Des Traynor (Intercom): How we rebuilt the company for the AI era

Key takeaways

1. Traynor says Intercom assumed Fin would pull customers off their existing help desks, but they only got "flirty eyes and longing looks"; adoption took off once Fin ran on top of Zendesk and Service Cloud, a step that took a year to see.
2. Because AI "massively decreases the cost of exploration," teams can prototype several ideas, pricing models or designs side by side before committing to one.
3. Pricing work that took two weeks of business operations modeling can now be explored in an afternoon with a model like Claude.
4. Intercom moved Fin to outcome-based pricing, charging per successful resolution, because Fin shrinks the seat counts its core business was built on.
5. On build versus buy, Traynor says reaching a 20% to 30% resolution rate is easy but not worth deploying; keep your best talent on strategy rather than replacing internal tools to save a few grand.