Tezign founder Fan Ling: enterprises need two 'world models', not a better LLM
量子位 · wechat · 2026-10-22
Reporting on Tezign founder Fan Ling's WTCC Shanghai talk: MIT NANDA data shows 95% of enterprise GenAI pilots yield no measurable P&L impact while 90%+ of employees use personal AI tools — AI stalls at the individual level. Fan argues companies must build two distinct world models: an outward 'customer world model' capturing subjective preferences, and an inward 'enterprise world model' encoding processes, assets, rules and tacit knowledge. Merging them blurs both.
Tezign has enlisted domain experts to capture real judgment paths (CreativeTrajectory), already at tens of thousands of entries, and calibrates AI personas against humans in identical behavioral economics experiments. It launched an Agentic Transformation co-build program and frames enterprise agents as a Signal-Plan-Create-Evals loop: 'agents push work to 80 points, humans take it to 120.'
More from Companies & People
- xAI launches Super Intelligence Partner Network with Microsoft, Google, NVIDIA among launch partners — NicoVerderosa · 2026-10-23
- Anthropic's Alex Albert: Claude Duet could matter more to small businesses than Square — alexalbert__ · 2026-10-23
- FAS launches a $174K Metascience Fellowship embedded high up at agencies like DOE — Afinetheorem · 2026-10-23
- Sierra Summit London feels underwhelming vs SF; production agent metrics still rare — DimlyDimwitted · 2026-10-23
- Bindu Reddy: I'd trade my Meta stock for Anthropic at $1.5T or OpenAI at $1T — bindureddy · 2026-10-23
- CTO exits blood-testing startup of Elizabeth Holmes' partner, dubbed "Theranos 2.0" — annatonger · 2026-10-23