Why Cognition bought Poke: AI personality is becoming a competitive advantage
Cognition acquired The Interaction Company, developer of the AI assistant Poke, in a deal valued at low nine figures. The acquisition aims to integrate Poke’s conversational personality and interaction model into Cognition’s coding agent, Devin. Poke leverages Cognition’s infrastructure to improve speed and reliability while utilizing Cognition’s SWE-1.7 model for future tasks. The deal highlights a strategic industry shift where user experience and agent personality are becoming as critical as
Analysis
TL;DR
- Cognition acquired The Interaction Company, developer of the AI assistant Poke, in a deal valued at low nine figures.
- The acquisition aims to integrate Poke’s conversational personality and interaction model into Cognition’s coding agent, Devin.
- Poke leverages Cognition’s infrastructure to improve speed and reliability while utilizing Cognition’s SWE-1.7 model for future tasks.
- The deal highlights a strategic industry shift where user experience and agent personality are becoming as critical as underlying model capabilities.
- Long-term plans include using Poke to orchestrate multiple Devin sessions and maintain persistent memory across coding tasks.
Why It Matters
This acquisition signals a pivotal trend in AI development: the differentiation of products will increasingly rely on interaction design and persona rather than just raw computational power. For practitioners, it underscores the importance of building agents that feel like collaborative colleagues rather than sterile tools, which can significantly impact user adoption and retention. Furthermore, it demonstrates how specialized interaction models can be integrated into broader engineering workflows to enhance productivity and context management.
Technical Details
- Acquisition Structure: Cognition acquired The Interaction Company (developer of Poke) for a valuation in the low nine figures, combining Poke’s interaction layer with Cognition’s coding infrastructure.
- Model Integration: Poke is set to utilize Cognition’s latest software engineering model, SWE-1.7, for specific tasks, aiming to improve reliability and performance in orchestrating complex workflows.
- Platform Compatibility: Poke operates across major messaging platforms including iMessage, SMS, Telegram, and WhatsApp, having secured approval as the first AI agent on Apple’s Messages for Business platform.
- Workflow Orchestration: The technical vision involves Poke acting as an orchestrator for Devin, managing multiple pull requests (PRs) simultaneously and maintaining persistent memory across different coding sessions, addressing current limitations where Devin handles only one PR at a time.
- Usage Metrics: Over a three-month period, Poke users exchanged more than 100 million messages, indicating high engagement despite challenges with operational costs.
Industry Insight
- Personality as a Feature: Companies should invest in refining the "soft skills" of their AI agents, such as humor, slang, and proactive behavior, as these traits drive user preference and perceived value alongside technical accuracy.
- Hybrid Agent Architectures: There is significant potential in decoupling interaction layers from core execution engines, allowing specialized agents (like Poke) to manage context and user relationships while powerful models (like Devin) handle complex task execution.
- Sustainability Challenges: High user engagement does not guarantee profitability; developers must address the high cost of running interactive AI agents to ensure long-term business viability, potentially through optimized infrastructure or hybrid monetization strategies.
Disclaimer: The above content is generated by AI and is for reference only.