How does Decagon handle voice, chat, and email in a unified way?

Decagon unifies all three channels within a single intelligence layer. This means your AI agent's knowledge, memory, and workflow logic are consistent whether a customer interacts via phone, web chat, or email, providing a seamless cross-channel experience.

Can I integrate Decagon with my existing CRM or support tools?

Yes, Decagon provides an 'Integrations' feature with support for tool connectors. While specific tools aren't listed on the homepage, the platform is built to connect with external systems to execute workflows.

What are Agent Operating Procedures (AOPs) and how do they work?

AOPs are a proprietary framework for defining AI agent logic. Instead of writing complex rules or scripts, you describe workflows and decision trees in natural language. The Decagon engine interprets these AOPs to guide the AI's behavior, making it easier for non-technical teams to build and modify agent flows.

What kind of testing and optimization tools are included?

Decagon includes a comprehensive optimization suite: 'Experiments' for live A/B testing of agent behaviors, 'Testing & QA' for running simulations at scale, and 'Watchtower' for always-on automated quality assurance monitoring.

How does the 'Duet' AI partner feature work?

Duet is described as a 'self-improving agent' for conversational AI. Based on the context, it likely acts as an AI copilot that analyzes performance data, identifies optimization opportunities, and suggests or implements improvements to your agent's workflows and knowledge, creating a continuous improvement loop.

Is Decagon suitable for small businesses or startups?

The platform is heavily marketed towards 'world's leading enterprises' with features for scaling and deep optimization. Its focus on enterprise-grade ROI metrics (like 95% cost reduction) and a likely complex setup suggests it may be overkill and cost-prohibitive for small businesses with simpler needs.

Directory/Chatbots/AI Assistants/Decagon

Decagon

Decagon is an enterprise conversational AI platform that builds, deploys, and manages AI concierge agents across chat, voice, and email channels. It uses natural language 'Agent Operating Procedures' (AOPs) to define agent workflows, enabling non-engineers to build and iterate quickly. The platform includes omnichannel deployment, robust testing & QA, A/B experimentation, and a unified analytics suite for measuring customer experience and operational metrics like deflection rate and CSAT. Targeted at enterprises in industries like financial services, travel, retail, and health & wellness, Decagon is designed for customer experience, operations, and support teams seeking to automate complex customer interactions. It emphasizes brand customization, cross-channel memory, and self-improving agents via features like 'Duet Autopilot'.

Chatbots

Visit website

Added on

Sep 22, 2026

Monthly Visits

278.3K

Voice AIChat AIEmail AINatural Language AOPsA/B TestingCustomer InsightsEnterprise Platform

Product information

Everything worth knowing about Decagon

The complete picture — from everyday features to the technical detail builders and IT folks dig for.

Core features

What it actually does

Omnichannel Support

Unified platform for building and deploying AI agents across voice, chat, and email channels.

Agent Operating Procedures (AOPs)

Define agent workflows using natural language for faster iteration and transparency.

Live A/B Testing

Validate and iterate on AI logic with robust experimentation tools.

AI-Powered Insights & Reporting

Turn every conversation into actionable customer insights.

Conversational AI Concierge

Build agents that deliver personalized, trusted experiences across the customer lifecycle.

Technical capabilities

Unified Intelligence Layer

Ensures consistent customer experiences across all deployed channels from a single build.

Self-Improving Agent (Duet Autopilot)

AI partner that continuously optimizes agent performance.

Must watch videos

The New AI-Native Stack: Founders on Building Agents That Work at Enterprise Scale

31:17

Tutorial31:17

The New AI-Native Stack: Founders on Building Agents That Work at Enterprise Scale

AI agents are application-layer products, not base-model outputs—and the value being captured in the AI era comes from companies that solve enterprise governance, latency, cost, and multi-model orchestration on top of foundation models. This panel brings together founders of leading AI-native companies to explain why a base model alone does not solve enterprise problems, and what it actually takes to ship agents into production. Patrick Wendell (Databricks co-founder) moderates a conversation with Ashwin Sreenivas (Decagon, AI agents for enterprise customer support), TR Vishwanath (Glean, enterprise AI assistant), and Jeff Wang (Cognition, Devon software engineering agent). The founders discuss why model routing across a dozen providers—including open-source—delivers better ROI than committing to one lab, how governance and permission layers prevent AI from leaking sensitive documents, why sycophancy and hallucination make current AI unsuitable for high-stakes automated decisions, and how software engineering employment has actually increased as coding agents have gotten better. The panel covers real augmentation vs substitution data and concludes that unlimited ambition, not automation anxiety, is the defining characteristic of teams winning with AI today. Talk By: Patrick Wendell, Co-Founder, Databricks; Ashwin Sreenivas, Co-Founder, Decagon AI; Jeff Wang, President New Enterprise, Cognition; TR Vishwanath, Co-Founder & CTO, Glean 00:00 - Introduction and panel overview 00:59 - Decagon: AI agents for enterprise customer support 01:48 - Glean: governed enterprise AI assistant 02:51 - Cognition and Devon: agentic software engineering 04:30 - Why base models are not enough for enterprise AI applications 13:57 - Multi-model strategy: routing across a dozen providers for cost and performance 21:53 - AI augmentation vs substitution: where are we today? 26:17 - Software engineering employment empirics in the agent era 28:47 - Impact on everyday knowledge workers and organizational productivity 🤝 Stay connected: Website: https://databricks.com Twitter: https://x.com/databricks/ LinkedIn: https://www.linkedin.com/company/databricks/ Instagram: https://www.instagram.com/databricksinc/ Facebook: https://www.facebook.com/databricksinc/

0:29

Tutorial0:29

Introducing Decagon Assist

E49: Where AI Meets Customer Experience with Bihan Jiang (Decagon)

34:47

Tutorial34:47

E49: Where AI Meets Customer Experience with Bihan Jiang (Decagon)

This week, we are chatting with Bihan Jiang, a product lead at Decagon. Decagon is reimagining customer service with AI agents. The company recently raised a $65M Series B and is trusted by companies such as Eventbrite, Substack, ClassPass, Rippling, Notion, and more. We dive into her journey starting at Scale AI after Stanford, moving to the application side of AI, LLM capabilities powering Decagon, and how customers are using Decagon to elevate their customer experience. We also dive into how Decagon is powering email, SMS, and voice and how they adhere to customers' requests for accuracy and reliability. Episode Chapters: 1. Growing up in Texas - 1:53 2. The AI powered PM - 3:22 3. Career progression - 6:12 4. Customer support market - 8:58 5. Users of Decagon - 13:13 6. Can we automate 100%? - 16:28 7. Moving into voice AI - 17:25 8. Customer research - 22:22 9. Product roadmap decisions - 26:44 10. Joining an AI startup - 29:15 11. Ending questions - 32:37 As always, feel free to contact us at partnerpathpodcast@gmail.com. We would love to hear ideas for content, guests, and overall feedback. This episode is brought to you by Grata, the world's leading deal sourcing platform. Our AI-powered search, investment-grade data, and intuitive workflows give you the edge needed to find and win deals in your industry. Visit grata.com to schedule a demo today. Fresh out of Y Combinator’s Summer batch, Overlap is an AI-driven app that uses LLMs to curate the best moments from podcast episodes. Imagine having a smart assistant who reads through every podcast transcript, finds the best parts or parts most relevant to your search, and strings them together to form a new curated stream of content - that is what Overlap does. Podcasts are an exponentially growing source of unique information. Make use of it! Check out Overlap 2.0 on the App Store today.

Who uses it

Real use cases, no hype

Customer Support Automation

Deploy AI agents across chat, voice, and email to handle common queries, deflect tickets, and provide 24/7 support.

Omnichannel Experience Unification

Ensure consistent brand messaging and context as customers move between phone, chat, and email channels.

High-Volume Contact Center Optimization

Achieve massive cost reductions and deflection rates (e.g., 95% cost reduction, 80% deflection) for large enterprises.

Rapid AI Workflow Development

Use natural-language AOPs to build and refine conversational flows quickly without extensive coding.

Customer Insights & Voice of Customer

Leverage conversation data and analytics to understand customer needs and improve products/services.

Brand-Safe AI Interactions

Deploy fully customizable voice and chat agents that adhere strictly to company guidelines and values.

Scalable QA and Performance Monitoring

Use automated testing and Watchtower to continuously monitor and ensure AI agent reliability at scale.

What's great

Omnichannel platform (chat, voice, email) in one system
Natural-language Agent Operating Procedures (AOPs) for easy workflow definition
Rapid iteration without requiring engineering sprints
Proven high deflection rates and cost reductions

Technical strengths

Self-improving Duet AI partner for continuous optimization
Built-in A/B testing (Experiments) and simulation-based QA at scale
Always-on Watchtower for automated quality assurance

Where it falls short

Enterprise-focused, likely complex and expensive for small teams
Steep learning curve for full platform mastery
Limited transparency on specific AI model architectures used

Technical limitations

Proprietary AOPs syntax adds another layer to learn
Integration depth with specific third-party tools not fully detailed

For technical folks

The deep-dive specs

FAQ

Includes technical Q&A

Decagon focuses on building an 'AI concierge' for the entire customer lifecycle, not just support deflection. Its key differentiator is Agent Operating Procedures (AOPs), which allow you to define complex agent workflows in natural language for faster, more transparent iteration compared to traditional configuration languages.