BlueRun Ventures Leads $8 Million Seed Round in Yuanli Intelligence | BlueRun Portfolio

Yoolee AI, an enterprise agent company, recently raised an $8 million seed round led by BlueRun Ventures, with Lightspeed China Entrepreneurs Fund participating.

BlueRun Ventures believes that Yoolee AI builds its Enterprise Agent system around reinforcement learning training, using centralized training with distributed deployment to enable agents to self-optimize within each company's unique data environment. This approach, the firm argues, solves the persistent problems of difficult migration and limited generalization that plague AI software in enterprise settings, achieving a critical leap for enterprise intelligence from "knowledge reuse" to "capability growth."

Yoolee AI was founded by Fan Zhang, former COO of Zhipu AI, and focuses on building reliable, self-evolving agent workers through RL for Business as its core methodology. The company aims to convert foundational model capabilities into quantifiable business outcomes through commercial scenario reinforcement learning, driving real-world deployment of AI agents.

Fan Zhang is a serial entrepreneur in AI. He studied artificial intelligence in France and later worked at a national research center on machine translation. After returning to China in 2010, he led intelligent voice product development (Siri-like products) at Sogou and Tencent, then founded Miaoji Travel and served as CTO at SouChe. In 2022, he founded Yuanyin Intelligence, which was later acquired by Zhipu AI. He joined Zhipu AI as COO in 2023 and left in June this year to start his next venture.

Zhang said that current mainstream Agent-to-B models mostly rely on "customization plus SFT (supervised fine-tuning)" with fixed workflows — high deployment costs, slow scenario adaptation, strong coupling with the base model, and massive upgrade costs with every model iteration that prevent scaling. Yoolee AI sees reinforcement learning as introducing a new learning paradigm: business-outcome-oriented commercial reinforcement learning to build self-evolving productivity agents that continuously create commercial value in real business environments.

Reinforcement learning has achieved breakthroughs in scenarios with clear feedback like code generation and mathematical reasoning, but faces far more complex environments in enterprise operations. Yoolee AI's core goal is to build authentic reward functions and optimization mechanisms by combining business priors with tacit knowledge, enabling AI to continuously evolve in complex commercial decision-making.

There is growing industry consensus that reinforcement learning-driven AI agents represent a new global direction for the AI industry. Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, has similarly proposed using reinforcement learning to help companies customize AI models based on KPIs like revenue and profit.

Yoolee AI is reportedly already collaborating with multiple leading companies across sectors to explore applications and deployment of commercial reinforcement learning in different domains.

ZeroG Aircraft Industry Raises Over 100 Million RMB in Series A+ Round | BlueRun Portfolio BlueRun Ventures Continues Backing RevoSight in 100 Million RMB Series A+ Round | BlueRun Portfolio Tashi Zhihang Raises $122 Million in Angel+ Round | BlueRun Portfolio