Decagon Pricing Explained (2026): What It Costs and What Drives the Bill

Decagon Pricing Explained (2026): What It Costs and What Drives the Bill

Decagon publishes no price. There is no rate card, no free tier and no self-serve signup — decagon.ai/pricing returns a 404, checked again on 28 August 2026. But "no published pricing" is not the same as "unknowable": Decagon publishes an unusually detailed argument for how it prices, and procurement platform Vendr publishes what buyers actually signed. Put those together and you can budget for Decagon before the first sales call.

TL;DR:

• Published rate card: none. Every commercial path routes to a demo request.
• Pricing model: per-conversation by default, per-resolution offered as an alternative — Decagon says the vast majority of customers choose per-conversation.
• What buyers actually pay: median $432,750/year, range $105,000–$923,183 (Vendr, read 28 August 2026; sample size unpublished).
• The structural cost driver: you pay per conversation, so you pay for the ones the AI does not resolve.
• Effective cost per resolved ticket: roughly $2.01 at 10,000 conversations a month, falling to ~$1.60 at the volume implied by the median contract.
• The rule that decides your model: per-conversation beats per-resolution above a 66% resolution rate. Below it, the reverse.
• Reported discount lever: a buyer-reported "30% discount off list price for 1mil+ conversations."

Compiled by Alex Gomez, Sr. Analyst. Disclosure: this page is published by Aissist.io, which sells an agentic AI layer that competes with Decagon and publishes its own per-resolution rate. Every Decagon figure below is sourced to Decagon's own pages, to Vendr, or to named third parties, and labelled with which; our own figures are self-reported and marked as such. No vendor reviewed this page before publication. Sources read 28 August 2026.

How does Decagon pricing work?

Decagon offers two billing models and recommends one of them. Its pricing post (10 December 2024, by Bihan Jiang, Director of Product) sets both out in its own words:

And it says plainly which one wins: "the vast majority of our customers choose per-conversation pricing."

Decagon's reasoning is worth reading rather than paraphrasing, because it is a real position and not marketing fog. It argues per-conversation is predictable, and that it keeps incentives clean by focusing on conversation volume rather than parsing the definition of an outcome. Its blunt version: "You never want to be in a situation where you're arguing over what a 'resolution' is."

Its glossary page on resolution-based pricing goes further, listing that model's weaknesses: "Defining what a resolution is can be tricky, as not all cases end wrapped up in a bow," month-to-month variance that makes forecasting hard, and "vendors predisposed to push issues toward 'resolved.'"

That is a company publicly arguing against the pricing model its closest rivals use — including us. It is a fair argument, and we take the other side of it below. Both positions are self-interested; the volume and resolution rate in your own account decide which is right for you.

What do buyers actually pay for Decagon?

The median signed Decagon contract is $432,750 a year, ranging from $105,000 to $923,183. Vendr is a procurement platform whose figures come from contracts its buyers actually signed, which makes it the strongest cost evidence available for a vendor with no rate card.

Two caveats to carry. Vendr does not publish the sample size or date range, so this is transaction-derived data of unstated volume. And the numbers move: in June 2026 competitors citing Vendr reported a median near $386,000 with a ceiling around $590,000; our 28 August 2026 read shows $432,750 and $923,183 — a median up roughly 12% and a ceiling up roughly 56% in about ten weeks. If you see a confident "$386K" quoted as current Decagon pricing, it is a stale read of the same source.

The drift tracks the company. Decagon raised $250 million at a $4.5 billion valuation on 27 January 2026 and added "more than 100 new global enterprise customers" that fiscal year, including Deutsche Telekom, Block and Avis Budget Group (Decagon, vendor-claimed). Sacra estimates Decagon reached $100 million in annualized revenue in July 2026, up from $44 million at the end of 2025 (Sacra, independently reported). A vendor tripling revenue while selling to airlines and banks is not optimising for the buyer who wants a number on a web page.

What per-unit rates do people report?

Three different per-unit rates appear in print, and they do not agree. Decagon has never published any of them, so read the table below as the spread of what buyers and analysts have reported, not as a rate card.

Note the contradiction. Decagon says per-resolution is the higher rate, yet one reported contract shows $0.50 per resolution against $0.99 per conversation. Both can be true — rates are negotiated per deal and committed volume moves them — but it means no one outside a Decagon contract can hand you your number.

What drives your Decagon bill?

Three variables move a Decagon invoice more than the headline rate does.

1. Conversation volume, not resolution volume. Under per-conversation pricing your bill is set by how many conversations arrive, and improving the agent does not reduce it. Contrast that with an outcome-based rate, where a better agent costs more only when it does more work you wanted done.

2. Your resolution rate — but inverted. At a reported $0.99 per conversation and a 70% resolution rate, three in ten paid conversations still reach a human. You pay for the conversation and the agent who cleans it up. Effective cost per resolved conversation is the rate divided by your resolution rate: $0.99 at 70% is $1.41; at 50% it is $1.98.

3. The platform fee underneath. A platform fee reported around $50,000 a year sits below usage, and Decagon includes no helpdesk — so your existing helpdesk bill continues in parallel rather than being absorbed.

What does Decagon cost at my volume?

At the reported $0.99 per conversation plus a $50,000 platform fee, Decagon runs about $2.01 per resolved ticket at 10,000 conversations a month and about $1.60 at 32,000. The platform fee is the reason the number falls — it amortises across a bigger base. All three rows below assume a 70% resolution rate; the arithmetic is ours, run on the reported rates above.

The third row is the useful one. $430,160 lands within half a percent of Vendr's observed median of $432,750 — which says the median Decagon contract represents roughly 32,000 conversations a month, not 10,000. If you are at 10,000, do not budget from the median. If you are under 5,000, the $50,000 floor alone may exceed your entire automation budget.

Which is cheaper — per conversation or per resolution?

Per-conversation pricing wins above a 66% resolution rate; per-resolution wins below it. With the reported rates of $0.99 per conversation and ~$1.50 per resolution, the two cost exactly the same when the AI resolves 66% of conversations, because $0.99 ÷ $1.50 = 0.66. That single division is the whole decision.

• If your AI resolves more than 66% of conversations, take per-conversation, because you stop paying a premium on the resolutions you are already good at.
• If your AI resolves less than 66%, take per-resolution, because you stop paying for conversations that ended with a human anyway.
• If your resolution rate is still unknown — which is most first-time buyers — take per-resolution for year one, because it caps the downside while you find out.

The threshold moves with your quoted numbers. Run the same division on the two rates Decagon gives you and you have your own break-even, in about four seconds. Decagon's own argument against per-resolution pricing is strongest exactly when you are above that line, which is worth noticing.

How does Decagon compare on cost per resolution?

Effective cost per resolved conversation is the only figure that compares across pricing models, so that is the column that matters. All effective figures below assume the same 240,000 conversations a year at a 70% resolution rate.

Rates are drawn from our AI agent pricing benchmark, which tags every figure as published, reported or estimated and normalises 18 vendors to the same denominator. Our own $0.60 is published on our pricing page but is self-reported on our own definition of resolution and has not been independently audited — weigh it accordingly.

The honest summary of the trade: Decagon buys you forecasting certainty and pays for it with the cost of failure. If your volume is stable and your resolution rate is high and steady, per-conversation is defensible and Decagon's argument for it is a good one. If your volume spikes, or your resolution rate is still climbing, you are paying full freight for every conversation the agent could not finish.

What is not in the rate?

Four costs sit outside the per-unit rate, and they are where quotes get bigger than expected.

Implementation time. G2 buyers report two months to implement Decagon and four months to return on investment (G2, independently reported, read August 2026). Two months of internal project time is a real cost even when nobody invoices you for it.

Your helpdesk. Decagon layers onto your existing stack rather than replacing it, so Zendesk, Salesforce or Intercom seat costs continue alongside the Decagon contract. Compare it against a helpdesk-replacement product and you are comparing different totals.

The annual commitment. Contracts are annual by default, with quarterly payment reported as negotiable and Net 30 / Net 60 terms. There is no published free trial and no self-serve tier — Capterra lists Decagon's starting price as "Contact vendor," with no trial and no free version.

The spike month. Under per-conversation billing, a Black Friday or an outage bills at full rate on conversations the agent was never going to resolve. Model your worst month, not your average one.

Is Decagon worth it?

Decagon is worth its price for a large enterprise with high, predictable conversation volume and a dedicated internal owner for the agent. Its results are specific and real: Duolingo reports an 80% deflection rate, Chime 70% resolution across chat and voice, and ClassPass a 95% cost reduction (vendor-claimed, decagon.ai, read August 2026). Buyers rate it 4.7 out of 5 across 31 reviews on G2 (read August 2026). It is also one of the few vendors in this category you can budget for in advance, because it publishes its pricing logic and its contracts show up in procurement data.

For SMB and mid-market teams, the entry economics are the obstacle rather than the technology: a ~$50,000 platform fee before the first conversation, no trial, no self-serve, and an evaluation that starts with a sales cycle. If that describes you, our Decagon AI alternatives guide covers the field, and the full product assessment — capability, reviews, security, incident history — is in our Decagon AI review.

Want the same resolution depth without the platform fee? Aissist.io is up to $0.60 per resolution, published, with no platform floor and no annual commitment — and it runs on the helpdesk you already have. Get a free demo →

What to ask before you sign

Get seven answers in writing before you compare quotes. Most of what makes an AI agent contract expensive is not the headline rate.

• What is the per-conversation rate at my committed volume, and what is the rate above it? Volume tiers are where the negotiation lives — the buyer-reported 30% discount attaches to 1m+ conversations.
• What counts as a "conversation"? Get the definition in writing, including re-contacts, channel switches, and whether a conversation spanning days bills once or twice.
• Is the per-resolution option genuinely available to me, and at what rate? Decagon offers it. Price both and divide one by the other to find your break-even.
• What is the platform fee, and what does it include? Confirm whether it is separate from usage and whether it steps up with volume.
• What is the voice rate versus the chat and email rate? Channels rarely price the same.
• What happens in a spike month? Model Black Friday, not your average month.
• What is the trailing twelve-month uptime, and is there an SLA with credits? Ask this of any vendor in this category, us included.

A vendor that cannot define a billable conversation in one sentence will define it in the invoice.

FAQs

How much does Decagon cost?

Decagon publishes no rate card — decagon.ai/pricing returns a 404. Vendr, whose figures come from signed contracts, reports a median contract value of $432,750 per year, ranging from $105,000 to $923,183 (read 28 August 2026; sample size unpublished).

Does Decagon charge per conversation or per resolution?

Both are offered, but per-conversation is the default in practice. Decagon's own pricing post describes per-resolution as "a higher fixed rate for each fully resolved conversation, with no charge for escalations," then states that "the vast majority of our customers choose per-conversation pricing."

What is Decagon's platform fee?

Around $50,000 per year, reported independently by Quiq in March 2026 and The WITN in June 2026. The fee is reported to apply regardless of which usage model you choose, and it sits on top of per-conversation or per-resolution charges.

Does Decagon have a free trial or free plan?

No. There is no free trial, no free tier and no self-serve signup. Every commercial path routes to a demo request, and Capterra independently lists Decagon's starting price as "Contact vendor" with no trial and no free version.

Why do some articles say Decagon costs $386,000 a year?

Because they are quoting a stale read of the same Vendr page. In June 2026 Vendr's Decagon median read near $386,000 with a ceiling around $590,000; on 28 August 2026 it read $432,750 with a ceiling of $923,183. Any Decagon cost figure is only as good as its read date — ask for one.

What is the effective cost per resolution with Decagon?

Divide the per-conversation rate by your resolution rate, then add amortised platform fees. At $0.99 per conversation and 70% resolution, usage alone is ~$1.41 per resolved conversation; adding a ~$50,000 platform fee across 120,000 conversations a year brings it to roughly $2.01.

Which is cheaper, per-conversation or per-resolution pricing?

Per-conversation is cheaper above a 66% resolution rate; per-resolution is cheaper below it, using the reported $0.99 per conversation and ~$1.50 per resolution. Divide your quoted per-conversation rate by your per-resolution rate to find your own break-even.

How long does Decagon take to implement?

G2 buyers report two months to implement Decagon and four months to reach return on investment, read August 2026. Decagon deploys onto your existing helpdesk rather than replacing it, so your helpdesk costs continue in parallel.

Is Decagon cheaper than Intercom Fin?

Usually not, at mid-market volumes. At 240,000 conversations a year and 70% resolution, the reported Decagon per-conversation model works out near $1.71 per resolved ticket against roughly $1.11 for Intercom Fin including seats — though Fin replaces your helpdesk while Decagon layers onto it.

Why is Decagon so expensive?

Decagon sells to large enterprises — airlines, banks, telecoms — where implementation, security review and voice deployment carry real cost. It raised $250 million at a $4.5 billion valuation in January 2026, and its contract sizes reflect that segment rather than the mid-market.

Changelog

• 28 August 2026 — Updated. Added the 66% break-even rule between per-conversation and per-resolution pricing, a volume-tiered cost table (10k / 20k / 32k conversations a month), a normalised seven-vendor comparison, G2 implementation and ROI data, Series D and revenue context, and two original graphics. Corrected the Decagon quote on customer preference to its verbatim wording. Vendr figures re-read on this date, unchanged at $432,750 median.
• 28 August 2026 — First published. Verified on this date: decagon.ai/pricing returns 404; Decagon's pricing post and resolution-based-pricing glossary page are live and quoted directly.