Unique Research · Insight / Founder story · June 21, 2024 · Source interview: Kyle Poyar (OpenView) with Paul Yacoubian, Copy.ai founder & CEO
“In a new-tech adoption cycle, incumbents often need two to three years to fold innovation into existing products. That gives innovators a two-to-three-year head start — but only if they build distribution in that window. The whole team has to move with urgency.” — Paul Yacoubian
As of March 2024, AI writing tool Copy.ai had >15 million users worldwide. SaaS growth partner Kyle Poyar sat down with Paul on how they got there.
2020 spark — copy-paste business processes
“If you invent something and can’t distribute it, it has no value for humanity,” Paul said. He believed go-to-market processes could be “copy-pasted” across companies — traditionally by hiring veterans or consultants that only the top ~1% can afford. GPT-3’s July 2020 beta made a scalable, tech-native version feel real: continuous valuable generation; could a company converse and sell the output?
MVP in 90 days
Strategy: ship MVPs in weeks on Twitter, iterate on feedback. The fifth try became Copy.ai. Each launch grew followers and interest for the next. Twitter hosted early adopters for real-time loops. In three months: >40,000 users; seed closed in three weeks. Stack of seven tools: Webflow (site/UI — Paul wouldn’t recommend that path now), Magic auth, Firebase, Stripe, Notion, Slack, and Copy.ai itself for copy. Claim: a promising company can start in 48 hours on a $100 budget.
Monetize from $3/mo
They charged for a fourth MVP, taglines.ai: would anyone use it, return, pay $3/mo? Then $5, then $10 — users paid and applied it to more marketing/sales work. Paul bought Copy.ai for $6,000 on a $167/mo lease-to-own plan and asked what a fuller platform could command vs. a $2,000/mo agency. $49/mo (with annual discount) worked for returning prosumers, not casual consumers. Later price experiments underperformed the original ladder. Early: 7-day unlimited trial; later freemium (compute cheaper) with 2,000 words + 200 bonus credits free.
Path to 1M — then 10M
2,700 users in three days; 40k in three months. After Twitter, other social and heavy SEO — ≥30 free tools (à la HubSpot headline / Shopify name generators); “people need Instagram caption help.” Domain authority compounded into a flywheel, not a funnel. Paul built in public while competitors stayed quiet — he credits that visibility. Early “marketing” was Paul alone on Twitter; toward 10M they hired lean specialists (social, SEO, email) — one person, one channel. Prosumer brand personality: founders should inspire, not only inform.
How Copy.ai uses AI internally
AI drafts free SEO tools, descriptions, landing pages, and I/O examples; humans lightly edit. Blog workflow now starts from real internal/advisor conversations rather than blank pages. On signup/demo, a live AI workflow profiles the person, role, and company for personalized inbound sequences, case suggestions, and tailored plans — “every step of sales accelerates with strong data,” vs. generic personas.
PLG
One-click Google login; freemium for discoverers who aren’t ready to buy (pure 7-day trials expired before evaluation). Enterprise deals average ~five stakeholders — five entry points; end users often tell you who to meet and in what order.
Advice to AI founders
Joke first: “You’re four years late” — pointing at Attention Is All You Need (>110k citations) and >90% of AI capital linked to that line of work. Seriously: build a self-contained vertical app, own as much of the value chain as possible, sell it as an operated service, and repeat across industries and sub-functions.
Original interview: Kyle Poyar. 5 CN images; attach on ingest.
— END —