- New framework measures AI-era entrepreneurial capabilities through real-world outcomes rather than degrees or test scores
- Move comes as Zhejiang expands funding, policy support and talent programs for AI-driven solo founders
Hangzhou has introduced what organizers call China’s first standardized framework for measuring the capabilities of AI-powered solo entrepreneurs.
This rollout comes as Zhejiang Province has gone into overdrive to build an ecosystem around so-called “one-person companies,” or OPCs for short.
The Artificial Intelligence Open Source Ecosystem Academy at Hangzhou Dianzi University released the OPC Nine-Level Talent Competency Standard on June 8, creating a system to assess, test and certify skills associated with OPC, short for One Person Capability.
The framework reflects a broader push by Zhejiang to support individuals who use OpenClaw-inspired AI tools to build businesses with minimal staffing.
Policymakers and industry groups have rolled out new funding programs, regulatory initiatives and talent-development schemes in recent months.
Three broad categories
The standard divides OPC capabilities into three broad categories — learners, builders and creators — and nine levels, from L1 to L9. Each level includes measurable assessment criteria.
At the lower end, the framework evaluates a person’s understanding of AI tools and their ability to use them for tasks such as writing documents and creating presentations.
Mid-level assessments focus on identifying business scenarios for AI adoption, building automated workflows and developing AI applications.
Higher levels measure commercial outcomes, including securing customer orders, completing business transactions and coordinating teams of AI agents.
“The old evaluation systems use old rulers to measure new capabilities,” said Fan Jingjing, founder of AI startup Datawhale and vice chairman of the academy. “Today, if we want to assess one-person AI full-stack capabilities, we need a new ruler.”
Fan said the framework differs from traditional talent evaluation systems that emphasize educational credentials, résumés or exam scores.
Instead, it relies on practical indicators. Level 2 candidates, for example, are assessed through publicly available platform data.
Meanwhile, Level 6 candidates must demonstrate that they have secured real customer orders. Level 8 assessments track token consumption linked to AI-agent operations.
Hub for OPC startups
The release comes amid a rapid expansion of Zhejiang’s OPC ecosystem. In April, 39 organizations, including universities, industrial parks, large-model developers and OPC entrepreneurs, formed the Zhejiang OPC Innovation Ecosystem Service Alliance.
Authorities in Hangzhou’s Shangcheng District have also launched a policy initiative aimed at turning the district into a hub for OPC startups.
The program includes a dedicated 1 billion yuan ($147 million) fund, with an initial 100 million yuan already committed.
Meanwhile, Hangzhou’s market regulator has introduced 12 measures to support the sector, including workplace registration mechanisms, data ownership protections and regulatory sandbox programs.
Xu Xuchu, dean of the Artificial Intelligence Open Source Ecosystem Academy, said earlier that the school operates as “a college without walls” that brings together government agencies, communities, companies and investors to train talent.
In May, Hangzhou released a draft three-year AI-plus-OPC action plan that targets the creation of more than 100 OPC communities and the attraction of more than 30,000 innovation and entrepreneurship professionals by 2028.
