Meta’s Muse Has 1.8 Million Daily Users. The Next Search Gatekeeper Can Complete the Purchase.

On October 6, Reuters reported that Citi counted 6.6 million downloads and 1.8 million daily users for Meta’s Muse, one month after its U.S. launch. That matters because Muse is not only answering questions: it can search, compare and transact while connected to services such as email, calendars, shopping and payments. The decision now facing marketers is whether to keep optimizing for a customer visit—or prepare for a journey in which an agent can narrow the choice and finish the purchase before the brand’s website earns a click.

The update is adoption, not another demonstration

Meta launched Muse in the United States on September 8 through a dedicated app and WhatsApp. According to Reuters, the agent can work across connected applications, continue tasks in the background and ask for authorization before certain actions. The product is deployed, but its reach is still geographically limited and its reliability is not settled; Reuters also documented mixed internal tests and security concerns around sensitive data.

Citi’s October 6 note supplied the new commercial signal. The brokerage said daily usage had reached 1.8 million and projected that Muse could generate more than $27 billion in annual revenue by 2030, mostly from transaction-related revenue. That revenue number is a forecast, not a result. The more useful fact for marketers is that an early audience is repeatedly using an assistant designed to act, not merely recommend.

The website may no longer be the first meeting

Search advertising assumes a familiar sequence: a person expresses intent, sees options, clicks and reaches a property the brand controls. Social advertising starts earlier, but still tries to produce a visit, lead or purchase that the advertiser can observe. An agent that can compare offers and make a payment compresses those stages into a private conversation.

The commercial consequence is larger than “zero-click search.” A summary can withhold traffic; a transaction-capable agent can withhold the whole consideration set. The winning brand may receive an order without ever presenting its carefully designed landing page, while the losing brands may never know they were compared.

That changes what a product page is for. It still needs to persuade people, but it also needs to expose clear facts an intermediary can evaluate: price, availability, delivery timing, cancellation conditions, compatibility, warranties and the proof behind a claim. Beautiful ambiguity becomes expensive when another system is reducing ten options to one.

Permission becomes a distribution advantage

Muse’s position inside WhatsApp gives Meta something search engines have traditionally lacked: an ongoing, conversational relationship tied to a user’s chosen services. If people grant an agent access to calendars, shopping accounts or payments, the agent can interpret constraints that a keyword never reveals. A request such as “replace this before Friday and keep it within my usual budget” contains timing, preference and purchasing authority in one instruction.

That gives leverage to the platform holding the permission layer, and to merchants whose offers can be understood and fulfilled with minimal uncertainty. It reduces leverage for publishers, affiliates and advertisers that depend on inserting themselves between query and checkout. It also pressures brands that rely on emotional copy while leaving basic commercial terms fragmented across pages, PDFs and customer-service scripts.

The strategic question is therefore not whether to “advertise in Muse.” Meta has not announced a conventional ad auction inside the agent. The immediate question is whether a brand’s offer remains competitive when the intermediary is paid to complete the task rather than maximize browsing.

Measurement loses another visible handoff

Agent-mediated commerce creates an attribution problem. A customer may learn about a product from a creator, ask Muse to compare it days later and authorize payment without generating the familiar sequence of search, site visit and retargeting events. Last-click reporting will overvalue whichever endpoint records the order and undervalue the earlier signals that made the brand eligible.

This is where OrionPilot’s organizing role becomes relevant. Keeping the offer, audience assumptions, campaign logic, creative evidence and measurement questions connected gives a team a stronger basis for judging demand when the visible click path disappears. The goal is not to claim credit for every sale; it is to preserve a coherent marketing explanation across channels while intermediaries reveal less of the journey.

Reconsider the asset your marketing is building

For years, marketers treated traffic as the bridge between attention and revenue. Muse suggests that the durable asset may be something else: a trustworthy, legible offer that can survive comparison without a salesperson, a landing-page sequence or a retargeting campaign to rescue it.

Businesses should audit the questions a purchasing agent would need answered before it could recommend them with confidence. Are prices and terms consistent? Can availability be verified? Are product differences concrete? Is evidence easy to distinguish from promotional language? Can a human take over when the request is unusual or consequential?

Meta’s early numbers do not prove that personal agents will replace search, browsers or social feeds. They do show that the intermediary capable of acting has moved from demonstration to repeat use. Marketing strategy should now account for a customer journey in which the most influential visitor may be software—and the customer may arrive only after the decision has already been made.