Listen now on
Spotify . YouTube . Apple Podcasts
Amos Bar Joseph is co-founder and CEO of Swan, and the loudest voice behind the three-founder, $30M ARR experiment. A year later Swan is six people. He's fine with that, and the reason is the whole episode.
The North Star hasn't moved: $10M in ARR per employee, the most leveraged company ever built. What moved is his read on the constraint. It was never the work. It's the deciding.
What you’ll learn
Why Swan doubled from three founders to six people, and the hiring rule about surface areas with no mental owner
What the $10M ARR per employee North Star does to every small decision inside the company
How one inbound process, defined once, now runs 30,000 times a month on a $100K AI budget
His two-lens answer to build versus buy, and why buying software is buying thousands of decisions, not developer hours
What he stopped believing about AI autonomy, and the specification work that never went away
Episode Highlights
01:24 - The three-founder, $30M story, one year later
03:40 - Doubling headcount without dropping the North Star
03:55 - Scaling is decisions, not work
05:08 - The solo-founder comparison and annualized revenue
09:17 - The first thing that broke was sales
13:55 - Halligan, Dorsey, and rebuilding the org chart
17:15 - The GTM brain as the central nervous system
19:30 - Scoring decides where humans stay in the loop
22:01 - A $100K AI budget for three GTM people
22:44 - One process, 30,000 runs a month
24:12 - Build versus buy, two lenses
27:53 - Buying software is buying thousands of decisions
30:56 - What he stopped believing about autonomy
33:14 - What the unicorn playbook costs you
35:42 - The one line to post for your team
Key Takeaways
1. The bottleneck is decisions, not work
Swan launched with three founders, seven figures in ARR, and a public bet to reach $30M without hiring. A year in, the constraint wasn't output. AI can do the work, but every piece of it generates micro decisions, and automating those away means someone else is building your business. They doubled to six people to buy decision-making capacity.
2. Hire when a surface area has no mental owner
His rule: if nobody goes to sleep thinking about a part of the business, that part is broken or about to be. Sales broke first at Swan, because his energy went to growth and noise while pipeline work deflated him. AI doesn't think about areas no human owns.
3. Scoring decides where humans stay in the loop
He treats scoring as resource allocation. High-scored leads get human review on the outreach, the deal room, the follow-up. Low-scored leads get the automated version at 80 to 90% of the quality, because decision capacity is the scarce resource, not the work.
4. Define a process once, run it 30,000 times
GTM work is async, which makes it easier to scale than software. Their inbound flow, de-anonymize, research, score, sequence, alert, got reviewed for its first hundred runs, refined, and now fires 30,000 times a month. The AI bill for three GTM people: around $100K a year.
5. Buying software is buying decisions
Two lenses. Humans versus tokens: never hand AI your core IP, which for Swan is its LinkedIn game. The tech stack: almost never build.
"Buying enterprise software was never about buying developer hours. It was buying the thousands of decisions that were incorporated into building that product."
6. Specification is the work that stayed human
What he stopped believing: that AI does much with little specification. Production costs collapsed, but the thinking around production, specifying the work upfront and reviewing it after, got bigger. His one line for GTM leaders to post for their team is at the end of the episode.
If you're enjoying these episodes, reply and let me know what you'd love to learn more about: leverage math, build versus buy, lead scoring, decision velocity, anything.
Catch you next week,
Jared
P.S. Amos writes about building Swan at theautonomousage.com, and he's on LinkedIn. Also, always growing our RevGenius community; you can join here to follow along.
Upcoming events
Revolt! 2026 (Digital Conference)
(14-15 October, 11 AM to 5 PM ET)
Two days, dozens of sessions, 9,000+ leaders.
Last year drew registrants from 3,400+ companies across 40+ countries.
Revolt! 2026 line-up:
October 14th | 11:15 AM ET - What is GTM Engineering
Florin Tatulea of ZoomInfo on what the role actually is and where it sits.October 14th | 12:45 PM ET - The New Revenue Tech Stack: What to Build, Buy, and Integrate in the Age of AI
Tom Andrews and Chad Boersma of Airspeed on where your stack goes next.October 14th | 1:45 PM ET - What Happens When AI Starts Doing the Work in RevOps?
Brandon Bussey (Couchbase), Matt Flotard (Gong), and Katie Daggett (CADRE AI), with Mollie Bodensteiner of ZoomInfo moderating.October 14th | 2:45 PM ET - Unique Buying Signals for Outbound
Tori Seidenstein of Tadata on signals your competitors aren't watching.October 14th | 3:30 PM ET - Planning Done Right: The 6-Step Framework for Plans That Actually Work
Stephanie Martin of Lative on building plans that hold past January.October 15th | 11:15 AM ET - You Bought the AI Tools. Why Isn't Pipeline Going Up?
Mintis Sow of HubSpot on why the spend isn't showing in pipeline.October 15th | 12:00 PM ET - 5 Bold Predictions for GTM in 2027
Sangram Vajre of GTM Partners on where GTM lands next.October 15th | 12:45 PM ET - The Connective Tissue Between Sales and Marketing
Davide Grieco of Clay on making the two teams run as one motion.October 15th | 2:30 PM ET - The AI-Enabled Sales Leader: Coaching, Structure, and Metrics in an AI-First Org
Kevin Dorcey of Leanscaper on what sales leadership looks like now.October 15th | 3:15 PM ET - The New Marketing Moat: If Everyone Has AI, What Makes You Different?
Mark Kilens, Natalie Marcotullio, Devin Reed, and Melissa Moody on differentiation when everyone runs the same stack.
More sessions dropping soon!
New Perks just dropped
Firmable: 20% off any annual plan
AI-native sales platform with company and contact data, buying signals, and agentic AI for B2B sales teams across the US, Canada, and APAC.
TitanX: 20 reserved TitanX 2.0 waitlist spots plus a 10-point bonus
Phone intent platform that tells outbound reps who is likely to answer before they dial.
Want to get your brand in front of 60,000+ revenue leaders?
Partner with us and put your name at the center of the conversation.
