Prentis Raises $100M to Build AI Agents for Office Automation

Summary

Prentis, an AI research lab co-founded by serial entrepreneur Ritankar Das alongside LinkedIn co-founder Reid Hoffman and gaming executive Mark Pincus, is in advanced talks to raise $100 million at a $1 billion valuation. The startup, launched in April 2026, focuses on developing computer-use models and AI agents designed to automate routine office workflows—such as handling insurance claims and processing customs duty refunds—by learning how workers navigate documents and business systems.

The company has already demonstrated significant market traction, securing contracts worth up to $50 million from customers in healthcare, manufacturing, and retail & e-commerce sectors. According to investor materials, Prentis projects a $75 million annualized run rate by Q3 2026, though the company notes these figures are performance-dependent estimates based on contracted fees equal to 20% of realized savings.

Prentis claims its Hive-32B model outperforms frontier models including OpenAI's GPT-5.4 and Anthropic's Claude Opus 4.6 on computer-use benchmarks (WindowsAgentArena and ScreenSpot-v2). The company's competitive advantage centers on deploying a significantly smaller, more cost-efficient model—claiming approximately 10 times lower cost per task than frontier APIs—making it more practical for everyday office automation workflows.

The startup enters a crowded market where Anthropic, OpenAI, and Mira Murati's Thinking Machines Lab are also developing computer-use AI agents. Notably, Anthropic has already acquired Seattle-based Vercept to accelerate its capabilities in this category, signaling the strategic importance of office automation as a major AI use case.

• Founded by Ritankar Das (CEO) with backing from Reid Hoffman and Mark Pincus, the lab competes directly with Anthropic and OpenAI in the high-stakes computer-use AI agent space