Poke makes using AI agents as easy as sending a text

A new startup called Poke is offering a personal AI agent accessible through common messaging apps like iMessage, SMS, and Telegram, aiming to bring complex automation to a non-technical audience. The agent, which launched publicly in March, allows users to perform actions like managing calendars, tracking health data, and controlling smart home devices directly via text. The service arrives amidst growing industry interest in agentic AI systems, such as OpenClaw, but seeks to bypass the technical barriers and security concerns associated with those more powerful, developer-focused tools.

The Interaction Company of California, the Palo Alto-based startup behind Poke, recently added $10 million in funding on top of a $15 million seed round, bringing its post-money valuation to $300 million. Backed by Spark Capital and General Catalyst, the company built Poke to be model-agnostic, using the best AI for a given task from various providers or open-source projects. Users interact with the agent through pre-made automations called "recipes" that integrate with services like Gmail, Strava, and Notion. For more demanding real-time tasks, the company has a flexible pricing model where the agent itself helps determine a personalized cost based on usage, though the basic service is free.

By embedding its agent within existing messaging platforms, Poke is making a strategic play to lower the barrier to entry for AI-powered automation. Instead of requiring users to download a new app or learn a complex interface, it leverages the familiarity of texting to drive adoption. The company is also building an ecosystem around its platform by allowing users to create and share their own recipes, and it plans to incentivize creators by paying them for new user sign-ups generated from their automations. This approach positions Poke less as a direct competitor to foundational models and more as an accessible distribution layer for practical, everyday AI actions.

By abstracting away the underlying AI models and eliminating the friction of app downloads, Poke is betting that the path to mass adoption for AI agents lies in meeting users where they already are: their text message inbox. Its success will depend on whether this convenience can deliver enough value to become an indispensable part of a user's daily routine.