Decagon raises Series C at $1.5B to deploy AI customer service agents for large consumer businesses
Jun 23, 2025 with Jesse Zhang
Key Points
- Decagon raises Series C at $1.5B valuation from Andreessen Horowitz and Accel to deploy full-action AI agents that handle transactional customer service flows like refunds and rebooking within live interactions.
- LLM-powered agents close exploitable gaps in legacy decision-tree chatbots by adapting to context and user history, shifting customer service from a cost center into a potential revenue channel through proactive upselling.
- Sharp drop in BPO stocks after large language models emerged signals that customer service is AI's most obvious displacement target, validating Decagon's timing to capture contact-center volume before incumbents close the capability gap.
Summary
Decagon, an AI customer service agent platform, raised a Series C led by Andreessen Horowitz and Accel at a $1.5 billion valuation. Co-founder Jesse Zhang describes the product as a full-action agent rather than a chatbot. It can book hotel rooms, process refunds, reorder credit cards, manage loyalty points, and handle cancellation flows within live voice or chat interactions.
Decagon's advantage over legacy chatbots rests on nuance at scale. Old decision-tree systems were trivially gameable: someone would post a Reddit workaround and the shortcut would go viral, hitting retention metrics directly. LLMs let Decagon build flows that are context-sensitive and user-specific, closing that loophole while covering far more ground than any scripted tree.
From cost center to revenue channel
Zhang argues the more interesting medium-term opportunity is flipping customer service from a cost center into a revenue-generating surface. Most deployments start with low-hanging fruit like transaction processing and FAQ lookups. A genuinely good AI experience increases engagement rather than just deflecting tickets. That higher engagement opens the door to proactive outreach, upselling, and retention saves that previously required skilled human agents.
Zhang describes the endpoint as a product concierge: an interface so capable that customers use it instead of navigating an app. Whether that framing converts into measurable revenue uplift for clients is not quantified, but it is clearly the sales narrative Decagon is leading with.
BPO market impact
Call center BPO stocks dropped sharply when large language models first emerged. Zhang treats that as validation that customer service is one of AI's most obvious displacement targets. Better-run BPOs are pursuing AI-native strategies, either by partnering with vendors like Decagon or building internally. Partnership should dominate, since enterprise buyers typically view BPOs as service providers rather than software builders.
Decagon competes against Intercom and Sierra, among others. The round signals that investors see room for a well-funded specialist to capture significant contact-center volume before larger incumbents close the capability gap.