On 31 October 2026, Wired ran a story under the headline “Meta’s Muse Could Upend All Kinds of Businesses. It Hasn’t Yet.” The piece reports that companies have spent the past several weeks reassuring investors that Muse and other AI agents are more of an opportunity than a threat. That is the news. The headline grants the upheaval in the first sentence and takes it back in the second.

The claim, and the question I ask of it

The claim is that a personal AI agent from Meta could step between businesses and their customers. The excitement rests on an assumption: that customers will route their asking, buying and booking through one new agent, and that any business that is not ready for it loses.

I stay out of most of these conversations because they rest on hype and groundless assumptions. The first question I ask of any new tool is the same one I ask of a new hire or a new subscription. Is there a need for it? What problem does it solve that I could not already solve, and is it worth adding?

Put that to Muse. What can a customer do with it on a normal Tuesday that they could not already do with ChatGPT or Claude? Wired cannot answer that yet. Neither can the executives calming their investors. Until that answer exists, Muse is one more tool in the pile, and a pile is not a shift to plan around.

Fewer tools, used deeply

I use Claude for almost everything. I have found one tool used well to be more efficient and more effective than a stack of them, each used a little. The rush to personal agents such as Muse and Instinct, sitting next to Codex and Claude, raises an obvious question. Do we need a different agent for each job, or a pile of tools to run our lives and work?

Capabilities outlast interfaces. The name on the agent matters less than what it can actually do, and what it can do is what I test. If a new agent does something a real business needs and could not get before, I will say so and show how I use it. So far, with Muse, I cannot.

Impressive and important are different things

A new agent can be technically impressive and interesting, and still not be where things are going. Both can be true of the same product at the same time. The mistake is reading a good demo as a direction. A direction shows up in what a business can do day to day that it could not do last month. Wired’s own reporting says that, for Muse, nothing has changed yet.

What this looks like for an operator

Take a clinic, an online shop, or a marketing team of three. The question put to them this month is “what is your plan for Muse?” That is the wrong question, and it costs real time to answer.

The customer who eventually arrives through Muse, or through any other agent, will arrive with the same questions customers ask today. Where is my order. Do you have this in a medium. Can I book Thursday afternoon. What is the price for two people. The agent is a new doorway. The questions behind it do not change.

So the work that needs doing does not depend on which agent wins. It is the same work I build around at AeroChat: a customer’s question, whichever chat channel it arrives on, gets checked against the business’s own systems and turned into an action, a status, a booking or a handover, within the limits the business set. That work is worth doing whether Muse takes off or disappears.

The one thing I think is already decided for smaller businesses is the move to being data-centric. Wherever your information shows up, it needs more precision and more detail, and an accurate account of what you are, so that any agent reading it can quote you correctly. That is true today for ChatGPT and Google’s AI Overview. It will be true for Muse if Muse matters. Nothing about that work is Muse-specific.

What to do

Do not pivot for Muse. Watch it, and write down, in one line, what a customer could do with it that they could not already do. If you cannot write that line, there is no plan to make yet.

List what your customers actually ask, and in which channels. That list is the real requirement, and it does not change when a new agent launches.

Make your information precise and complete wherever it appears, with specific facts an engine or an agent can quote. That serves every agent, including the ones not built yet.

Pick one AI tool, learn it properly, and put it to work on those questions. Add a second only when the first cannot do the job.

Re-ask the need question in three months, against what Muse does then, not what the launch said it would do.

Source: Wired: Meta’s Muse Could Upend All Kinds of Businesses. It Hasn’t Yet