Intercom's $250M AI Support Bet
Intercom's $250M/year AI bet
Intercom is choosing to win by owning the whole support workflow, not by being a chat widget. The bet is that if AI can answer most routine questions instantly, the product that controls the messenger, help center, customer record, and human handoff becomes the system customers build around. That is a sharper position than Drift’s sales focused bot layer and more disruptive than Zendesk’s seat based ticketing core.
-
Earlier bots worked best in narrow phone tree use cases, like delivery apps with a small set of repeat issues. Fin changes the economics because teams can point it at existing docs and start resolving open ended questions without months of scripting and rule maintenance.
-
This also forces Intercom to remake its business model. Traditional help desks charge by agent seat. Fin introduced per resolution pricing, first at $1.90 and later $0.99, because the value is removing a conversation from the queue, not giving another rep a login.
-
The strategic opening is real, but the market is getting crowded fast. By February 2026, AI support agents like Sierra and Decagon were growing quickly, while some customers were already layering AI agents on top of Intercom or Zendesk. That makes ownership of the underlying support stack more valuable.
From here, customer service software is likely to consolidate around platforms that combine AI resolution, agent assist, knowledge management, and workflow actions in one place. If Intercom keeps increasing the share of chats it resolves instantly, its messenger footprint across thousands of sites can become both its distribution edge and its moat.