Tsinghua Prof Li Ning: Low AI ROI Tied to Unchanged Orgs: what are the key facts?

Tencent's WorkBuddy has surpassed 10 million monthly active users, while ByteDance merged Feishu and Doubao to launch Doubao Work.

Tsinghua Prof Li Ning: Low AI ROI Tied to Unchanged Orgs: what are the key facts?

Alibaba consolidated three agent product lines into Qianwen Office, and 360 and Kingsoft also released AI office products.

Tsinghua Prof Li Ning: Low AI ROI Tied to Unchanged Orgs: what are the key facts?

MIT research shows that about 95% of enterprise generative AI investments in 2025 yielded no measurable financial returns.

Tsinghua Prof Li Ning: Low AI ROI Tied to Unchanged Orgs: what are the key facts?

A BCG report indicates that only 5% of enterprises achieved meaningful AI returns.

Tsinghua Prof Li Ning: Low AI ROI Tied to Unchanged Orgs: what are the key facts?

Tsinghua professor Li Ning argues that enterprises need to transform their organizations, not just adopt AI tools.

Physix Frontier · News Briefing Card (enbrief · Sep 23, 2026)

Tsinghua Prof Li Ning: Low AI ROI Tied to Unchanged Orgs

KEY FACTS

Tencent's WorkBuddy has surpassed 10 million monthly active users, while ByteDance merged Feishu and Doubao to launch Doubao Work.
Alibaba consolidated three agent product lines into Qianwen Office, and 360 and Kingsoft also released AI office products.
MIT research shows that about 95% of enterprise generative AI investments in 2025 yielded no measurable financial returns.
A BCG report indicates that only 5% of enterprises achieved meaningful AI returns.
Tsinghua professor Li Ning argues that enterprises need to transform their organizations, not just adopt AI tools.

KEY DATA

10 millionTencent WorkBuddy MAU

95%MIT research: share of enterprises with no financial returns

5%BCG research: share of enterprises with meaningful AI returns

30-40 billion USDGlobal enterprise generative AI spending

PHYSIX OBSERVATION

Big tech is piling into AI office tools, yet overall enterprise returns remain extremely low, showing that stacking tools cannot fix organizational misalignment. Li Ning nails the key point: AI productivity does not match Industrial Age organizational structures, and without rebuilding task flows and incentive mechanisms, even the strongest agents are just expensive add-ons.

Source: enbrief original report ↗