Last updated: 2026-06-13
Decagon is a $4.5B AI concierge automating chat, voice, email and SMS for 100+ enterprise customers. Pricing starts near $95K/year plus a $50K platform fee.
Decagon is an enterprise AI customer support platform valued at $4.5 billion as of January 2026, automating chat, voice, email, and SMS for 100+ companies including airlines and banks. Pricing starts with a $50,000 annual platform fee plus usage, with typical contracts between $95,000 and $590,000 per year. Decagon Voice 2.0 adds sub-second latency phone support and persistent cross-channel memory.
Decagon is an enterprise AI customer support platform founded in 2023 by Jesse Zhang and Ashwin Sreenivas, valued at $4.5 billion after a $250 million Series D led by Coatue Management and Index Ventures in January 2026. The company calls its product an "AI concierge": autonomous agents that resolve customer issues end to end across chat, email, voice, and SMS rather than just pointing users to a help article. Decagon runs on a multi-model architecture, combining foundation models from OpenAI, Anthropic, and Cohere with its own fine-tuned models. As of March 2026, 80% of production traffic runs on Decagon's in-house models trained specifically on support conversations. A group of specialized agents cross-checks responses, with a supervisor model flagging hallucinations before a reply reaches the customer, and a persistent "user memory" layer carries context across channels so a chat conversation can continue directly in a phone call without the customer repeating themselves. Decagon Voice 2.0 handles inbound and outbound calls with sub-second latency, interruption handling, branded caller IDs, and customizable tone, supporting use cases like renewal calls, appointment reminders, and proactive outreach. The platform connects directly to Zendesk, Salesforce, Intercom, and Kustomer, giving agents full Customer 360 context to create cases, update opportunity stages, and tag tickets automatically. In one published case study, Decagon agents reached a 70% combined resolution rate across chat and voice. Decagon serves 100+ enterprise customers in airlines, banking, telecom, and retail, and is best suited to organizations processing 10,000+ support tickets per month with 50+ agents. Pricing is fully custom: a $50,000/year platform fee plus usage-based per-conversation or per-resolution charges, with deployments typically landing between $95,000 and $590,000 per year and a roughly six-week setup. There is no free tier or self-serve signup; the product is sold through enterprise sales with dedicated implementation support. In March 2026, Decagon launched Duet, an AI-assisted tool for drafting and refining the "Agent Operating Procedures" (AOPs) that define how its agents behave, aimed at cutting the manual work CX teams spend tuning agent logic. G2 reviewers note strong satisfaction with response quality and implementation speed, but flag limited visibility into why the AI made a specific decision and rudimentary role-based permissions for audit purposes.
Custom enterprise pricing only, no public rates and no free tier. A $50,000/year platform fee applies regardless of plan, on top of usage-based per-conversation or per-resolution pricing. Typical annual contracts range from $95,000 to $590,000+ depending on ticket volume, with median deals around $400,000/year. Implementation takes roughly 6 weeks.
Decagon is an enterprise AI customer support platform, founded in 2023 by Jesse Zhang and Ashwin Sreenivas, that builds autonomous AI agents to handle full customer interactions across chat, email, voice, and SMS. The company calls its product an 'AI concierge' because the agents resolve issues end to end, such as processing refunds or updating accounts, rather than just linking to a help article. Decagon reached a $4.5 billion valuation in January 2026 after a $250 million Series D round led by Coatue Management and Index Ventures. It now serves 100+ enterprise customers across airlines, banking, telecom, and retail. The platform uses a mix of foundation models from OpenAI, Anthropic, and Cohere alongside Decagon's own in-house fine-tuned models. As of March 2026, 80% of its production traffic runs on those in-house models trained specifically on support conversations.
Decagon does not publish pricing and sells exclusively through custom enterprise contracts. Every deal includes a flat $50,000 per year platform fee, regardless of which pricing model is chosen. On top of that fee, Decagon charges either per-conversation pricing, where every AI-touched interaction is billed, or per-resolution pricing, which has a higher per-unit rate but only bills for conversations the AI fully resolves. Based on market data, total annual contracts typically range from $95,000 to $590,000, with median deals landing around $400,000 per year. There is no free tier, no public self-serve checkout, and no monthly subscription option. Implementation usually takes around six weeks before an account goes live. Buyers should budget for usage growth, since the per-conversation and per-resolution rates have no baseline cap and can scale quickly with ticket volume.
Decagon's core feature is its AI concierge: agents that work across chat, email, voice, and SMS under one intelligence layer with persistent 'user memory', so a conversation that starts in chat can continue in a phone call without the customer repeating themselves. Decagon Voice 2.0 handles inbound and outbound calls with sub-second latency, interruption handling, branded caller IDs, and customizable tone, and can run outbound campaigns for renewals or appointment reminders. The Duet tool, launched in March 2026, uses AI to help CX teams draft and refine the 'Agent Operating Procedures' that define agent behavior. Decagon also connects directly to Zendesk, Salesforce, Intercom, and Kustomer, giving agents full Customer 360 context to create cases, tag tickets, and update opportunity stages. A supervisor model reviews other agents' responses to catch hallucinations before they reach customers.
No, Decagon has no free tier and no free trial available to the public. The platform is sold only through custom enterprise contracts that start with a $50,000 per year platform fee before any usage charges. This makes Decagon impractical for solo users, small businesses, or startups looking for a low-cost or free AI chatbot. Companies wanting to test AI customer support without a large budget commitment typically look at lower-cost alternatives such as Intercom Fin, which starts around $29 per seat per month, or other chatbot tools with self-serve free plans. Decagon's sales process involves a demo and a roughly six-week implementation period, which is standard for enterprise software but not suited to quick, no-commitment evaluation. Anyone evaluating Decagon should expect to go through a procurement and legal review process before signing.
Intercom Fin is a strong alternative if you want a much lower entry price, starting around $29 per seat per month instead of Decagon's six-figure annual contracts. Sierra is the closest comparable in approach, since it is also an LLM-native conversational AI platform built for end-to-end customer experience automation at enterprise scale, with a similar custom-pricing model starting around $95,000 to $150,000 per year. Zendesk AI is worth considering if your support team already runs on Zendesk and you want AI features bundled into that subscription rather than a separate platform. Ada and Forethought are also commonly evaluated by teams that want a faster, lower-touch chatbot rollout than Decagon's multi-week implementation. Teams with under 10,000 monthly tickets or under 50 support agents will generally find better value outside Decagon's enterprise-only pricing.
Decagon is best for large enterprises in industries like airlines, banking, telecom, and retail that process 10,000 or more support tickets per month and employ 50 or more support agents, since that volume is needed to justify the $50,000 per year platform fee plus usage costs. It suits organizations with dedicated procurement and legal teams that can manage a custom enterprise contract and a roughly six-week implementation. Companies that already run Zendesk or Salesforce benefit most, since Decagon's deepest integrations give agents full Customer 360 access and automatic ticket tagging. Decagon is not a good fit for startups, solo founders, or small support teams without a dedicated Agent Engineer to build and maintain its Agent Operating Procedures. It is also a weaker fit for companies that need detailed audit logs or granular permission controls for compliance, since G2 reviewers note these areas are still rudimentary.
Decagon connects to a customer's existing stack through direct API integrations rather than offering a standalone public developer API for building new applications. It has confirmed direct integrations with Zendesk (knowledge base, ticketing, and Sunshine-based escalation), Salesforce (full Customer 360 access, case creation, and opportunity updates), Intercom (email routing and chat), and Kustomer (knowledge base sync). Voice channels integrate with Amazon Connect, RingCentral, and SIP trunking, while knowledge sources can include Confluence, Contentful, Guru, Slack, and custom databases. E-commerce integrations with Shopify and Stripe support order lookups and payment processing. There is no public documentation describing Model Context Protocol (MCP) support, so teams that specifically require MCP-based tool orchestration should confirm directly with Decagon's sales team before purchasing.
Decagon and Sierra are the two platforms most often compared head to head, since both are LLM-native AI agent platforms aimed at enterprise customer experience automation with custom, six-figure annual pricing. Decagon's main differentiator is its cross-channel 'user memory' and Decagon Voice 2.0, which lets a single conversation move between chat, email, voice, and SMS without losing context, plus deep native write-access into Salesforce and Zendesk. Sierra is generally positioned around its own agent-building workflow and has its own set of enterprise integrations and pricing starting in a similar $95,000 to $150,000 per year range. Neither platform publishes detailed public pricing, so the real decision usually comes down to a sales evaluation of each vendor's voice quality, integration depth with your specific helpdesk or CRM, and how transparent each platform is about why its agents make the decisions they do. Teams should request side-by-side pilots before committing to either.