Reid Hoffman and Mark Pincus are co-founding Prentis, a new AI lab in early-stage discussions to raise $100 million. The startup centers on automating routine computer tasks, positioning this capability as the next frontier beyond code generation.
Prentis' thesis diverges from the current AI consensus. Most attention flows toward large language models writing code or assisting developers. Prentis instead targets workflow automation, where AI systems handle repetitive digital work that doesn't require custom programming. Think data entry, email sorting, spreadsheet management, document processing, and browser-based task execution.
This category of work represents enormous economic value. Knowledge workers spend significant time on these tasks despite their low cognitive demands. Automating them creates immediate ROI for enterprises without requiring deep technical integration or retraining.
The timing signals confidence in agentic AI. Recent advances in reasoning models and tool-use capabilities have made it technically feasible to build systems that understand user intent, navigate software interfaces, and execute multi-step workflows autonomously. Companies like Anthropic, Google, and OpenAI have released agents in alpha or limited releases. Prentis bets this becomes the dominant application layer.
Hoffman's track record with LinkedIn and investments in OpenAI lends credibility. Pincus, the Zynga co-founder, brings gaming expertise in systems design and user engagement. Neither founder has announced a CEO yet, suggesting the lab remains in early assembly.
The $100 million raise, if completed, positions Prentis as a well-funded competitor in a crowded space. Startups like Anthropic's Claude-powered automation tools, Replit's AI agent initiatives, and internal efforts at major cloud providers already pursue similar goals. The differentiation will depend on execution speed and whether Prentis can build agents that reliably handle domain-specific