The AI landscape is currently an alphabet soup of protocols like A2A, ACP, and MCP. We’ve spent the last year perfecting how agents talk to other bots and commerce systems, but the industry has largely ignored the most critical connection: The one with humans.
In today’s lead story, Frederic Lardinois reports that Twilio is bridging that gap with its new open-source Agent-2-Human (A2H) protocol, which streamlines the transition from autonomous task work to human intervention, ensuring agents can hand off to a human.
Our story has exclusive comment from Twilio’s Rikki Singh, the company’s VP of product and engineering for emerging technology, who says that while agents are becoming increasingly autonomous, you will always need a human in the loop: “Not because the AI is inefficient, but because there is an element of human judgment that informs a lot of decisions we make, and there is an element of trust that comes with that human judgment.”
Read more here: Twilio’s A2H is a new protocol that helps agents talk to humans
More stories of note today:
✧ Google’s Gemini 3.1 Pro is mostly great
✧ Prometheus and OpenTelemetry finally play nice
✧ Anthropic: You can still use your Claude accounts to run OpenClaw, NanoClaw and Co.
