Decagon: What It Actually Does, What It Costs, and Who It Fits
An agentic AI customer support platform built to autonomously resolve complex, multi-step support conversations across chat, email, voice, and SMS, sold through custom enterprise contracts rather than published per-seat pricing.
Decagon is built for enterprises that want an AI agent to fully own and close out complex support conversations across channels, not just answer simple FAQs in a website widget. It fits companies with real engineering resources to integrate it deeply and a support volume large enough to justify a six-figure annual contract. It's not the right fit for a small team that wants transparent, published pricing or a lighter chat add-on bolted onto an existing helpdesk.
What Decagon Is For
Decagon is an agentic AI customer support platform, not a chat widget with an AI feature bolted on. It's built to autonomously resolve full support conversations, including multi-step actions like processing refunds, updating orders, or troubleshooting account issues, across chat, email, voice, and SMS, rather than simply answering questions and handing everything else to a human. Founded in August 2023 by Jesse Zhang and Ashwin Sreenivas, the company raised a $250 million Series D in January 2026 that tripled its valuation to $4.5 billion, and it reported crossing $100 million in annualized revenue in July 2026.
It's built for enterprises, particularly AI-native SaaS, fintech, and consumer tech companies, that have the engineering resources to integrate it deeply into their existing systems and the ticket volume to justify a large annual contract. Decagon is used by more than 100 companies including Notion, Duolingo, Rippling, Substack, Eventbrite, Bilt, Oura, Affirm, and Chime, alongside larger enterprise logos added through 2025 such as Avis Budget Group, Mercado Libre, and Deutsche Telekom. A small team without dedicated engineering support, or one that just wants a chat widget with AI answers, is generally better served by a lighter, self-serve product.
Key Features
- Agentic resolution of full support conversations, including multi-step actions like refunds, order changes, and account updates, not just answering questions
- Deployment across chat, email, voice, and SMS from a single platform rather than a single-channel widget
- Structured, deterministic workflow logic that CX teams can update directly with shorter feedback loops and less engineering dependency once deployed
- Deep integration work with existing support and backend systems, tailored per enterprise deployment
- Usage-based pricing structured around per-conversation or per-resolution outcomes rather than flat per-seat fees
- Analytics on resolution rate and conversation outcomes built for enterprise reporting needs
Pros & Cons
Pros
- ✔ Built specifically for autonomous, multi-step resolution rather than a support-suite AI add-on
- ✔ Strong enterprise traction and funding, with a large and growing named customer base as of 2026
- ✔ CX teams reportedly get faster, more direct control over agent logic once integration work is done
Cons
- ✕ No published pricing; every deal requires a sales process, and third-party estimates suggest six-figure annual minimums
- ✕ Meaningful engineering lift to integrate deeply, particularly where core systems (like Salesforce Service Cloud) need custom connector work
- ✕ Built with an AI-native SaaS and tech focus, so less out-of-the-box breadth for industries with heavier legacy system requirements
Pricing
| Component | Price (approx.) | What it covers |
|---|---|---|
| Platform fee | Reportedly around $50,000/year minimum | Base access to the platform regardless of usage-based pricing model chosen |
| Per-conversation or per-resolution usage | Reportedly in the ballpark of $0.50+ per resolution in some negotiated deals | Usage-based fees layered on top of the platform fee, billed per interaction or per successfully resolved conversation depending on contract terms |
| Typical total annual contract | Third-party deal data suggests a median in the mid-hundreds-of-thousands range, commonly cited around $95,000 to $590,000+/year | Combined platform fee plus usage at typical enterprise support volumes |
Decagon does not publish pricing. These figures are drawn from third-party marketplace and deal-tracking sources (not confirmed by Decagon directly) current as of late 2026, and actual contract terms vary significantly by volume, channel mix, and integration scope.
Pricing is entirely quote-only and negotiated per enterprise contract; the figures above are third-party estimates, not confirmed list prices, and should be treated as directional only. Confirm current terms directly with Decagon's sales team before budgeting.
Common questions
Intercom and Zendesk are broad customer support suites, shared inbox, ticketing, help center, with an AI agent layered on as an add-on feature inside that suite. Decagon is a dedicated agentic AI platform built specifically to autonomously resolve full, multi-step support conversations across channels; it isn't a helpdesk with a chatbot bolted on, and most enterprise buyers deploy it alongside an existing ticketing system rather than as a replacement for one.
Decagon does not publish pricing. It's sold through custom enterprise contracts, typically combining a flat annual platform fee (third-party sources put this around $50,000/year at minimum) with usage-based fees per conversation or per resolved outcome. Total annual contracts commonly land in the mid-six-figure range based on third-party deal data, though actual pricing depends on volume and negotiation.
Enterprises with real engineering resources to integrate it deeply and enough support volume to justify a large contract. Its customer base leans toward AI-native SaaS, fintech, and consumer tech companies like Notion, Duolingo, and Chime, alongside larger enterprise logos like Deutsche Telekom, though it also serves companies further from that core profile.
Not natively as of this writing; teams whose account, case, or customer relationship data lives primarily in Salesforce Service Cloud reportedly need custom connector work to bring that data into Decagon. This is worth scoping directly with Decagon's team before signing a contract if Salesforce is central to your support stack.
Both are fully agentic AI customer support platforms sold through custom enterprise contracts, but they lean toward different strengths: Decagon is generally associated with structured, deterministic workflow control and faster CX-team-driven iteration on agent logic, while Sierra is generally associated with voice-first, multi-channel deployment across a broader spread of industries. Neither difference is absolute, and both should be evaluated against your own channel mix and integration needs.
Considering Decagon for autonomous, multi-step support resolution?
We'll walk through your ticket volume, channel mix, and existing systems (including whether Salesforce or another CRM needs to connect) before recommending whether a fully agentic platform like Decagon is worth the enterprise contract, or whether a lighter option fits better.
Ready to get started?
We usually reply within 24 hours.