$131 Million Says Decagon’s AI Customer Experience Is Here Now

Earlier this year, Decagon was said to be in talks to raise $100 million at a $1.5 billion valuation, with multiple sources confirming that Andreessen Horowitz and Accel were expected to co-lead the round. The company had signed over $10 million in annual contracts, and while the deal hadn’t yet closed at the time, it was already drawing strong investor interest. That interest proved prescient. Decagon has now closed a $131 million Series C round at the same $1.5 billion valuation, bringing its total funding to $231 million just one year after emerging from stealth. The round was oversubscribed by a factor of five, according to the company, underscoring the level of market confidence in its product and trajectory. The funding was led by Andreessen Horowitz and Accel, with participation from returning investors Bain Capital Ventures, A*, and BOND, and new backers Avra, Forerunner, and Ribbit Capital. Founded in 2023 by Jesse Zhang and Ashwin Sreenivas , Decagon started with a deliberate search. The co-founders spent months interviewing dozens of companies to find a problem that was both critical and uniquely solvable with the latest advances in AI. They landed on customer support—an area that continues to drain resources across industries, with teams often overwhelmed by tickets, long queues, and repetitive queries. “Ashwin and I had seen firsthand how limited, rigid, and inefficient legacy CX systems had become,” said Zhang. “We knew the latest generation of LLMs were uniquely suited to change that.” Rather than follow the typical playbook of professional services and custom integrations, Decagon built a software-first model that lets companies create and update customer-facing AI agents using a system it calls Agent Operating Procedures, or AOPs. These are written in natural language by customer experience teams and automatically compile into code, allowing the agents to respond to a range of complex requests. Businesses don’t need to rebuild workflows from scratch for every scenario. AOPs have become central to Decagon’s product appeal. Customer experience operators can define how agents should behave in specific situations using plain language. That logic is then converted into machine-readable instructions that adapt to real-world conversations. Technical teams can still access the underlying code to manage integrations, enforce compliance, and refine the system as needed. The result is a system that companies can deploy within weeks, not months, and continue to update without external help. Agent performance is visible in real-time, across channels, giving teams the tools to troubleshoot, experiment, and improve outcomes quickly. “Decagon’s relentless focus on customer outcomes and differentiated approach to human-agent collaboration make them a breakout leader in AI agents for customer experience,” said Ivan Zhou, Partner at Accel. The product has attracted a wide mix of clients, including Hertz, Duolingo, Oura, Bilt, Eventbrite, Notion, ClassPass, Substack, and Rippling. While many companies in this space tout efficiency gains, Decagon’s reported numbers offer more concrete validation. Duolingo has achieved deflection rates above 80 percent. Oura has seen customer satisfaction scores triple. ClassPass has processed over 2.5 million conversations with Decagon’s agents and cut its cost per reservation by 95 percent. Bilt, a credit card and loyalty platform, reduced its customer support team from hundreds to 65. At Hertz, Decagon’s product has changed how the company manages its support workflow. “Our partnership with Decagon has taken us from idea to impact without sacrificing control, visibility, or enterprise standards,” said Vikram Rajagopalan, Vice President of Customer Experience. “Its AI agents are flexible, reliable, and built for scale, which is enabling more personalized interactions and helping transform how we serve our customers.” The company’s infrastructure spans chat, voice, email, and SMS. Once support logic and brand guidance are written, they apply across all channels. That consistency allows Decagon to handle tasks such as identity verification, refunds, and ticket escalations without context loss. A centralized system governs knowledge and execution across the platform, eliminating the need to build separate bots for each medium. Technically, Decagon relies on a mix of foundation models from OpenAI, Anthropic, and Cohere, and trains its agents using internal documentation, help center content, and historical support transcripts. Human reviewers are part of the process, rating and refining agent responses to improve quality over time. In February, the company announced a partnership with ElevenLabs to develop voice agents capable of more natural interactions. The company appeared on the Forbes AI 50 list in April, just months after it raised a $65 million Series B round led by Bain Capital Ventures at a $650 million valuation. Since then, it has more than quadrupled its customer base

$131 Million Says Decagon’s AI Customer Experience Is Here Now

Keep reading. It's free.

Related stories