You’re reading an issue of “The AI Economy,” my newsletter exploring the forces shaping the AI era—tracking how AI is rewriting business, work, technology, and culture. Subscribe to get expert insights and curated updates delivered straight to your inbox.
How many times have you called customer service, explained your problem, got transferred, and had to start all over again? With AI agents in the mix, that experience doesn’t get better—it multiplies. Twilio calls it the “conversation gap”: the disconnect between fragmented systems that leaves AI agents without the context they need to pick up where the last interaction left off. To solve this, the communications infrastructure platform is introducing four capabilities built to eliminate “agent amnesia” and improve customer engagement.
This is a problem that Twilio argues most businesses aren’t even equipped to recognize today. “Most businesses are operating in the dark,” Omar Paul, Twilio’s vice president of product management, said during a press briefing last week. “They have siloed data—data sitting in different systems inside that enterprise. They struggle with fragmented channels, and the integration complexity that is required to link these things together is at a breaking point.”
He added that while AI was viewed as the so-called “easy button” for customer engagement, “this reality has not lived up to the hype.” Without a shared persistent memory of previous interactions, Paul said, agents—human or AI—are effectively starting blind every time. The fix, Twilio argues, requires infrastructure built from the ground up for the agentic era.
Meet the CX ‘Nervous System’
To close that gap, Twilio is launching four new platform capabilities designed to enable context-rich conversations between brands and their customers. Paul describes these products as the “nervous system for modern customer engagement.”
Twilio Conversation Memory
This gives AI agents persistent context about a customer across every interaction and channel — maintaining history, preferences, behavior, and conversation state in real time. The idea is straightforward: when a customer contacts a brand, the agent, human or AI, already knows who they are and where things left off. No cold starts, no repeated explanations.
Twilio Conversation Orchestrator
Where Conversation Memory handles context, Orchestrator handles continuity. It connects individual calls and messages across channels into a single, unbroken conversation—managing routing, escalation, state, and handoffs between humans and AI. The goal: a customer should never feel like they’re starting a new conversation just because they switched from chat to phone.
Twilio Conversation Intelligence
This turns live conversations into real-time signals that businesses can act on immediately. Think of it as a layer that watches interactions as they happen and triggers automated responses—escalating a call to a human agent, for instance, when a conversation with an AI is going sideways. “A relevant but surprisingly difficult thing to do,” Paul said, “is for a business today to detect that a call or a chat interaction is going south, and then to proactively bring a person into that interaction.”
Twilio Agent Connect
This self-hosted open-source framework lets companies plug their own AI agents and models directly into Twilio’s voice and messaging channels—without rewiring their existing stack. Businesses keep their preferred AI and models while Twilio handles the underlying complexity of real-time voice streaming, session management, and identity. “Think of it like an easy button for developers who build agents handling the physics of your communication,” Paul explained.
The Real CX Battle Lies in Infrastructure
Competition within the customer engagement space is fierce. Salesforce has moved aggressively with Agentforce, launching an AI-powered contact center that grounds agents in unified customer profiles. Amazon Web Services expanded its Connect offering last week into a broader agentic AI suite with its own persistent context capabilities. Zendesk is also making a case for what it calls “contextual intelligence” through its Resolution Platform, though its AI agents don’t yet maintain memory natively across interactions.
But Twilio’s bet is that the battle won’t be won at the feature level. Bells and whistles are table stakes. The real advantage, the company argues, lies deeper—at the infrastructure layer, where conversations are actually built, routed, and remembered.
“At the center of the CPaaS, CDP, and AI convergence, Twilio is redefining what a Customer Engagement Platform (CEP) looks like—one that remembers, adapts, and orchestrates across every touchpoint,” Mila D’Antonio, principal analyst for customer engagement at Omdia, writes in a statement. “That combination of vision, execution, and ecosystem leverage is what solidifies Twilio’s place at the top of the category.”
All four capabilities are generally available starting today.
Featured Image: Twilio signage recreated using Microsoft Copilot.
