Unpopular opinion: I’m not convinced Meta Muse has a great business model

A lot of people seem extremely bullish on Meta Muse right now.

I’m a little more skeptical.

Not because I think Muse is a bad product — actually, I think Meta has done a few things extremely well.

But I’m not convinced the long-term business model is as obvious as the initial hype makes it look.

The first thing is that the personal AI agent category itself isn’t new.

Before Muse, we already had products moving in the same general direction:

• OpenAI Operator / ChatGPT agent
• Google’s agent products
• Microsoft Copilot agents
• Anthropic Cowork

So to me, the interesting question isn’t:

“Can Meta build a personal AI agent?”

Clearly it can.

The harder question is:

“What advantage does Meta have that makes Muse sustainably better or more profitable than everyone else?”

And this is where I’m less convinced.

Consumer AI is hard to monetize

Most AI companies have increasingly pushed toward enterprise customers for a reason.

Businesses are willing to pay meaningful amounts of money when AI saves employee time, replaces workflows, or directly increases revenue.

Consumers are much more price sensitive.

A personal agent that books restaurants, finds deals, fills out forms, or manages your calendar sounds useful.

But how many consumers will eventually pay $20, $50, or $100 every month for it?

Especially if several companies offer similar functionality.

Commerce sounds attractive, but incentives get complicated

The obvious alternative is transactions.

If your agent helps you find a hotel, buy a product, order food, or book a service, the agent platform could potentially take a referral fee or transaction cut.

Conceptually, this sounds great.

But there’s a problem:

The most valuable merchants already own their customer relationships.

Amazon wants you on Amazon.

Booking wants you on Booking.

DoorDash wants you inside DoorDash.

Large platforms make money from customer engagement, ads, subscriptions, data, upsells, and repeat purchases.

Why would they happily give an AI agent control of that relationship?

The merchants most willing to give an agent generous economics may end up being the merchants that have the least bargaining power.

That makes building a truly neutral “agent layer for everything” much harder than it sounds.

Where I think Meta absolutely nailed it: distribution

This is where Muse becomes much more interesting.

Meta has something most AI startups don’t:

massive consumer distribution.

It can put Muse in front of people through Instagram, WhatsApp, Facebook, creators, ads, social sharing, promotions, and eventually hardware.

That is a huge advantage.

Muse could become the first personal AI agent that normal people — not just people following AI every day — actually use.

And I think Meta deserves a lot of credit for that.

The product also feels much more consumer-oriented than many previous agent attempts.

Instead of selling “agentic workflows,” Meta can simply tell someone:

Give Muse something annoying to do, and it’ll do it for you.

That’s a much easier pitch.

So my take is basically:

I’m bullish on Muse’s ability to get users.

I’m much less convinced about its ability to turn those users into a durable, high-margin business.

And those are two very different things.

Meta may eventually prove this wrong through subscriptions, commerce, ads, hardware, or some model we haven’t seen yet.

But right now, distribution looks like the strongest part of the story to me — not monetization.

Curious what others think.

What is actually going to be Meta Muse’s long-term business model?