Decagon
Enterprise AI agents for customer support across chat, email and voice, configured in plain language with Agent Operating Procedures
Overview
Decagon builds AI agents that handle customer conversations for large companies over chat, email and voice. Support teams describe how the agent should behave in natural-language Agent Operating Procedures (AOPs), which the platform turns into reliable workflows that can look up accounts, process refunds, change orders and escalate to humans. Analytics and testing tools show how conversations go, where customers get stuck and what to improve.
Founded in San Francisco in 2023 by Jesse Zhang and Ashwin Sreenivas, Decagon grew very quickly: it raised $250M at a $4.5B valuation in January 2026 and reports enterprise customers including Block, Chime, Notion, Duolingo and Avis Budget Group. Pricing is per conversation by default, with per-resolution pricing available.
The Verdict
Who Should Use Decagon?
Best For
- Enterprises with large, complex support operations
- Teams that want support leads, not engineers, to shape agent behavior
- Fintech and consumer tech with many account actions
- Omnichannel support including voice
Not Ideal For
- Small and mid-size businesses (six-figure contracts)
- Teams wanting a free trial or self-serve setup
- Simple FAQ bots
What's Great
- AOPs make complex procedures editable in plain language
- Strong roster of well-known enterprise customers
- Agents take real actions in back-end systems
- Detailed conversation analytics and testing
- Fast product development, backed by heavy funding
Watch Out For
- No published pricing, trial or self-serve option
- Contracts commonly run into the hundreds of thousands per year
- Per-conversation pricing charges for unresolved conversations too
- Young company, though growing fast
Pricing
View all features & details
Key Features
- Agent Operating Procedures in natural language
- Chat, email and voice agents
- Actions via API integrations
- Conversation analytics and insights
- Testing and simulation before launch
- Human escalation with full context
Integrations
- Zendesk, Salesforce, Intercom helpdesks
- Internal account and order systems
- Knowledge bases and help centers
- Telephony for voice
How It Compares
| Feature | Decagon | Sierra | Intercom Fin | Ada |
|---|---|---|---|---|
| Configuration | Plain-language AOPs | Agent SDK plus no-code | Guidance in plain language | No-code builder |
| Pricing model | Per conversation or resolution | Per resolution (outcome) | $0.99 per outcome | Per resolution |
| Public pricing | No | No | Yes | No |
| Voice | Yes | Yes | Yes | Yes |
| Valuation | $4.5B | $15.8B | Part of Intercom | $1.2B (2021) |
Is this your tool? Add a badge