Anthropic Tests Conway An AI Agent That Never Stops Working

Anthropic is testing a new AI system codenamed Conway. Unlike today's chatbots that wait to be asked, Conway is designed to work continuously in the background and take initiative on its own.

Conway represents a fundamental shift in how we think about AI assistants. Instead of responding when you ask, Conway functions as a digital twin residing on your computer, automatically activated by events. It can navigate the web, gather information, process data, and execute complex workflows from start to finish without requiring you to supervise.

Anthropic describes Conway as an always-on agent. It supports webhook integrations, allowing external services to trigger tasks. Developers will also be able to build custom tools and UI elements on top of the system.

This is not just a technical update. It is a strategic move by Anthropic to compete in the agent market and transform Claude from a passive conversational tool into an operational partner that actually does things, not just talks about them.

Conway is still experimental, but the signal is clear: the future of AI does not wait. It acts.