Decagon pricing, per conversation or per resolution

Decagon publishes no price. Its site has no pricing page, and the way in is a Get a demo form, so your figure comes from its sales team. What Decagon does publish is how it charges. Its post on pricing AI agents offers two models, a fixed rate for every conversation or a higher rate for each conversation its AI resolves, with the rate tied to your volume. We read Decagon’s own pages and Fin’s pricing page on 25 September 2026, and each claim here links to where we read it.

Deacon is ours, and it’s built for a founder who wants AI answers without a sales call. It’s an AI customer support agent that answers your customers in seconds from your own docs, then shows you what they’re stuck on and what they want next. There’s a free plan with every feature, and it needs no card.

How Decagon charges, in its own words

Decagon’s Director of Product set out its pricing in Pricing the AI Agent Economy, posted on 10 December 2024 and still on its blog. The post argues that seats suit software people use, while an AI agent does the work itself, so it should be measured against the human labour it takes on. It judges an agent by resolution rate, the share of conversations the agent resolves without a person, and by customer satisfaction. Decagon says it supports two pricing options.

• Per conversation. A fixed rate for every incoming conversation, with flexible pricing at higher volumes.
• Per resolution. A higher fixed rate for each conversation the AI fully resolves, and no charge when it escalates to a person. Committing to more resolutions lowers the rate.

Most of Decagon’s customers choose per conversation, the post says. Its reasons are a bill that’s predictable because it follows volume, and no arguments over what a resolution is. The case it has in mind is a frustrated customer who stops replying and gets counted as resolved.

Decagon’s glossary entry on resolution-based pricing describes the second model in general terms. You pay only when the AI resolves a conversation from start to finish, each resolution costs more than plain usage pricing would, and volume discounts are often available. It names the trade-offs too, that a resolution can be hard to define and that the bill can vary from month to month, and it says the model suits high-volume support.

Neither page gives a rate, and nor does any other page we read on decagon.ai. So any Decagon figure you see elsewhere didn’t come from Decagon’s own pages.

What each model means for your bill

Both of Decagon’s models bill a rate for a count. What changes is what gets counted, and so what makes the bill go up.

A per-conversation bill follows your traffic. A launch week that brings twice the conversations raises the bill with it, whether the AI solved them or passed them to your team. A per-resolution bill follows the AI’s results, so it climbs as the agent gets better even when your traffic is flat. It also rests on where a resolution ends, and Decagon’s glossary says those grey areas can lead to billing disagreements.

Fin also bills by outcome, and it publishes its rate, $0.99 an outcome in support, charged at most once a conversation. Zendesk AI pricing does the same sums for Zendesk’s automated resolutions.

Who Decagon is built for

Decagon sells to large companies, and says so. Its homepage describes “concierge experiences for the world’s leading enterprises”, and its About page names Avis Budget Group, Chime, Oura Health, 1-800-Flowers.com and Hunter Douglas. Its customer stories add Duolingo, Notion, Rippling, ClassPass, Perplexity and Hertz, among others.

In January 2026 Decagon raised $250 million at a valuation of $4.5 billion. It said more than 100 new global enterprise customers had joined it in its last fiscal year, among them Avis Budget Group, Block and Deutsche Telekom.

Its AI agents work across chat, voice and email, and its About page adds SMS. A team builds them from Agent Operating Procedures, workflows written in plain language, then tests them and runs experiments on them as they change.

The demo form shows the size of deal Decagon expects. Alongside your email, name and company, it requires your monthly support tickets, picked from four bands or “Not sure”. The smallest band is less than 9,999 tickets a month, and the others are 10,000 to 49,999, 50,000 to 249,999 and 250,000 or more.

What buying Decagon involves for a small team

Buying Decagon starts with its sales team, and three things follow from that.

• There’s no plan to start on. The site offers Get a demo and Sign in, with no sign-up, and we found no free plan or trial on its pages.
• The price is a rate you agree. Decagon ties it to your volume, and on per-resolution pricing to how many resolutions you commit to.
• It’s a partnership, in Decagon’s words. Its pricing post talks of long-term partnerships, and its agents are built from procedures your team writes and tests.

That suits a company with a support department and a budget line for it. If you answer support yourself, it’s a lot to arrange before you know whether an AI agent helps at all. Deacon vs NiCE and Deacon vs Five9 look at two more platforms built for large support teams, and what a founder needs instead.

Deacon, AI answers you can start on today

Decagon is built for enterprises, with a support team behind the AI and a rate agreed with sales. If support is you, answering the same questions between releases, you want those questions handled this week, without a demo or a commitment. That’s who Deacon is for.

Sign up on the free plan and point Deacon at your help pages, or add files, pasted text and answers you write. Then paste one script tag into your site, or hand Claude Code, Cursor or Codex the prompt that adds it and checks it’s live. There are no flows or decision trees to build. Deacon answers your customers in seconds, at any hour and in their own language, from your content. When your content doesn’t cover a question, it says it doesn’t know rather than guessing, and asks for the visitor’s email.

When a visitor leaves their address or asks for a person, you get an email. You reply from Deacon’s dashboard, and if they still have the chat open, your reply appears there under your name.

A question Deacon couldn’t answer waits for you with an Answer this button. What you write becomes part of what Deacon knows, and Save & check shows the reply it now gives, so Deacon needs you less the more you answer.

Deacon also shows you what your customers need. Topics group their questions, count how many people asked and label most of them as a how-to, a bug report, a feature ask and so on, on every plan.

Deacon’s price is on the pricing page, and it doesn’t depend on what counts as resolved. An answer is one reply Deacon writes on your live site, including one saying it doesn’t know. Free gives you 50 answers a month, with no card and no end date. Starter is $50 a month, or $40 a month billed yearly, for 1,000 answers a month. Growth is $150 a month, or $120 a month billed yearly, for 4,000. Every feature is on every plan, nothing is billed past your plan’s answers, and no plan upgrades itself.

Questions about Decagon pricing

Does Decagon have a free plan or a free trial?

Not that its site shows. When we read it on 25 September 2026, decagon.ai had no pricing page, no sign-up and no trial, and its buttons led to Get a demo or Sign in. Deacon’s free plan needs no card, has no end date and includes every feature.

How much does Decagon cost?

Decagon doesn’t say. Your rate comes from its sales team after a demo, and its own pages tie that rate to the model you choose and your volume. Chatbot pricing sets out what the tools that do publish a price charge a small team.

Is Decagon priced per resolution?

It can be. Decagon offers per-resolution pricing, with no charge for a conversation handed to a person, and per-conversation pricing, which counts every conversation. Its pricing post says most of its customers choose per conversation.

What are the alternatives to Decagon for a small team?

Look for an AI agent with its price on the page and a plan you can start yourself. The best AI agents for customer service compares them with their prices, and the best AI chatbots for customer service ranks chatbots for small teams.

Can Deacon answer calls and emails, or take actions, like Decagon?

No. Deacon is a chat widget on your website, with no phone line, email inbox or SMS channel. It takes no actions in other systems, so it won’t change a booking or issue a refund, and it declines questions about a visitor’s own account, invoice or data. Decagon’s agents work across chat, voice and email.

What happens when Deacon’s answers run out?

Nothing more is billed. Every owner gets an email when the month’s answers reach 80% and again at 100%. After that, visitors get a short reply saying Deacon can’t look into it right now and are asked for their email, unless they’ve already left it, and you can still reply to every conversation from the dashboard.