Just as brands are getting to grips with appearing in ChatGPT and Google AI, Meta has added another consideration into the mix.

Its new personal AI agent, Muse, launched in the US in September and is already turning heads.

Unlike a traditional chatbot, Muse doesn’t just answer questions. It can browse the web, use connected apps and complete tasks on your behalf.

Early signs suggest that people are interested. In a like-for-like US and Canada comparison, Muse recorded an estimated 1.8 million iOS downloads in its first 12 days, compared to 1.3 million for ChatGPT’s mobile launch.

But for brands, one feature in particular stands out - Muse can shop for you.

‘Hey Meta, buy me the best lawnmower for under £150’

A shopper can give the agent a brief, such as finding a lightweight lawnmower for a small garden for less than £150, and Muse can browse the web, compare options and prepare the purchase.

It’s an important step-change. Until now, the AI visibility conversation has largely been about whether ChatGPT, Gemini or another platform recommends your brand when somebody asks a question.

Agentic AI takes this further. Instead of simply recommending which lawnmower to buy, the AI could increasingly carry out the research and comparison itself.

That means brands may eventually need to convince two different sets of customers: the person buying the product and the AI agent acting on their behalf.

Could this be good news for challenger brands?

Potentially. People naturally bring existing preferences and brand awareness when deciding whether to buy. An AI agent given a specific task may instead compare options against price, specifications, availability, reviews and suitability.

Accenture found that 37% of loyal consumers would allow an AI agent to switch from their usual brand if it found a better fit.

That could create opportunities for challenger brands to break into consideration, particularly when they can clearly demonstrate why their product meets the brief.

It also adds another layer to AI visibility.

Product pages obviously matter, but AI systems do not operate solely from what brands say about themselves. Reviews, buying guides, YouTube content, expert commentary and coverage from trusted media can all contribute to the information available about a product.

For PR teams, that makes building authority on the websites and channels influencing AI increasingly important.

If AI starts moving from recommending products to helping purchase them, AI visibility starts becoming commercial visibility.

The trust problem

There is, however, a sizeable hurdle to overcome.

Muse becomes more useful when people connect more of their digital lives to it. On Mac, for example, users can give it access to files, Messages, calendars, Notes and Mail.

Letting an AI find a product is one thing. Giving it access to your messages and payment information before allowing it to shop for you is another.

Accenture found 74% of consumers would let an AI agent handle routine tasks and 32% would let one make purchasing decisions, but just 9% are currently open to fully autonomous shopping.

Meta has clearly anticipated some of those concerns. Purchases still require approval, while its Stripe integration can generate a single-use virtual card so Muse does not need access to the user’s actual card details.

There have already been privacy questions too. Meta recently disputed a claim that Muse accessed a user’s private messages without permission, stating that access must be granted by the user.

The interesting question is whether Meta can make this technology easier to adopt than previous AI agents.

It certainly has one advantage. Muse sits within an ecosystem people already use. More than 95% of early Muse users were also Facebook regulars and 63% had an Instagram log-in, according to Apptopia.

Rather than teaching consumers an entirely new behaviour, Meta can bring AI agents into apps and services already part of their day-to-day lives.

This is bigger than Muse

Not everybody is embracing that future just yet.

Amazon has already blocked Muse from shopping on its website, saying access from an unauthorised AI agent breached its conditions of use.

But Meta is far from the only company moving towards agentic shopping, with Google also developing technology capable of completing tasks and purchases for users.

Whether Muse itself becomes the winner is almost beside the point.

AI is moving from answering questions to taking action. If consumers increasingly hand over parts of the research, comparison and buying process, brands will need to make sure they are visible to the systems making those decisions.

At Democracy, we’re already helping brands understand where they appear in AI search, the sources influencing those answers and how PR, content and social can improve their visibility.

Get in touch if you want to understand how visible your brand is to AI and what could be influencing whether you make the shortlist.

Check out our other insights