Intercom's Integrated AI and Human Support
Eoghan McCabe & Des Traynor, CEO and CSO of Intercom, on the AI transformation of customer service
The real advantage is owning the handoff between automation and people, not just owning a bot. Intercom puts the AI agent, inbox, help center, workflows, reporting, and agent assist in one stack, so the same system can answer simple questions, escalate edge cases with full context, and learn from what human agents do next. That is different from bolting an AI layer onto a separate ticketing system.
-
Intercom has been building toward this from a chat first foundation. By 2023 it was already serving more than 25,000 companies through chat surfaces embedded on customer sites, which gave it a natural place to deploy Fin and measure percent of chats resolved instantly as a core product metric.
-
Zendesk historically centered the back office system of record, tickets, agent seats, and omnichannel queues. Its AI rollout initially fit that posture, focusing on triage and agent assistance inside the help desk, while Intercom pushed harder on full resolution in the front end conversation itself.
-
The market has since split into two camps. AI native agents like Sierra and Decagon focus on autonomous resolution and often sit on top of another help desk, while Intercom sells the agent and the human support stack together. That bundled model supports seat pricing, usage pricing, and tighter product lock in.
Going forward, the winners in support software are likely to be the companies that control both the conversation layer and the human fallback layer. As AI handles more of the repetitive work, the remaining value shifts to orchestration, routing, QA, analytics, and action taking across systems, which favors integrated platforms like Intercom over point bot vendors.