7 Free Ways To Get Customers With AI
TL;DR
Distribution is now the moat — Greg Isenberg says the Silicon Valley hierarchy has flipped from engineers-first to marketers/distribution-first, because code is commoditized and 200,000 new vibe-coded projects are being created every day on Lovable.
Start with audience, then build the product — his core framework is “distribution first, product second,” using Peter Levels as the model: 750,000+ followers, strong SEO, 125,000 tweets over years, and $3M+ revenue with zero employees.
MCP servers are a new zero-CAC acquisition channel — Greg frames building an MCP server in 2026 like building for mobile in 2010, citing a fintech friend who got 150+ installations in 30 days with $0 ad spend by showing up inside AI assistants.
Programmatic SEO still works if the pages are genuinely useful — his math: 10,000 pages × 30 visits/month = 300,000 visitors; at a 2% conversion rate and $10 per conversion, that can become $60,000/month, but only if the content is structured, valuable, and not obvious AI slop.
Free tools and shareable outputs can become self-propagating marketing — examples like Ahrefs’ backlink checker, Spotify Wrapped, GitHub contribution graphs, and Duolingo streaks all create artifacts people want to share, turning product usage into distribution.
You can buy or manufacture attention faster than building from scratch — Greg recommends either acquiring niche newsletters with 5,000–50,000 subscribers for roughly $5,000–$20,000 or using AI to turn one 30-minute “pillar” piece of content into tweets, LinkedIn posts, short clips, newsletters, and more.
The Breakdown
The big flip: marketers are becoming the winners
Greg opens with a pretty blunt thesis: everyone is teaching people to vibe code, but almost nobody is teaching distribution. He says Silicon Valley used to worship engineers, then product people, and now the people who understand brand, content, advertising, and audience-building are moving to the top because code is cheap and attention is scarce.
Why “build it and they will come” keeps failing
He describes the classic founder loop: build a product, launch to silence, add more features, launch again, get more silence. His alternative is to build an audience first — even 1,000 people — ask them what they want, build it in 24 to 72 hours, then launch into a warm room instead of shouting into the void.
MCP servers as an AI-native sales team
His first tactic is building MCP servers so AI assistants like Claude or ChatGPT can surface your product directly when users ask relevant questions. He calls it “zero CAC” and compares building for MCP in 2026 to building for mobile in 2010; one fintech example he mentions got 150+ installs in 30 days with no ad spend.
Programmatic SEO, but not the lazy spam version
Greg’s second tactic is creating large numbers of SEO pages around structured keyword patterns like “best CRM for dentists,” using scraped datasets, templates, and AI-generated copy. He knows the audience is rolling their eyes at “just make 10,000 pages,” so he stresses that the content has to feel useful and human — start with 100 pages, monitor indexation, then scale.
Free tools that do the marketing for you
Next he argues the tool itself can be your top-of-funnel, using Ahrefs’ free backlink checker as the example: give users a taste, capture their info, and upsell them into the main product. His point is that AI coding tools make this dramatically easier now — something that used to take months can be built in a day and keep attracting users long after launch.
From SEO to AEO: getting cited by ChatGPT and Perplexity
Greg then pivots from ranking on Google to being the source AI cites, calling this answer engine optimization. He says the winning format is direct, structured, citation-worthy answers with schema markup and FAQs, and cites Peter Levels’ claim that AI referrals jumped from 4% to 20% in a month.
Viral artifacts: make users want to show off
One of the more memorable sections is about designing outputs people naturally share — his examples are Spotify Wrapped, GitHub’s green contribution squares, Stripe Atlas incorporation milestones, and Duolingo streaks. The key question he keeps coming back to is simple: what does your user want to brag about, and can you make that output beautiful, branded, and one-click shareable?
Buying attention or manufacturing it with AI
He closes with two attention plays: buy a niche newsletter instead of starting from zero, or build an AI-powered content repurposing engine. On newsletters, he says 5,000 to 50,000-subscriber lists can often be bought for $5,000 to $20,000; on content, he recommends recording one 30-minute podcast, video, or even a voice memo, then turning it into tweets, LinkedIn posts, short-form videos, a newsletter, and quote graphics — more shots on goal without manually doing all the work.
The closing message: pick two and start this week
Greg wraps by repeating the core point: code used to be the moat, now distribution is. His energy here is basically part pep talk, part intervention — stop vibe coding in private and hoping customers magically appear; pick two of these seven tactics, ship this week, and build a real growth engine around the thing you made.