Enterprises need two world models for AI

What happened: At the WTCC forum in Shanghai, Tezign founder Fan Ling said enterprises need two world models for AI: a customer world model for understanding external preferences and an enterprise world model for internal processes, rules and tacit knowledge. Key numbers: An MIT Why it matters: The content argues that combining the two models makes both weaker, and that enterprise agents need a long-running loop of Signal, Plan, Create and Evals. Tezign also announced an Agentic Transformation co-building plan with universities, research institutions, enterprises and technology partners.

At the WTCC forum in Shanghai, Tezign founder Fan Ling said enterprises need two world models for AI: a customer world model for understanding external preferences and an enterprise world model for internal processes, rules and tacit knowledge.

The content argues that combining the two models makes both weaker, and that enterprise agents need a long-running loop of Signal, Plan, Create and Evals. Tezign also announced an Agentic Transformation co-building plan with universities, research institutions, enterprises and technology partners.