March 2026 to focus on discovery, TechCrunch reported, so merchants check out shoppers on their own store.
• Amazon. Alexa for Shopping includes Buy for Me, which can purchase from other retailers' sites for US customers.
• Perplexity. Instant Buy is available to US users, with the order and payment going to the merchant.

Each one reads product data differently, which our explainer on agentic commerce covers in more depth.

What should retailers do now?

1. 1.Build one product record with request-ready facts: recipient, occasion, use case, material, fit and season.
2. 2.Publish the same facts to every surface an agent reads: the Meta catalog, Shopify Catalog and your product pages.
3. 3.Keep live price, size and stock accurate across all of them, especially through holiday promotions.
4. 4.Make checkout wallet-ready, and test it as a guest.
5. 5.Measure each agent on its own. One agent's results say little about another's.

What we don't know yet

• Neither Meta nor Instinct has published how its agent ranks products.
• Instinct's transaction and usage figures come from its founder and are not audited.
• What is known about Muse's sources comes from Shopify's documentation and Meta's terms, not from Meta describing what Muse reads.
• Both products are changing weekly. Check the facts above against current documentation before acting on them.

How Lily helps

We treat each personal agent as its own channel. Lily Max builds one product intelligence layer with the facts every agent needs, and publishes it wherever each agent reads. Then we measure each agent separately, because results on one don't predict the next.

Book a demo to see how your products read to the agents your shoppers are starting to use.