What is Decagon AI? If you lead a service team in Germany or the EU, you have probably read the headline: a US unicorn that raised 250 million USD in January 2026 at a 4.5 billion USD valuation to build AI agents for customer support, with Deutsche Telekom, Chime, Duolingo and Rippling on its logo wall14. This article explains what Decagon actually does, who built it, how its platform is put together, where it hits limits for a European team, and how Superkind’s ready-made AI team for customer service gives you a different answer: one AI employee in Outlook, Teams and Slack, with approval wherever you want it, hosted in Europe.
Key takeaways
What it is: Decagon AI is a US customer-service agent platform. Autonomous agents on chat, email, voice and SMS, aimed at large English-first enterprises on Zendesk, Intercom, Salesforce or Kustomer12.
What it costs: Decagon publishes no prices. Third-party sources cite 40 to 80 USD per seat per month at the low end and six-figure annual contracts (200,000 to 500,000 USD and more) for enterprise deployments, through sales only8.
Get a ready-made AI team: Superkind comes with a customer-service team (Service Helper, Return Pro, Knowledge Keeper, Complaint Router) that runs automatically in Outlook, Teams and Slack. Start for free with unlimited users, or build your own.
Keep control, keep data in Europe: hosted in Europe, GDPR compliant, EU AI Act aligned, no model training on your data, and you decide what needs your approval.
Your AI team for customer service
What is Decagon AI?
Decagon AI is a US customer-service agent platform that runs autonomous AI agents across chat, email, voice and SMS, aimed at large enterprises in financial services, travel, retail, healthcare, telecommunications and technology1. The company calls this an “AI concierge”: a layer that reads a ticket or picks up a call, pulls what it needs from the helpdesk and the CRM, and resolves the case without a human for the ones that fit a defined workflow15. One example of a different kind of answer is Superkind: a ready-made AI team that handles service requests across Outlook, Microsoft Teams and Slack, hosted in Europe, and goes beyond service into accounting, sales and operations.
Decagon was founded in August 2023 in San Francisco by Jesse Zhang (CEO, previously founder of the gaming clip app Lowkey, which was acquired by Niantic) and Ashwin Sreenivas (previously founder of the AI startup Helia, acquired by Scale AI)67. Backers include Andreessen Horowitz, Accel, Bain Capital Ventures, Coatue and Index Ventures4. In January 2026 Decagon raised a 250 million USD Series D at a 4.5 billion USD valuation, bringing total funding reportedly to around 481 million USD across five rounds45. Named customers on the public page include Deutsche Telekom, American Airlines, Delta Airlines, Square, Chime, Hertz, Ticketmaster, Cash App, Duolingo, 1-800-Flowers, Hunter Douglas, Riot Games, Rituals, Soho House, Notion, Snapchat, Perplexity, Fanatics, ClassPass, Oura, Curology, Valon and Rippling1.
How Decagon works
Decagon’s public product page groups the platform into four parts1. In plain terms:
- Channels - the agents sit on chat, voice, email and SMS. There is also a separate channel called “personal agents”, a route for AI agents acting on behalf of a customer, with permissions the brand controls1.
- Build - AOPs and Integrations - AOPs (Agent Operating Procedures) are natural-language workflows that define what an agent should do, step by step, for a given case. Pre-built connectors plug into the helpdesk, the CRM and the knowledge base1.
- Optimize - Experiments and Testing & QA - live A/B testing between agent versions, plus simulation at scale so you can rehearse changes before they go live1.
- Scale - Insights, Watchtower and Suggestions - reporting, always-on QA on conversations (Watchtower) and AI-powered knowledge suggestions pulled from recent tickets1. On top of this sits Duet, which Decagon positions as an “AI partner” for building and tuning the agents themselves1.
The integration list is a service-stack one: Zendesk, Intercom and Salesforce for ticketing and CRM; Confluence, Contentful and Kustomer for knowledge; Amazon Connect and RingCentral for voice (plus SIP trunking); Zendesk Sunshine and Salesforce for live-chat escalation; Zendesk and Intercom again for automated email routing; and MCP open connectivity plus APIs for extensibility2. This is a deep plug into the English-speaking helpdesk world. It is not a general enterprise tool kit.
Who is Decagon for?
Decagon is pointed squarely at large customer-service operations in English-first markets. The customer wall and the independent reviews point to the same profile19.
- Enterprise service teams on Zendesk, Intercom, Salesforce or Kustomer - the integration depth is the main technical pull.
- High-volume English-language support - fintech (Chime, Square, Cash App), travel (American, Delta, Hertz), consumer (Oura, Duolingo, Rituals), SaaS (Notion, Rippling, Perplexity)1.
- Teams that want chat, email and voice from one vendor - the voice channel with Amazon Connect and SIP trunking is a differentiator against pure-chat tools2.
- Teams happy to buy through annual contracts and sales - there is no self-serve, no public price and no monthly plan8.
It is less obviously built for mid-sized service teams, non-English markets or teams that live in Outlook, Teams or Slack rather than in a dedicated helpdesk. That is where the European alternative question comes in.
Where Decagon hits limits in Europe
Decagon is a strong US enterprise tool and does the service-only job well where it fits9. For a service team in Germany or the EU, five limits come up again and again. None of them are a disqualification on their own; together they are the reason to look at a European alternative first.
- No stated EU hosting region - Decagon’s security page lists GDPR and SOC 2 Type II and talks about “multi-region infrastructure”, but it does not publish a dedicated EU region or EU data residency option3. For a DSGVO review, your DPO still has to request the DPA, run a transfer impact assessment and ask Decagon in writing where EU data is processed and stored.
- English-first by design - the product, the AOP examples and the voice agent are tuned for English-speaking markets. German, French and Dutch service work on paper, but your service manager will not find a native setup aimed at the DACH context.
- Lives in the helpdesk, not in Outlook or Teams - Decagon connects to Zendesk, Intercom, Salesforce and Kustomer2. If your service happens in a shared Outlook mailbox, a Microsoft Teams channel or a Slack workspace, you are either migrating onto a US helpdesk first or running two tools in parallel.
- Service-only - the agents resolve service tickets and calls. Accounting, sales, HR and operations work needs a separate tool, which means a separate contract, a separate price and a separate integration project.
- Enterprise-only pricing and sales motion - annual contracts in the hundreds of thousands of dollars, no public price, no monthly cancellation8. For a Mittelstand service team of five to twenty people that is the wrong budget shape.
Zero-data retention with OpenAI and Anthropic is a real plus3. But data retention at the AI provider is not the same as EU data residency at the platform, and that distinction matters when a European works council asks the question.
How it works with Superkind: your AI team for customer service
Superkind is the AI employee for your service team: you write to it in Outlook or any email, in Microsoft Teams, in Slack or in the Superkind web app, and it does the task across your helpdesk and your other systems. Hosted in Europe, GDPR compliant, with 1,500+ integrations into Zendesk, Freshdesk, HubSpot, Salesforce, lexoffice, DATEV, SharePoint and more. No new tool to learn. No migration onto a US helpdesk. The ready-made customer-service team runs automatically on triggers (new ticket, new email, new return request, every Monday morning) and asks you only where you asked it to.
You do not start from scratch: Superkind comes with ready-made AI employees for customer service. They run automatically in the background, triggered by new tickets, calls or emails, and you decide where they ask you first. Use them right away, adapt them, or build your own.
Service Helper
AI employee for first-line customer messages
Service questions arrive everywhere: Outlook, the shared mailbox, the contact form, the chat. The Service Helper picks up every new message on its own, pulls the order from your shop and the invoice from DATEV or lexoffice, writes the reply in your tone and sends it. Only the one case where it is not sure, or a complaint, lands with you.
+ 1,700 moreReturn Pro
AI employee for returns and refunds
Refunds eat up more time than any other service task because the data sits in three systems: the ticket, the shop and the bookkeeping. The Return Pro reads every incoming return request, checks the policy and the window, creates the DHL label and the credit note in DATEV. Routine refunds run on their own. Refunds above your limit wait for your OK with one click.
+ 1,700 more
Knowledge Keeper
AI employee for your help centre and standard replies
Every solved ticket is an answer waiting to be reused, and nobody has time to write it up. The Knowledge Keeper reads every closed ticket, groups questions that come back and updates your help centre and standard replies in SharePoint, so the next agent and the Service Helper always answer the same way.
+ 1,700 moreEvery Monday: look at last week’s closed tickets, group the questions that came back more than three times and draft help centre entries for them.
Reasoning ⌄Three new help centre entries drafted from 11 recurring questions last week: “Change delivery address after order”, “Return a gift”, “Combine two orders”. Published in SharePoint. Service Helper now uses them in replies.
Runs automatically every Monday
Complaint Router
AI employee for complaint triage
A complaint that lands in the wrong inbox costs you a customer. The Complaint Router reads every incoming message, spots the complaints, understands whether it is about a product, a delivery or an invoice, and routes it to the right person in Slack with the order, the invoice and a short summary. Nothing is sent on your behalf, you only decide.
+ 1,700 more
Every employee is ready to use right away. Or build your own on the platform.
Safe by default
Superkind is hosted and processed in Europe, GDPR compliant, aligned with the EU AI Act and does not train models on your customer data. The AI models come from OpenAI, Anthropic, Mistral and Google, all processed in Europe. A data processing agreement is available through the Trust Center. See the Data protection block below for how this plays out for a service team handling customer data every day.
- Hosted and processed in Europe
- GDPR compliant
- Aligned with the EU AI Act
- No training on your data
- DPA through the Trust Center
Send Superkind your first service email
Start for free in Outlook, Teams or Slack. Approval wherever you want it.
From the first task to your own Super App
- Start - sign up for free at app.superkind.ai with up to EUR 150 in credits, forward the first customer email to Superkind in Outlook, Teams or Slack and watch it get answered.
- Relief - what you explain once keeps running as a routine: every morning the overnight tickets are triaged, every return request is pre-processed with a draft refund, late DHL parcels get a reply, with your approval on the steps you marked as sensitive.
- Knowledge - every closed ticket makes Superkind smarter. Your tone, your return policy, your shipping windows and the way you handle complaints become your Company Brain, and it stays in the business even when a service agent moves on.
- Growth - when it gets more complex, you don’t start over. Build a custom returns workflow or a VIP-customer app on the Superkind platform, on top of what Superkind already knows about your service. Your own Super App, built with the Superkind team or by yourself, and it all belongs to you.
What does Decagon cost?
Decagon does not publish a pricing page. Every number below comes from third-party reviewers or analysts, which disagree with each other. For the authoritative figure for your volume, ask Decagon directly.
As of November 2026, published third-party ranges:
| Source | Reported figure | Scope |
|---|---|---|
| Enterprise seat pricing (third-party) | 40-80 USD per seat per month for typical enterprise deployments8 | Lower end, estimate from reviewer |
| Six-figure annual contracts (third-party) | 200,000 to 500,000+ USD per year depending on volume8 | Enterprise deployments, signed-contract data |
| Historical pricing model | Per-conversation and per-resolution billing (noted in older write-ups) | Unconfirmed for 2026, verify with sales |
| Sales motion | Annual contract, mandatory sales call, no public plan, no self-serve1 | Vendor’s own site |
| Superkind (for comparison) | Free: EUR 0 with up to EUR 150 in credits. Business: from EUR 50 / month for the whole company, unlimited users. Enterprise: custom | Vendor’s own public pricing (as of October 2026) |
Two things are stable across the sources. First, Decagon is sold through annual contracts and a sales conversation, not a signup page. Second, the entry point for an enterprise deployment is in the tens of thousands of dollars per year at the low end, rising into six figures at scale8. For a European service team evaluating Decagon against a general AI employee, Superkind’s EUR 50 per month for the whole company with unlimited users is a different budget conversation.
Decagon vs. Superkind
Two different products for two different jobs. The table shows where they overlap and where they do not.
| Criterion | Superkind | Decagon AI |
|---|---|---|
| What it is | One AI employee for the whole team, in Outlook, Teams, Slack and the web app | Customer-service agent platform (chat, voice, email, SMS) for US enterprises1 |
| Where you work with it | Outlook, Gmail, Microsoft Teams, Slack, web app | Zendesk, Intercom, Salesforce, Kustomer, Amazon Connect, RingCentral2 |
| Approval before send | You decide what needs approval; sensitive steps wait for your OK, routine runs on its own | Agents designed to resolve autonomously; approval not the default pattern |
| Beyond customer service | Service, accounting, sales, HR and operations in one tool; 1,500+ integrations | Customer service only1 |
| Hosting / Data protection | Hosted in Europe, GDPR, EU AI Act, no training, DPA via Trust Center | GDPR + SOC 2 Type II; zero-data retention with OpenAI and Anthropic; no stated EU hosting region3 |
| Pricing | Free start with up to EUR 150 in credits; Business from EUR 50 / month for the whole company, unlimited users (as of October 2026) | Annual contract through sales; 40-80 USD per seat / month at low end, up to 200,000-500,000+ USD / year at scale (third-party)8 |
| Language focus | German and English native, more languages on request | English-first |
| Who it fits | Service teams in Germany and the EU that want approval and want their data in Europe | Large US enterprises on Zendesk, Intercom, Salesforce or Kustomer with high English-language ticket volume |
Data protection and EU hosting
A customer-service agent reads names, order numbers, invoices and sometimes health or financial data. Where that data is processed matters as much as the deflection rate. Decagon is a US vendor, lists GDPR and SOC 2 Type II credentials and enforces zero-data retention at the AI providers, but does not publicly name an EU hosting region on its security page3. Superkind is the opposite: a European vendor built for European teams.
What that means for a service team: customer names, order histories, return reasons, invoices and call transcripts stay inside European processing, moved between Zendesk or Freshdesk, HubSpot, DATEV, lexoffice and SharePoint by Superkind’s 1,500+ integrations. If Decagon is on the shortlist, make the DPA, the EU hosting location and the transfer impact assessment part of the first demo, not an afterthought.
Further resources
- Decagon AI alternatives for European customer-service teams - the full side-by-side comparison.
- AI for customer service - the broader guide.
- What is 11x.ai? - the sales equivalent, same question for an outbound tool.
- What is an AI employee? - the category Superkind sits in.
- AI voice agents for inbound calls - where voice fits in for a service team.
Conclusion
Decagon AI is a well-funded US bet on autonomous customer-service agents: an AI concierge on chat, voice, email and SMS, built on AOPs, Watchtower and a deep plug into Zendesk, Intercom, Salesforce and Kustomer, aimed at large English-first enterprises on annual contracts124. For a service team in Germany or the EU, the interesting questions are not whether it works at scale - it clearly does, at Deutsche Telekom and American Airlines scale - but whether an English-first US helpdesk tool with no stated EU hosting region is the right shape for a Mittelstand service operation that runs in Outlook, Teams or Slack3. If the answer is “probably not”, start Superkind for free: one AI colleague across service, accounting, sales and operations, with 1,500+ integrations, approval wherever you want it, hosted in Europe. See the Decagon alternatives article for a side-by-side on the tools in this space.
Frequently Asked Questions
Decagon is a US customer-service AI platform that runs autonomous agents across chat, email, voice and SMS for large enterprises. The company calls this an “AI concierge”. Its platform includes natural-language workflows (AOPs), pre-built integrations into Zendesk, Intercom, Salesforce, Kustomer and others, live testing, QA (Watchtower) and an analytics layer. It is pointed at one job: deflecting and resolving customer tickets at volume in English-first markets.
Decagon lists GDPR and SOC 2 Type II as compliance credentials on its security page and says it enforces zero-data retention with the AI providers it uses (such as OpenAI and Anthropic). The page describes AES-256 for storage, TLS 1.2+ in transit and “multi-region infrastructure”, but it does not name an EU hosting region or dedicated EU data residency. For a European service team processing customer data, that means you still need to request the DPA, run a transfer impact assessment and confirm where EU data is stored before you sign. Superkind takes the opposite position: hosted and processed in Europe, GDPR compliant, aligned with the EU AI Act, no model training on your data, DPA through the Trust Center.
Decagon does not publish a price list. Third-party sources name very different figures: typical enterprise deployments at 40 to 80 USD per seat per month, or six-figure annual contracts of 200,000 to 500,000 USD and more depending on volume, with a per-conversation or per-resolution meter in older write-ups. All deals go through sales, usually on an annual contract. For the current number, ask Decagon directly. Superkind, by comparison, starts at EUR 0 with up to EUR 150 in credits, and Business plans start from EUR 50 per month for the whole company with unlimited users (as of October 2026).
Decagon was founded in August 2023 in San Francisco by Jesse Zhang (CEO, previously founder of the gaming app Lowkey, which was acquired by Niantic) and Ashwin Sreenivas (previously founder of the AI startup Helia, acquired by Scale AI). The company is backed by Andreessen Horowitz, Accel, Bain Capital Ventures, Coatue and Index Ventures. In January 2026 Decagon raised a 250 million USD Series D at a 4.5 billion USD valuation.
Superkind is the European alternative. Where Decagon is a specialist customer-service platform for US enterprises, Superkind is an AI employee for your whole team: you write to it in Outlook, Microsoft Teams, Slack or the web app, and it does the task in your programs through 1,500+ integrations such as Zendesk, Freshdesk, HubSpot, Salesforce, DATEV and SharePoint. The same AI employee that answers service tickets can also book invoices, write quotes and chase late payments, so you do not need a separate tool for every department. Hosted in Europe, GDPR compliant, from EUR 50 per month for the whole company with unlimited users (as of October 2026).
Decagon is a customer-service-only agent platform aimed at large US enterprises on Zendesk, Intercom, Salesforce or Kustomer, with annual contracts and sales-led pricing. Superkind is one AI employee for the whole team, in Outlook, Teams, Slack and the web app, with 1,500+ integrations and ready-made AI employees for service, accounting, sales, HR and operations. You decide what needs approval, routine work runs on its own, and your Company Brain keeps the knowledge when people leave. Hosted in Europe, from EUR 50 per month for unlimited users.
Decagon is English-first and sits in the service helpdesk, not in Microsoft Teams or Outlook. German-language service is possible on request, but the product is tuned for English-speaking markets. If your team works in Outlook, Teams or Slack and needs German-language replies by default, Superkind lives inside those tools and speaks the language your team uses.
Sources
- Decagon - Product page (voice, chat, email, Duet, AOPs, integrations)
- Decagon - Integrations (Zendesk, Intercom, Salesforce, Kustomer, Confluence, Amazon Connect)
- Decagon - Security (SOC 2 Type II, GDPR, AES-256, TLS 1.2+, zero-data retention)
- Built In - Customer Experience Platform Decagon Raises $250M at $4.5B Valuation (January 2026)
- SiliconANGLE - Decagon AI raises $250M at $4.5B valuation to scale AI concierge (28 January 2026)
- Wikipedia - Decagon (company)
- VentureBeat - Decagon emerges from stealth to provide human-like AI agents for enterprise customer support (June 2024)
- Sacra - Decagon profile and revenue analysis
- Helpshift - Decagon vs Zendesk AI: A Detailed 2026 Comparison
- EU AI Act - Regulation (EU) 2024/1689, Article 4 AI literacy and implementation timeline
Send Superkind your first service task.
Start for free in Outlook, Teams or Slack, with approval wherever you want it. Or see in 30 minutes how it runs in your systems.