Skip to content
Agents tracked: 484 Releases (7d): 690 Agent status: 2 with issuesSince last limit reset: Claude 7d · Codex 5d For AI agents: Markdown Updated Nov 18, 2026
Case study RipplingResults as reported by Decagon

How Rippling scaled multi-product customer support with Decagon AI agents

Rippling, an HR, IT and finance platform, replaced a decision-tree support tool with Decagon's AI agents to handle questions across many products. As reported by Decagon, chat deflection rose and ticket routing improved.

The problem

Rippling's earlier support tool relied on decision trees and scripted answers, which needed heavy manual oversight and struggled with vague questions such as "W-2 issue". It often routed tickets to the wrong teams, and customers could self-serve in chat only about 38% of the time. There was no AI agent available in email.

How they did it

  1. Connect the AI agent to existing knowledge bases and past tickets as learning sources.
  2. Integrate with Rippling's internal APIs so answers reflect the user's own situation, such as healthcare enrollment status.
  3. Work with Decagon engineers to define 75+ tags, the routing for each, and the cases that require escalation.
  4. Enable the agent to ask clarifying questions and take actions on behalf of the user.
  5. Roll out AI-enabled email deflection and use Agent Assist tooling for human agents.
  6. Build custom API workflows with Decagon's forward-deployed engineering team.

Results as reported by Decagon

  • As reported by Decagon, chat deflection rose from 38% to over 50%, and the case study headline cites a 32% increase in deflection.
  • As reported by Decagon, the new tagging and routing setup gave an immediate 7% improvement in conversation routing.
  • As reported by Decagon, the deployment supports 12+ product lines and 400,000 users.
  • Rippling launched AI-enabled email deflection, which it did not have before.

These figures are as reported by Decagon in Decagon customer story ↗. Decagon sells the product used here. AgentGid did not measure or independently verify them.

Takeaway. For support teams with many distinct products, connecting an AI agent to internal APIs and defining detailed tags and routing rules up front can improve both deflection and ticket routing.
AgentGid's take

This setup suits a larger support team with several products, internal APIs it can expose, and engineers who can define tags and escalation rules with the vendor. Decagon has no public pricing (custom, sales only), and the deflection and routing gains are vendor-reported, so ask for a pilot baseline. A smaller team could try Rasa (Free), which is open source but limited to one bot and 1,000 external conversations a month.

The agent used here

Similar use cases

Guides