Twilio used its annual SIGNAL conference in San Francisco on May 6 to announce a new conversation layer that welds together AI agents, human agents, and every channel a business runs into a single persistent experience. Four capabilities launched into general availability at the same time: Conversation Memory, Conversation Orchestrator, Conversation Intelligence, and Agent Connect.

The announcement marks a significant pivot in how Twilio is positioning itself. Rather than being a communications API vendor, the company is now pitching itself as the infrastructure layer for the agentic era of customer engagement.

What Twilio Actually Shipped

Conversation Memory is a real-time context layer that maintains consistent memory across every interaction, whether handled by a human agent or an AI. It draws on customer profile data, conversation history, preferences, and behavioral signals, and makes all of it available to whichever agent picks up the conversation next. The idea is that a customer who called last week and then sent a message today gets continuity, not a cold start.

Conversation Orchestrator handles routing, escalation, and state management across multiple channels and multiple agents. When a conversation moves from an AI agent to a human, or from messaging to voice, the orchestrator keeps the context intact and ensures the handoff is seamless rather than disjointed.

Conversation Intelligence adds a generative AI layer that analyzes live conversations in real time, surfacing signals like sentiment shifts and escalation risk while the call or chat is still in progress. Rather than reviewing recordings after the fact, teams can see what is happening and trigger automated workflows mid-conversation.

Agent Connect lets businesses plug their own AI models and frameworks directly into Twilio’s voice and messaging channels without rebuilding their communications infrastructure. Critically, it is model-agnostic, so teams can swap or combine models from different providers without getting locked into a single vendor’s AI stack.

All four capabilities were available in private beta since January 2026. Beta customers including Car Finance 247, Centerfield, and Constellation Dealerships reported immediate operational benefits from the persistent memory and orchestration features before the public launch.

Why This Matters for Businesses Running AI Agents

The core problem Twilio is solving is one that every business deploying AI agents runs into: AI agents are effective in isolation but break down in handoffs. A customer who speaks to an AI on your website, then calls your support line, then texts your team, typically has to re-explain their situation every time. The conversation layer is designed to eliminate that reset.

CEO Khozema Shipchandler described the launch as a next-generation platform that will make customer conversations persistent, contextual, and continuous across every channel.

That framing matters. The companies winning with AI right now are not just deploying individual AI agents, they are building systems where AI and human agents hand work to each other fluidly. That requires shared memory, shared context, and a routing layer that keeps the conversation coherent regardless of who picks it up.

What This Means for Business

If your business runs a contact centre, uses AI for intake or customer support, or is evaluating voice AI for the first time, Twilio’s conversation layer changes the calculus in two ways.

First, it lowers the barrier to deploying AI agents alongside existing human teams. Agent Connect’s model-agnostic design means you are not forced to rebuild around a single provider’s AI. You can integrate what works for your use case without a wholesale infrastructure change.

Second, Conversation Memory solves a real problem that damages customer trust. Customers who have to repeat themselves lose confidence in the experience. Persistent memory across channels is not a nice-to-have feature, it is baseline customer service in 2026.

The companies that figure out how to blend AI agents and human agents into a single coherent conversation will outperform those still treating AI as a separate, siloed tool. That is the shift Twilio is betting on, and given their position as a core communications layer for thousands of enterprises, it is a bet worth watching.

---

Enterprise DNA put together a free field guide on exactly this: the full Claude ecosystem, Claude Code, and how to roll agents out without breaking things. Get the guide.