Lenny's Podcast · Aug 21How Intercom rose from the ashes by betting everything on AI | Eoghan McCabe (founder and CEO)
Built into a beta within six weeks of GPT-3.5 and grew from 1 to 12 million ARR in its first year. It charges 99 cents per resolution, initially cost about 120 cents each, is the biggest customer experience agent by customers and revenue, is rated number one on G2, passed 300% growth and will pass 100 million ARR in under three quarters.
Tips
- Price AI agents per outcome, not per seat▶ 23:45Fin charges $0.99 per resolution and has a higher resolution rate than competitors; SaaS businesses were spending $20 to $30 per human-resolved ticket.
- Set price from customer value, not from your costs▶ 25:23Early cost was 120 cents per resolution at a 99 cent price; costs later dropped a lot.
- Price digital work below human work to keep it palatable▶ 24:39Human-resolved tickets cost $20 to $30 at SaaS firms and about $5 at consumer businesses.
- Hire real AI scientists and young AI talent▶ 51:52A Fin beta existed six weeks after GPT-3.5 launched; Fergal Reid, a long-time AI leader, became chief AI officer.
- Use your existing customer data to jump on new AI▶ 11:32A working Fin prototype came six weeks after GPT-3.5, and Fin went from 1 to 12 million ARR in year one.
- Position agents as more personal than slow human support▶ 13:54
- Remind the market that a late-stage company can be a startup▶ 1:12:15Fin is the biggest in its category by customers and revenue, wins direct bake-offs and is rated number one on G2.
