Intercom's AI Survival Bet
Eoghan McCabe & Des Traynor, CEO and CSO of Intercom, on the AI transformation of customer service
This was a survival decision disguised as a product bet. Intercom had built a seat based help desk business, but AI changes the basic unit of value from paying for human agents to paying for resolved issues. If Intercom had stayed focused on selling more seats, it would have been defending a shrinking pool of spend while customers, and newer AI vendors, moved toward automation that removes seats entirely.
-
Intercom had already seen the problem in its numbers. Growth slowed to 10% in 2023 as customer service teams got smaller, then re accelerated to $343M in 2024 by adding AI into higher tiers and charging 99 cents per AI resolved ticket. That shows the risk was not theoretical, it was already hitting the core model.
-
The workflow itself was changing. Older bots followed scripts or matched canned answers, but Fin could read support docs, understand free form questions, and handle a large share of conversations without constant rule writing. Once that becomes good enough, buyers expect fewer agents, not better agent tools.
-
Competition made waiting even riskier. Zendesk could push AI into a huge installed base, Gorgias already fit better with usage driven support economics in ecommerce, and newer players like Sierra and Decagon were built around replacing outsourced support and human queues from day one.
This pushes customer service software toward a winner that owns both the system of record and the automation layer. The strongest vendors will be the ones that can answer the question, take the action, and prove the outcome, while pricing on work completed instead of seats occupied. Intercom’s early full commitment gave it a path to stay in that top tier.