Greg Isenberg just had a baby. He needed a night nurse.
He couldn't find one through friends, so he called an agency. And the agency admitted, out loud, that their whole process was scrolling Facebook groups like "Miami night nurses" and charging him thousands of dollars for it.
That phone call turned into a 24-minute episode where Greg claims Meta's new personal agent will mint over 1,000 millionaires. And after watching it, I think his playbook to make money with Meta Muse is one of the clearest "first internet dollar" opportunities anyone has mapped out this year.
Here's the full episode, then the breakdown with every number he dropped.
The night nurse problem that started everything
Quick context if you've been heads-down shipping: Meta launched Muse on September 8, 2026. It's their personal AI agent. Free tier, $20/month Power plan, $100/month Maximum plan. Greg pays for the $100 one.
What makes it different from every other agent: it sits on top of Facebook groups, Marketplace, and Instagram. Billions of users posting "I need X" and "I can do X" every single day.
So when the night nurse agency told Greg they were just mining those groups manually, he opened Muse and typed the equivalent of: find me night nurses at $30 an hour in Montreal.
It came back with a ranked list. It had searched "Montreal moms looking for babysitters", "Babysitter needed in Montreal", "West Island Community Montreal", "Filipino caregiver Montreal". Ranked by quality, parent recommendations first.
The thing an agency charges thousands of dollars for, with a team of five people doing the grunt work.
His reaction: "That's crazy that I would pay a fee for that."
Why this is a real arbitrage (and not just a party trick)
Greg's argument is an old one, and that's exactly why I buy it.
He compares it to early eBay. The people who made millions there weren't geniuses. They just noticed, before everyone else, that connecting buyers and sellers in a brand new marketplace is worth a tremendous amount of money.
Same structure here:
- Buyers: billions of people posting needs in Facebook groups, on Marketplace, on Instagram.
- Sellers: people in those same places saying "I can do this thing, who needs it?"
- You: the person whose agent cleans, matches, and connects the two sides.
The economics flipped. The old agency needed five humans and charged thousands. You charge a few hundred dollars as a success fee, pass savings to the customer, and still keep what Greg calls "insane margins" because Muse does the structure, the matching, the DM drafting, even the calls and booking.
And the niches are endless. His examples: night nurses in Tampa. Furniture in Austin. Plumbers in Phoenix. One niche, one city, one set of rules for who gets approved.
Pick yours. That's the whole starting move.
The 4 levels: from side hustle to software company
This is the part of the episode worth stealing. Greg lays out four levels, and each one is a real business on its own.
Level 1: Mining
You're finding gold in other people's groups, on Marketplace, on Instagram.
One catch he demos live: Muse won't cannibalize Meta's own platforms by default. His first night nurse search pulled from generic public listing sites, not exactly a goldmine. You have to explicitly tell it to search Facebook groups. When he did, he found "some of the highest quality data is coming from Facebook groups."
→ Prompt it where to dig, or it digs in the wrong place.
Level 2: Own the list
Mining pays you once. The list it produces keeps paying.
Take the data off platform into Airtable or Google Sheets: names, needs, prices, trust signals. Every week it compounds. Greg's point is that this repository is yours to keep, and eventually it's sellable on its own (he's planning a whole episode on data brokering).
I'd add one thing from my side: a list only prints money if you work it. The follow-up, the check-ins, the "are you still looking?" messages. That's a lead nurturing problem, and it's where most people who get this far will quietly lose.
If the data-as-asset angle clicks for you, I broke down how Rashid built $32K MRR on exactly this model in database business ideas.
Level 3: Own the room
Here's the contrarian bit I haven't seen anywhere else.
If millions of Muse agents are about to crawl Facebook groups (1.8 billion monthly active users by Greg's count), then the groups themselves are undervalued. So are Facebook business pages and niche Instagram pages.
Why aren't you creating "The Montreal Night Nurse Agency" page right now? The agents are coming to crawl it.
Greg even floats buying existing Facebook groups, with a flag he raises himself: check Meta's terms of service first, because he's not sure how kosher that is. Neither am I. Treat that one as a maybe.
Level 4: Build technology
Once you've run levels 1 to 3, you know a niche's workflow better than almost anyone. That's when you build the SaaS or the vertical agent that automates it end to end.
That's the difference between "I make $2,000 a month" and enterprise value you can sell.
This ladder is becoming Greg's signature move. He ran the same "new AI tool, concrete money plays" structure earlier this month, and I covered it in Make Money With Gemini 4 Argon.
The tactical stack: phone numbers, flips, and content
Three tricks from the episode.
First, give your agent a phone number. People are connecting Muse to Bland AI's agent phone plan: a real US number with talk and text (Greg pegs it around $15 a month; Bland's listed plan runs $29.99 with a $14.99 first month). Then you prompt Muse: get yourself a phone number and call me. Suddenly your one-person agency reaches out, follows up, and closes sales by phone. No affiliation on his end, none on mine.
Second, Marketplace flips. Everything on Marketplace is public data. So have Muse hunt for mispriced or mislabeled listings, listings with no photos or details, stuff that's been sitting forever, pickup-only items (his example: pickup-only Mac minis), and price drops or relists where sellers are ready to negotiate. Then let Muse negotiate on your behalf.
Third, use it as your content hub. Muse can research what's working on Instagram in your niche and help you build the audience side. Greg's framing: groups are where Muse hears needs, Instagram is where it finds people and proves they're good, and your own content is how you end up owning trust and attention.
My honest take before you quit your job
I run this blog on receipts, so here's the other side.
Everything above is one founder's demo plus his pattern recognition. Nobody has publicly shown $1M earned through Muse yet. Greg says it himself: "I'm not sure how long it's going to last."
Three things I'd watch:
- Margins attract competition. The night nurse play works great until five other people run the same prompt in your city. Speed and list ownership are the only moats here.
- Platform risk is total. If Meta decides tomorrow that agents can't mine groups, levels 1 and 2 evaporate. (Levels 3 and 4 survive. That's the point of climbing.)
- The agent wars just started. Greg name-checks Dots and Instinct as rivals. He thinks Muse wins on Facebook groups, Marketplace, and Instagram access, and that Meta is pushing it hardest for mainstream adoption. Reasonable, not guaranteed.
The core logic holds even if Muse specifically loses the agent war, though. Demand data is sitting in public groups, agents can now clean it, and whoever ends up holding the cleaned list for a niche is the one both sides have to call.
Start with one niche, one city, this weekend. The free tier is enough to find out if your niche has a pulse.
FAQ
What is Meta Muse?
Meta Muse is Meta's personal AI agent, launched September 8, 2026 in the US and Canada. It can search and act across Facebook groups, Marketplace, and Instagram, plus public web data: finding people, drafting DMs, booking, and (with a connected phone plan) making calls and texts on your behalf.
How much does Meta Muse cost?
There's a free tier, a $20/month Power plan, and a $100/month Maximum plan. Greg Isenberg's advice: start free, land your first clients, and only upgrade once you're earning (he upgraded to the $100 plan for the better model and higher limits).
What's the fastest way to make money with Meta Muse?
The matchmaking agency play. Pick one service niche in one city (night nurses, plumbers, movers), have Muse mine public Facebook groups and listings for both sides, clean the results into a list, and charge a success fee of a few hundred dollars per connection. It's the exact job agencies already charge thousands for.
Is mining Facebook groups with Muse against the rules?
Muse itself only searches public data by default, and you can prompt it into public groups (private groups require you to join first). The grayer zone is buying groups to position for agent traffic. Greg flags that one himself: read Meta's terms of service before you try it.
Steal playbooks like this every week
I publish breakdowns like this because I'm doing the same climb: bootstrapping in public, stealing what works from founders a few steps ahead.
On the Profitable Founder Podcast I interview bootstrapped SaaS founders making $100K to $10M a year, and they walk me through exactly how. The numbers, the channels, the mistakes.
