Ownership shaped AI product velocity
Eoghan McCabe & Des Traynor, CEO and CSO of Intercom, on the AI transformation of customer service
This was really a claim about ownership structure shaping product velocity. Intercom was arguing that AI in support required a messy, risky rewrite of product, pricing, and workflow, while Zendesk, after its November 22, 2022 take private deal and founder exit, had stronger incentives to protect the existing per agent help desk business, raise monetization, and add AI more cautiously inside the current system of record rather than bet the company on autonomous resolution.
-
Intercoms view of the market was that real advantage came from joining the bot, the human inbox, the knowledge base, and customer data in one system. That matters because the hard part is not generating text, it is deciding when the bot should answer, when it should hand off, what customer context it can use, and how the human reply trains the system for next time.
-
Zendesks legacy center of gravity was the ticket. A customer writes in by email, chat, phone, or social, Zendesk creates a ticket, routes it to an agent, and sells seats and suites around that workflow. In that model, the safest AI is agent assist, triage, and pre written reply help, because it improves labor efficiency without immediately shrinking the seat base that drives revenue.
-
The market since then has validated that moving fast on AI changed the category. Intercom layered in per resolution pricing and reached an estimated $343M of revenue in 2024, while newer AI native vendors like Sierra and Decagon pushed containment to roughly 60% to 80% and built businesses around replacing human handled tickets outright. That raised the cost of waiting for every incumbent.
Going forward, the winners in support will look less like classic help desks and more like full stack resolution platforms. The control point is shifting from owning the inbox to owning the resolved outcome, which favors products that combine AI agent logic, human escalation, workflow automation, and customer context in one joined up system.