The Strategic Value of AI Personality

Cognition’s acquisition of The Interaction Company (the team behind the AI assistant Poke) for a reported "low nine-figure" sum highlights a shift in AI product strategy: moving beyond pure utility to focus on the quality of human-AI interaction. While Devin is a specialized coding agent, Cognition aims to evolve it from a robotic tool into a persistent, relatable colleague. By integrating Poke’s conversational model—which utilizes humor, slang, and a friendly tone—Cognition intends to make the experience of working with AI feel more natural and engaging.

Solving for Persistence and Orchestration

Beyond personality, the acquisition addresses critical functional gaps in current agentic workflows. Currently, Devin is limited to executing individual pull requests. The integration of Poke’s architecture offers two primary technical advantages:

• Task Orchestration: Poke can act as a high-level manager, coordinating multiple Devin sessions to handle complex, multi-step engineering projects.
• Contextual Memory: Poke’s interaction model enables a more persistent relationship with the user, allowing the agent to remember tasks and context across different sessions. This creates a "persistent co-worker" dynamic that is currently missing from most isolated AI coding tools.

The Economics of Agentic Platforms

Despite significant traction—with over 100 million messages exchanged by hundreds of thousands of users in just three months—Poke faced challenges in achieving profitability due to high operational costs. The acquisition provides the necessary infrastructure and model support (specifically Cognition’s SWE-1.7 model) to scale these interactions reliably. This move underscores a broader trend in the AI startup ecosystem: the difficulty of maintaining high-volume, low-latency consumer agent platforms without deep integration into specialized, high-performance backend infrastructure.