Meta's Muse Just Changed Who Picks Your Business

Sep 24, 2026 — By Christopher Lynch

Personal agents in 2026 and 2027: what's real, what's hype, and what to do before your customers' agents decide without you.

Two weeks ago, Meta launched Muse, a personal AI agent that doesn't just answer questions. It does things. Meta says it can send an email, book travel, buy movie tickets, schedule tennis lessons, even fill out a kid's field-trip permission slip (Meta, Sept 8, 2026).

People wanted it. Muse was downloaded more than 900,000 times in its first six days and hit No. 1 on the US App Store and Google Play (Bloomberg, Sept 21, 2026).

Here's the part most business owners missed.

When Muse books the tennis lesson, Muse picks the coach.

The customer may never see your website. They may never read your reviews or compare you to the place down the street. The agent can do all of that for them, in seconds, and hand them an answer. If you weren't in that answer, you weren't in the running, and you may never know they were looking.

Search, assistant, agent: the 10-second version

Search (the last 25 years): you type, Google shows ten results, you pick. Assistant (the last two years): you ask ChatGPT, it gives you one answer with two or three names. Agent (starting now): you ask, and it goes and does it. It books, orders and buys.

Each step hands more of the choice to the machine. With search, being #4 still got you seen. With an assistant, you're one of the names or you're invisible. With an agent, you're the business that gets booked, or you aren't.

What's actually live right now (September 2026)

This isn't a someday story. Agents are already doing business on people's behalf:

Meta Muse books, schedules and checks out on your behalf, and Meta has added shopping and booking partners including Expedia, Instacart, Shopify, PayPal, Best Buy, Walmart and Wayfair (TechCrunch, Sept 23, 2026). Google AI Mode finds and books restaurant tables across reservation platforms. It expanded in April to the UK, Canada, Australia, India and more, with no subscription required (9to5Google, Apr 10, 2026). Amazon's Alexa for Shopping can buy an item automatically when it drops to your target price, and Buy for Me purchases from other brands' own websites on your behalf (Amazon; Digital Commerce 360, May 2026). Perplexity added agentic shopping with PayPal checkout in late 2025 (Perplexity), and Microsoft launched Copilot Checkout in January (Microsoft).

Meanwhile, 45% of consumers say they've used AI to find a local business, up from 6% a year earlier (BrightLocal, Mar 2026, a survey of 1,002 US adults by a local-marketing software company). And 49% of US adults now use AI chatbots, up from 33% in 2024 (Pew Research Center, June 2026).

The honest reality check

It would be easy to write "agents will buy everything next year." That isn't true, and anyone who tells you it is is selling something.

OpenAI tried in-chat checkout and pulled it back within about six months, steering ChatGPT back toward helping people discover products and sending the purchase to the merchant's own site (CNBC, Mar 2026). Consumers want help choosing more than they want a robot buying. In Gartner's research, at most 11% of consumers were willing to let AI make the actual purchase decision, but around 3 in 10 were willing to let AI narrow their choices (Gartner, May 2026). Platforms are picking sides. Amazon blocked Muse from shopping on Amazon.com, saying Meta never disclosed the agent would access the site (CNBC, Sept 23, 2026).

So the real shift isn't that agents make every decision. It's more important than that.

The agent builds the shortlist. Even when a person makes the final call, they're choosing from the two or three options their assistant put in front of them. If you're not on the shortlist, the decision happens without you.

Our predictions for 2026 into 2027

These are our predictions, not facts. Here's what we're betting on, and why.

The shortlist moves into the agent. More buyers will start with "who should I use?" asked to an assistant, and pick from what it names. Gartner's finding that people will let AI narrow their choices long before they'll let it buy is exactly the behavior that makes the shortlist decisive. Booking goes agentic first where it's simple. Restaurant tables, appointments, services, reorders: low-risk, repeatable decisions. Google already books tables in nine countries. Expect appointments and local services next. Being reachable becomes a choice. The Amazon and Meta standoff is a preview. Some platforms will open up to agents, others will wall them off, and businesses will have to decide where they want agents to find and book them. Agents reward businesses they can read. This one isn't a guess. OpenAI's documentation says merchants provide structured feeds, and that "reviews, and performance signals" improve "ranking, relevance, and user trust" (OpenAI). Google's documentation says structured data makes a business's products eligible for its shopping experiences (Google). Our read: agents lean toward businesses whose facts they can read and verify. The long game is big. Gartner predicts that by 2030, 20% of digital commerce will run through AI platforms, via on-platform checkout or AI agents (Gartner). That's their forecast, not a guarantee. But the direction is set.

What to do now: a checklist for any business

You don't need to rebuild your company. You need the agents to find you, understand you, and trust you.

Find out what the assistants say about you today. Ask ChatGPT, Claude and Gemini the question your customers ask: "Who's the best [what you do] in [your area]?" Are you named? Who is? Make your facts machine-readable. Services, prices, hours, locations and service areas should be in structured data on your website, not just in a paragraph or a PDF. Be the same everywhere. Your name, address, phone and description should match across Google, Apple Maps, Yelp, industry directories and your own site. Agents cross-check, and inconsistencies cost you trust. Earn and answer reviews. Both OpenAI and Google name reviews and ratings as signals. Recent, specific reviews matter. Answer your buyers' real questions on your own site, in plain language, so an assistant has something accurate to quote. Make yourself bookable. If an agent can't see your availability or book you, it's more likely to book someone it can. Re-check every month. The answers change as the assistants learn. Measure, fix, measure again.

Find out where you stand, in about a minute

We built a free tool that does step one for you. Type your website, and we'll ask ChatGPT, Claude and Gemini the question your buyers ask. You'll see word for word who they name, whether it's you, and what getting named is worth at your own numbers.

Run the Free AI Check →

The agents are already building the shortlist. The only question is whether your name is on it.

Intuitive Context is a founder-led AI consulting and build practice. We measure how AI answer engines and personal agents describe a business, then fix what they read. Run the free AI check, or talk with our team.