- Revenue Creator
- Posts
- Amos Bar-Joseph: 1M impressions, millions in pipeline, zero ghostwriters
Amos Bar-Joseph: 1M impressions, millions in pipeline, zero ghostwriters
He almost let AI write it. That would've killed everything.
Amos Bar-Joseph built a content engine that generates 1M+ impressions a month.
The secret isn’t AI. It’s context.
Amos Bar-Joseph has built and scaled two B2B companies. He knows what the old model looks like: growth at all cost, headcount as leverage, hustle as strategy. Swan AI is his third company, and he's building it differently. Three people. An autonomous business OS. And a founder-led content motion on LinkedIn that, for a while, was threatening to eat him alive.
The posts were working. The problem was writing them.
The Challenge
Amos wasn't posting tech stack carousels or recycled listicles. He was documenting Swan's real journey to $10M ARR per employee: specific, opinionated, personal. That kind of content takes time. Two hours a post, by his estimate.
The obvious answer was to let AI write it. He didn't take that path. "We're gonna lose our unique advantage," he told me. "My personal touch. My personal storytelling."
So instead of automating the output, he went after the bottleneck differently. How do you build a human-AI collaboration workflow that speeds up the process and produces better work?
Here's what he figured out.
The Playbook
First, a prerequisite Amos won't let you skip: before any of this works, you need a point of view. The content pillars Claude eventually learned to work from didn't come from AI — they came from real strategy work Amos did upfront with advisors. The workflow amplifies your thinking. It doesn't replace the thinking you haven't done yet.
With that in place, the system runs on three layers.
Layer 1: Conversations that compound. Amos keeps a dedicated Claude project for post writing, nothing else lives there. Every conversation within that project is accessible to Claude across sessions. Posts build on posts. Hooks that landed inform the next round of hooks. Instead of starting from scratch each time, Claude picks up context from what came before. Most people let every AI conversation go to oblivion. Amos treats them like institutional memory.
Layer 2: Skills over stuffed prompts. Rather than cramming his entire context window with everything at once, Amos built reusable skills inside Claude — one for his content pillars, one for his best-performing hooks, one for his tone of voice, one for how to analyze Swan's narrative from different angles. Claude doesn't load all of it on every query. It surfaces what's relevant to the specific request. The result is cleaner, more accurate output without the noise.
Layer 3: A workflow designed for collaboration, not commands. This is where the actual writing happens, and it's deliberately iterative. Amos starts with a brain dump — no overthinking, just the raw idea. From there, Claude walks him through a defined sequence:
First, it suggests three angles to approach the topic, drawing on his content pillars. (Same story, different lenses: maybe it's about misaligned product strategy, maybe it's about valuation inflation over value creation.)
Amos picks one. Then Claude proposes three arc structures — the emotional journey the reader will take, step by step, at a high level.
Once the arc is chosen, Claude surfaces the top hooks it thinks will perform best given that specific angle and structure, informed by the skills built from Amos's historical post data.
Every step requires a human decision. It's not a vending machine. It's a thinking partner that extends your judgment rather than replacing it.
The Results
Posts that used to take two hours now get done in a fraction of that time. The output is consistently generating over 1 million impressions a month on LinkedIn, and Amos attributes millions of dollars in pipeline directly to the content motion.
The workflow didn't water down the voice. If anything, the structure forced more intentionality into every post — the right angle, the right arc, the right hook — not just the first thing that came to mind.
That's what context engineering actually looks like in practice.
- Amos Bar-Joseph, Co-Founder and CEO of Swan AI
Want to learn from more Revenue Creators like Amos? Join the RevGenius community and be part of the movement rewriting the GTM playbook.
Upcoming Events
GTM Demo Day
Experience the future of GTM with an exclusive look at the essential tech stack for 2026. This no-fluff showcase delivers high-value, 10-minute demos from brands solving real technical challenges, paired with lively community networking and exclusive partner discounts. It’s a straightforward, product-first event designed for immediate impact.
The SaaSpocalypse Debate: AI and the Future of RevOps
Dive into the "SaaSpocalypse" and the shifting landscape of RevOps in the age of AI. Rather than just managing tool sprawl, RevOps leaders are now architecting the data models and automated workflows that define modern revenue systems. Join us to discuss whether the SaaS stack is truly shrinking or simply evolving, and learn what it takes to design systems that AI can reliably navigate.
Funding Announcements
GTM Hot Jobs from RevOps Pipeline
Director, Sales Enablement & RevOps - Arlo - $150K - $200K - Apply
Customer Success Manager - PartnerStack - $70K - $95K - Apply
Director of Product, Automation - Retool - $298K - $412K - Apply
Business Operations & Strategy Manager - Scale AI - $176K - $220K - Apply
Sr Director, Product Management - Salesforce - $218K - $365K - Apply
Are you hiring and want your job featured? Submit a role here.
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.