Mike HeapMike is an experienced Product Manager who focuses on all the “non-development” areas of My AskAI, from finance and customer success to product design, copywriting, testing and more.
Table of Contents
- How does Decagon pricing actually work?
- Per-conversation vs per-resolution: which model does Decagon put you on, and what does each count?
- What does Decagon cost at typical volumes?
- At 1,000 and 10,000 tickets a month
- At 50,000 tickets a month
- What the Decagon pricing page doesn't tell you (the hidden costs)
- You still need a helpdesk
- Usage volatility with no ceiling
- The enterprise floor is itself a cost
- Agent Assist is Zendesk-only
- There's nothing to test before you buy
- What real Decagon customers actually pay (and why the invoices stay private)
- Does Decagon have a free trial, and what are the contract terms?
- How does Decagon pricing compare to the alternatives at the same volume?
- Is Decagon actually worth the money?
- FAQs
Created time
Jul 12, 2026 04:19 PM
Title length (<60)
Author
Last optimised
Ecomm?
Tags
Image
Publish date
Jul 13, 2026
Video
Slug
decagon-pricing-explained
Featured
Type
Article
Ready to Publish
Ready to Publish
Decagon publishes no price. The pricing page is a 404, and the homepage sends you to a demo. Real deals are enterprise-only annual contracts in the low-to-mid six figures, on usage-based terms, with no seat fees. The cost that surprises people is the sales-gated minimum, the separate helpdesk underneath, and a definition of "resolved" that Decagon controls.
OK, so you went looking for what Decagon costs, and where the pricing page should be, at
decagon.ai/pricing, you got a 404. The homepage just sends you to "Book a demo."Every third-party breakdown quotes a confident dollar figure. None of them can tell you where that number came from, because (I checked) Decagon has never published one.
Decagon sells enterprise-only by design: annual contracts negotiated per deal, with the number living behind a sales cycle.
Which is fine if you're an enterprise with a procurement team and six figures of AI budget to spend. It's maddening if you're a support leader trying to work out, before you give up an afternoon to a discovery call, whether this thing is a $30k tool or a $500k one.
I've reconstructed Decagon's real cost from the only citable sources that exist: Decagon's own pages, independent contract data from the Vendr marketplace, and Sacra's research estimate. That covers the two billing models Decagon uses and what each one counts, the math where it can fairly be modeled (I'm blunt about the volumes where Decagon won't quote you), the hidden line items that never make it into a demo, and what the alternatives cost at the same volume.
I'm Mike, co-founder of My AskAI. We help 200+ ecommerce and SaaS teams run AI customer support inside their existing helpdesk, so I spend most of my week inside these pricing models, including the ones that hide the number. Decagon is one of the hardest in the category to pin down, and that difficulty is most of the story.
How does Decagon pricing actually work?
TL;DR: Decagon is enterprise-sales-only with no public pricing. It bills usage-based. That's either per conversation (every incoming chat, resolved or not) or per resolution (only fully-resolved conversations, at a higher rate), with no per-seat or per-agent fees. Contracts are annual. Vendr's real-contract data runs roughly $105K to $923K a year, median ~$433K.
Decagon publishes no pricing. The
decagon.ai/pricing URL returns a 404 (I checked it live while writing this), and the homepage is a demo/contact-sales flow, with no self-serve plan, no tiers, no "starting at." Everything is quoted after a call.
What Decagon does publish is how the model works. In its own post on how AI agents should be priced, Decagon lays out two options it offers customers. Per-conversation is a fixed rate for every incoming conversation, resolved or not (so the chats it misses count too), "with flexible pricing for higher volumes."
Per-resolution is a higher fixed rate charged only when the AI fully resolves a conversation, with no charge for escalations to a human. Decagon adds that "the majority of our customers gravitate towards per-conversation pricing", which means most of its customers pay for the chats the AI couldn't handle too.
The one thing Decagon doesn't do is charge per seat. There are no per-agent fees at all. It's entirely usage-based.
That's a genuine point of difference from the per-seat helpdesk AI products (Zendesk AI, Intercom Fin), where the seat license stacks on top of the AI charge, and I think it's a fair thing to like about the model even if the total is bigger.
For the actual dollars, the only credible independent anchor is Vendr, which aggregates real purchase data. As of this writing its Decagon marketplace page shows a median annual contract of $432,750 and a range of $105,000 to $923,183, with the model described as per-conversation, variable on volume.
Separately, Sacra estimates Decagon's per-resolution rate at roughly $1.50 per resolution ("~10% the cost of human support agents"). That's Sacra estimating, though. Decagon has never published this rate, so treat it as directional.
Component | Price | How it's billed | Notes |
Platform / contract | Custom (annual) | Annually, sales-negotiated | No public rate. Vendr's real contracts: floor ~$105K, median ~$433K, top ~$923K/yr |
Per-conversation | Sales-quoted rate | Per incoming conversation, any outcome | The model most Decagon customers land on, with discounts at higher volume |
Per-resolution | Higher sales-quoted rate | Only when the AI fully resolves | No charge for escalations. Sacra estimates ~$1.50/resolution, a directional figure Decagon has never published |
Seats / agents | $0 | n/a | No per-seat or per-agent fees, a real contrast with per-seat helpdesk AI |
Channels (chat/email/voice/SMS) | Included in platform fee | n/a | Whether voice carries an extra per-minute charge is unconfirmed. Decagon doesn't say |
The contrast that made me want to write this: you can read My AskAI's price, $0.10 a ticket, on the page, without talking to anyone. Our tiers are public: Pro at $199/mo (1,000 credits included), Scale at $499/mo (2,000 credits), Enterprise from $999/mo, with usage a flat ~$0.10 per credit on Scale and up and no per-seat fee.
With Decagon, the number is a conversation you have to earn your way into. Neither is automatically right for you, but one of them lets you do the math tonight.
Per-conversation vs per-resolution: which model does Decagon put you on, and what does each count?
TL;DR: Per-conversation bills every incoming chat regardless of outcome. Per-resolution bills only the fully-resolved ones, but at a higher unit rate. The catch: Decagon decides algorithmically what counts as "resolved," and that's exactly where billing disputes start.
The two models sound similar until you follow the money, and the difference is where buyers pick the wrong one. Under per-conversation billing, every incoming conversation is billable: the ones the AI nails and the ones it fumbles and hands to a human alike. I've watched teams sign per-conversation without clocking that the misses bill too.
Under per-resolution billing, you only pay when the AI fully resolves something, but the per-unit rate is higher, and escalations are free. On paper, per-resolution looks like the buyer-friendly one: pay only for wins.
Decagon concedes the slippery part itself. In its glossary entry on resolution-based pricing, it grants that "defining what a resolution is can be tricky, as not all cases end wrapped up in a bow… Gray areas can lead to billing disagreements."
That is the vendor telling you, in writing, that the unit it bills you on is fuzzy. What counts as resolved is set by Decagon's own "AI resolved" tag (their algorithm, decided on their side), a decision the buyer can't independently audit. When the thing that determines your invoice is a definition you don't control, "pay only for wins" just becomes "pay for whatever they score as a win."
The vocabulary is a second place buyers get caught, because vendors use these words interchangeably when they shouldn't. A conversation is an incoming chat.
Deflection is the AI keeping a customer away from a human, which is not the same as solving their problem. A resolution is (supposedly) a solved case. An outcome is whatever the vendor decides to bill on.
Decagon's own marketing leans on an 80% deflection figure, but deflection isn't resolution: a high deflection rate can hide customers who gave up rather than got helped. If you're being quoted on a per-resolution model, make them define, in the contract, exactly which of those four things you're paying for.
Under a per-resolution model, your bill rises as the AI gets better, and it shows up at renewal.
Understand this before you sign one. When I talk to teams about it, I get them to ask two questions: what your resolution rate will be on day one, and how much it will change over the next 12 to 24 months.
If you start at 20-30% resolution and climb toward the 60%, 70%, 80% you're hoping for, your costs on a per-resolution model can double or triple. And most of that climb comes from your work: updating knowledge, connecting tools and APIs, tuning guidance, running weekly QA. A per-resolution model charges you more for improvements you paid for with your own effort.
(If you want the field numbers behind those resolution rates, we keep a running benchmark in our AI resolution-rate benchmarks. The median across ~195 rated deployments sits around 70%.)
This is the whole reason My AskAI prices per ticket. You pay a flat $0.10 whether or not the AI resolves the ticket, which sounds worse until you run it forward: because the bill doesn't move with the resolution rate, your effective cost per resolved ticket falls as the AI improves, and you keep the upside of your own work while the vendor never claws it back.
We count a conversation resolved simply when the AI handled it without escalating to a human, and we make escalation dead easy so we're never pretending we know an issue was truly solved. If you ever want to see why the AI answered a given ticket the way it did, or which knowledge source it used, you can ask Echo inside the dashboard, where there's no black-box "resolved" tag you can't see behind.
What does Decagon cost at typical volumes?
TL;DR: Decagon won't quote you at 1,000 or 10,000 tickets a month. Its floor is enterprise. Where it can be modeled at all (Vendr's contract bands plus Sacra's ~$1.50/resolution estimate), a large deployment lands in the low-to-mid six figures a year. My AskAI is transparent, and buyable, at every volume.
A pricing post is meant to nail the three-volume worked example: 1k / 10k / 50k tickets a month, with the effective cost per resolved ticket at each. I'll run it.
But for Decagon, two of those three rows can't fairly be filled with a Decagon number. Those two blanks are the finding here.
First, the guardrails. Decagon has no rate card, so I'm not going to invent per-volume prices or take Vendr's dataset maximum ($923,183) and drop it into a "50k tickets" row as if it were a per-volume quote.
That maximum is just the top of a whole-dataset range across many contracts. Nobody was quoted it at 50k.
What I can use is Vendr's figures as bands (floor ~$105K, median ~$433K, top ~$923K a year) and Sacra's ~$1.50/resolution as a clearly-labeled estimate for a sensitivity check. For the resolution rate I anchor to the ~70% field median from our benchmark corpus (195 rated deployments across roughly 55 vendors; it's a directional aggregate, so treat it as a reference point).
Decagon's own G2 Ticket-Resolution score is 7.9/10, the weakest of its category scores, and its headline 80% is a deflection number. So assuming 70-80% resolution for the model is a generous read, well beyond conservative.
Volume | My AskAI cost | Decagon (modeled / where sold) | Effective $/resolved |
1,000 tickets/mo | ~$199 (Pro) | Below Decagon's enterprise floor, so it won't quote this volume | My AskAI ~$0.28 @70% |
10,000 tickets/mo | ~$1,299 (Scale) | Still below the typical enterprise floor. If quoted at all, a six-figure annual minimum dwarfs the usage | My AskAI ~$0.19 @70% |
50,000 tickets/mo | ~$5,000 (Enterprise) | Illustrative only. Vendr band, upper half of $105K to $923K/yr. Sacra sensitivity ~$630K to $720K/yr (an estimate, never a published rate) | My AskAI ~$0.14 @70% |
At 1,000 and 10,000 tickets a month
Plenty of write-ups will quote you a Decagon price at 1,000 tickets. Decagon's own positioning says it won't take that deal.
It's built for organizations doing tens of thousands of conversations with six-figure budgets, and a team at 1,000 or even 10,000 tickets a month sits below the volume where it will engage. At those volumes it's below the floor and unsold, with no dollar figure to give.
My AskAI, by contrast, is a real answer at these volumes: a 1,000-ticket team starts on Pro at $199 a month today, no sales call, no floor.
At 50,000 tickets a month
This is the first volume where a Decagon conversation makes commercial sense, and even here I'll give you a band and no single point. A deployment this size lands somewhere in the upper half of Vendr's observed $105K to $923K range.
As a sensitivity check, Sacra's ~$1.50/resolution estimate against ~50,000 conversations a month at 70-80% resolution works out to roughly $630K to $720K a year. Notice that sits high relative to Vendr's ~$433K median, which is why I treat the Vendr band as the anchor and Sacra as a stress test, never a quote.
What I won't do is tell you "$720K is what Decagon charges," because nobody outside a Decagon contract can say that. My AskAI at 50,000 tickets is ~$5,000 a month, flat, and you can read that off the page.
What the Decagon pricing page doesn't tell you (the hidden costs)
TL;DR: The sticker is only half the story, and there's no sticker anyway. The real gotchas are the enterprise floor, the helpdesk you still have to run underneath, a usage bill that spikes with seasonal volume, and an "Agent Assist" that only works on Zendesk.
There's no pricing page to be quiet about the line items, but the demo is. Five of them land mid-deployment, when the budget's already committed.
You still need a helpdesk
Decagon isn't a ticketing system. It has no standalone human inbox, so it layers on top of and hands off to Zendesk or Salesforce, which means Decagon's bill is additive to whatever you already pay for the helpdesk underneath it.
Reviewers describe adopting Decagon as migrating your setup around it. When you model total cost, Decagon's bill sits on top of whatever you already pay for the helpdesk underneath (I'd budget for both from day one).
Usage volatility with no ceiling
Because you're billed per conversation (or per resolution), your cost moves with your volume, and support volume is seasonal and spiky. Teams flag unpredictable costs during seasonal spikes: a product launch or a Black Friday surge stretches the invoice as much as the team, with no flat ceiling to cap it. If you need a forecastable line item, a usage meter with no cap is the opposite of that.
The enterprise floor is itself a cost
The ~$105K+ annual minimum buys you admission and not much more. On top of the contract sits roughly six weeks of white-glove onboarding and a dedicated implementation engagement (Decagon runs a hands-on, forward-deployed model), plus the enterprise security and compliance, SOC 2 Type II and the rest, that a six-figure buyer expects and pays for. The "price," in other words, bundles a sales cycle and an implementation lift that a mid-market or SMB team can't easily absorb, in either money or calendar time.
Agent Assist is Zendesk-only
Decagon's Agent Assist capability is restricted to Zendesk, and non-Zendesk teams have flagged frustration at paying for a platform whose assist features they can't fully use. If you're on Salesforce or anything else, that's a gap I'd price in before you sign.
There's nothing to test before you buy
Decagon publishes no docs site and no self-serve entry point, so knowledge configuration happens inside the gated onboarding. To me the real cost sits in the variable that most determines your invoice: how much of your volume the AI handles. You can't read that in advance, and there's no way to scope your likely spend, or measure your own resolution rate, before you're on a contract.
What real Decagon customers actually pay (and why the invoices stay private)
TL;DR: There's almost no public Decagon billing chatter, and the reason is structural. Decagon sells enterprise-only, under annual NDA contracts, so the invoice never becomes a public artifact in the first place.
If you go looking for the genre that every self-serve vendor generates, the Reddit and G2 threads where customers vent about a surprise invoice, you won't find it for Decagon. That absence has a structural cause.
It comes down to who Decagon sells to (and I don't read the silence as a warning sign). When every customer is an enterprise on a negotiated annual contract signed under an NDA, the invoice never becomes a public thing.
Nobody screenshots a six-figure procurement agreement into a subreddit. Decagon has around 18 reviews on G2 and effectively no presence on Reddit, Capterra or the other places support buyers compare bills.
Nothing's being hidden. Enterprise software bought this way just doesn't leave that trail, and I've stopped expecting it to.
In this category the cheaper, transparent alternatives publish a rate on a page, while Decagon asks you to sign before you see one.
Where a figure can be sourced, it comes with its provenance attached. A G2 reviewer reported a cost-per-ticket moving from $78.43 to $73.92 across a partial first quarter. That's G2-reported and never invoice-verified, a single reviewer's number, so read it as one data point, nothing more.
On the partner and vendor side, TaskUs has publicly cited a 25-50% cost reduction from its Decagon partnership, and Decagon's own marketing quotes a $1.75M cost reduction at Bilt Rewards (we read vendor-published savings as a ceiling and discount hard from there). Both of those are vendor- or partner-published outcomes with no neutral invoice behind them. Weigh them accordingly.
For what it's worth, from my own time shopping this category: word-on-the-street is that Decagon (and Ada) sit at the expensive end. I've heard per-resolution numbers closer to $2-3, higher than Intercom Fin's $0.99.
That's anecdotal and no kind of official Decagon rate. But it lines up with the whole thesis of this post: you can't get a real number without signing, so the numbers that circulate are all second-hand.
Does Decagon have a free trial, and what are the contract terms?
TL;DR: No free trial, no free plan, no self-serve signup. The only path is demo → discovery → custom quote → annual contract. Quarterly payment is negotiable, and discounts of up to ~30% have been reported for very large (1M+ conversation) commitments.
There is no free trial and no free tier. The only door into Decagon is a demo booking, and what's behind it is a discovery call, a custom quote, and an annual contract.
You'll see the occasional third-party claim (on some aggregator sites) that Decagon offers a free trial. I couldn't corroborate it against Decagon's own site, so I wouldn't rely on it. Decagon does run a "Decagon University" sandbox, but that's an education resource for existing customers and no way to trial the product before buying.
On terms: contracts are annual by default, with quarterly payment reported as negotiable, and volume discounts of up to roughly 30% off list for commitments north of a million conversations. All of which is normal for enterprise software. It's just gated behind a sales process, so "what does it cost" and "can I try it" have the same answer: book a call.
For contrast, because "can I test it first?" is the most reasonable question a support leader can ask: the way we handle it at My AskAI is a 30-day free trial with every feature unlocked, unlimited tickets, and no credit card. You run the AI on your real tickets before you pay a penny (you do sign up with a business email), new customers commonly get 50% off their first three months, and the tiers are self-serve and published, so you can start today and read the price off the page. Decagon asks you to trust the model and negotiate the number; we ask you to prove it on your own tickets first.
How does Decagon pricing compare to the alternatives at the same volume?
TL;DR: At any volume a mid-market team can actually buy, Decagon sits out of the running as an enterprise floor, never a competitor at 10k tickets. Against transparent per-ticket pricing, My AskAI models at ~$1,299/mo for 10,000 tickets, and Decagon won't quote that volume at all.
Decagon won't sell at 10,000 tickets a month, so lining it up against everyone else at that volume isn't quite a fair fight, and I haven't invented a number to fill the gap. Here's the 10k picture with that said up front:
Vendor | Pricing model | Cost at 10k tickets | Effective $/resolved | Notes |
Decagon | Per-conversation/resolution, enterprise-only | Not sold at 10k (enterprise floor) | n/a | ~$105K+/yr minimum |
My AskAI | $0.10/ticket flat (Scale) | ~$1,299/mo | ~$0.13 | Transparent, self-serve |
Intercom Fin | $0.99/resolution + seats | ~$7,425/mo | ~$0.99 | Per-resolution, costs rise as the AI improves |
Zendesk AI | Add-on + Suite seat | ~$11,250/mo | ~$1.13 | Seat-stacked |
At a volume a mid-market team can put on a card, the transparent per-ticket vendors are an order of magnitude apart from the enterprise floor, and Decagon isn't even on the field. Where My AskAI lands at ~$1,299 a month for 10,000 tickets, Decagon's minimum annual commitment alone would swamp it several times over, and you'd have to sign to find out by how much.
Enterprise scale is where the picture flips. If you operate there (100k+ conversations, complex voice needs, a Zendesk or Salesforce stack you're building a CX-automation layer on top of, a procurement team that prefers one big contract), Decagon is a real, capable product, and this becomes a question of fit and not of price.
Decagon is built for a specific buyer. If you're not that buyer, you want a model you can read on the page and test before you pay.
You can see the cheaper-at-your-volume options in the Decagon alternatives round-up, and the full My AskAI numbers in our pricing explainer.
Is Decagon actually worth the money?
TL;DR: Worth it for enterprises with 50k+ conversations, an existing Zendesk/Salesforce stack, a six-figure AI budget and complex voice needs. Not worth it if you're price-sensitive, mid-market, on Freshdesk/HubSpot/Gorgias, or you want to try before you sign.
Decagon is a strong product priced for a specific buyer, and its own case studies report real enterprise results (Chime at ~70% resolution and doubled NPS, Substack at ~90% resolution, Bilt at that $1.75M cost reduction, all Decagon-published). Decagon is good. The open question is whether you're the buyer it's built for, and whether you can commit to a bill you can't fully compute in advance.
Decagon is worth it if:
- You're at true enterprise scale, tens of thousands of conversations a month and up.
- You already run Zendesk or Salesforce and want an AI automation layer on top of it.
- You have a six-figure AI budget and a procurement team that prefers one negotiated contract.
- You need complex multi-channel or voice support on one platform.
- You have a dedicated owner to run a white-glove, forward-deployed implementation.
Decagon isn't worth it if:
- You're price-sensitive or mid-market, sitting below the floor where Decagon will quote.
- You're on Freshdesk, HubSpot or Gorgias (integration and Agent-Assist gaps bite here).
- You want to test on your real tickets before you commit budget, and there's no trial.
- You want a transparent price you can read without a sales call.
- You want quick time-to-value without a multi-week onboarding and a sales cycle.
For everyone Decagon prices out, and that's most teams reading this, a flat, usage-based model (the boring-but-effective option) means the bill stays flat as the AI improves, and you capture the upside of your own knowledge and API work while the vendor never charges you more for it. My AskAI is deliberately the option for that buyer: it runs inside the helpdesk you already have (Zendesk, Intercom, HubSpot, Freshdesk or Gorgias), it's SOC 2 Type II and GDPR compliant, and the price is on the page.
If you want the full feature picture on Decagon itself, the complete Decagon guide covers it. If you want to compare the field, the alternatives round-up is the place, or just read our pricing without booking anything. If you want to see the flat model hold up at real volume, our high-volume support case study walks through the numbers, or you can start a free trial on your own tickets today.
FAQs
How much does Decagon cost?
Decagon publishes no public price. It's enterprise custom, usage-based, quoted after a sales call.
The best independent anchor is Vendr's real-contract data, which shows a median annual contract around $432,750 and a range of roughly $105,000 to $923,183. Which end you land on depends on volume, model (per-conversation or per-resolution) and how hard you negotiate.
Is Decagon free? Does Decagon have a free trial?
No on both. There's no free plan, no free trial, and no self-serve signup. The only path is a demo, a discovery call, a custom quote and an annual contract.
A "Decagon University" sandbox exists, but it's an education resource for existing customers and gives you no way to trial the product. (Third-party claims of a Decagon free trial couldn't be corroborated against Decagon's own site.)
Does Decagon charge per conversation or per resolution?
Both models are on offer. Per-conversation is a fixed rate for every incoming conversation regardless of outcome.
Per-resolution is a higher rate charged only when the AI fully resolves a conversation, with escalations free. Decagon says the majority of its customers gravitate towards per-conversation pricing.
What counts as a "resolution" in Decagon?
It's set by Decagon's own algorithmic "AI resolved" tag. Decagon itself concedes in its glossary that defining a resolution is tricky and that "gray areas can lead to billing disagreements." Practically, that means the unit you're billed on under a per-resolution model is a definition Decagon controls and you can't independently audit. So if you go that route, get the definition written into the contract (I'd make that a redline).
Does Decagon have per-seat or per-agent fees?
No. Decagon is entirely usage-based, with no per-seat or per-agent licensing. That's a genuine contrast with the per-seat helpdesk AI products (Zendesk AI, Intercom Fin), where a seat license stacks on top of the AI charge, though Decagon's usage-based total is still an enterprise-sized number.
Is there a minimum contract or spend for Decagon?
Effectively yes. Vendr's lowest observed Decagon contract is around $105,000 a year, and Decagon positions itself for organizations doing tens of thousands of conversations, so there's a de facto enterprise floor (I've never seen a smaller deal cited). Contracts are annual by default, with quarterly payment reported as negotiable.
Why doesn't Decagon publish its pricing?
Because it sells enterprise-only, with every deal a custom, usage-based contract negotiated per customer under an NDA. There's no standard rate card to publish, and the price really does vary by volume, billing model and negotiation. It's the same reason there's almost no public billing chatter about Decagon: those invoices never become public artifacts.
What discounts can I negotiate with Decagon?
Discounts of up to roughly 30% off list have been reported for very large commitments (north of a million conversations), and multi-year or annual-prepay terms are the usual levers. With no published price, what you pay is largely a function of how hard you negotiate, which cuts both ways: the number is movable, but two similar companies can end up paying very different amounts.
How does Decagon pricing compare to Intercom Fin?
They work in different ways. Intercom Fin publishes a transparent $0.99 per resolution (plus Intercom seats), so you can model it at any volume.
Decagon is opaque, enterprise-custom, with a floor around $105K+ a year and no public rate. At a mid-market volume Fin is buyable and Decagon typically isn't.
At true enterprise scale Decagon competes on capability and negotiated rate. Both charge in a way where cost can rise as the AI resolves more, which is the argument for a flat per-ticket model instead.
Is Decagon worth the money for a mid-market team?
Generally no (it's the question I field most from mid-market teams). The reason has nothing to do with product quality: Decagon is priced and built for enterprise.
A mid-market team usually sits below the volume and budget where Decagon will engage, and there's no trial to prove value first. If you're mid-market and want AI support you can read the price of and test on your own tickets, a transparent per-ticket tool is the better fit.
How much is Decagon AI worth (the company)?
That's a different question from what the product costs. Decagon the company was valued at roughly $4.5B in its Series D (early 2026), which is the company's valuation and says nothing about the price of its product. If you searched "how much is Decagon worth" hoping for a price, the answer you actually want is the contract cost above: no public rate, ~$105K to $923K a year in real contracts, quoted after a demo.
