Intercom Zapier Brex AI bets

These were not side projects, they were redesigns of each company’s core job to match what AI now makes possible. Intercom moved from selling support seats to selling resolved conversations, because AI can answer a large share of tickets directly. Zapier moved from a point and click automation builder toward plain language workflow creation, because AI lowers the skill needed to build automations. Brex embedded AI into expense review and documentation, because finance teams care most about faster, cleaner close and policy enforcement.

• Intercom had the most existential pressure. Support is one of the clearest AI replacement categories, so it reorganized around Fin, expanded its AI team, and changed pricing to $0.99 per resolution. That shifts the product from helping agents work faster to doing the work itself.
  1 sacra 2 sacra 5 intercom
• Zapier’s risk was subtler but just as real. Its old product asked users to wire together triggers and actions by hand. AI makes that interface feel outdated. Natural language generation of Zap outlines and a company wide roadmap reset were about preserving Zapier’s role as the control layer between apps.
  1 sacra 3 sacra 6 zapier
• Brex did not need AI to save its pricing model, because most of its revenue already came from interchange and usage. The bet was product depth. AI helps turn messy receipts, memos, and policy checks into structured finance data, making Brex more valuable to controllers and accountants, not just cardholders.
  1 sacra 4 sacra 7 brex

The next phase is each company pushing AI from feature to system of record advantage. Intercom can own more of the support stack if Fin keeps resolution quality high. Zapier can become the orchestration layer for AI agents across software. Brex can move from expense cleanup into broader autonomous finance workflows, where the real prize is controlling how company money is reviewed, coded, and approved.