CRM incumbents bundle AI into workflows

Salesforce’s advantage is distribution, not model novelty. A service team already running Salesforce keeps customer records, case history, workflows, and permissions in the same system, so adding AI is less about buying a new bot and more about turning on automation inside the software agents already use all day. That makes AI easier to approve, easier to deploy, and harder for standalone vendors to displace once the workflow is embedded.

• The practical edge comes from control of the system of record. If AI can read account history, update the CRM, trigger refunds, escalate cases, and log outcomes without leaving the help desk, the incumbent removes a lot of integration work that Decagon and Sierra often solve with forward deployed engineers.
  1 sacra 2 sacra
• Incumbents still differ in how far they go. Zendesk has leaned more toward routing, agent assist, and workflow automation, while Intercom pushed a more vertically integrated AI agent with outcome based pricing and even a standalone product for teams that stay on Zendesk or Salesforce. That shows how valuable the installed base is.
  3 sacra 4 sacra 5 sacra
• For AI native vendors, this creates a coopetition pattern. Decagon can sit on top of Intercom or other support systems today, but the long game is contested. As AI agents start resolving 60 to 80% of tickets and taking backend actions, they can begin as a layer on top of the CRM and eventually challenge which product owns the customer workflow.
  1 sacra 2 sacra 5 sacra

The market is heading toward a split where incumbents use distribution and embedded data to bundle AI into existing seats, while AI native companies push deeper automation and simpler outcome based pricing. The winner will be the product that becomes the default operating layer for customer conversations, actions, and handoffs across the full support stack.