No. 117
2026-11-13
the AI pulse — brewed, not raw
 AI CRITIQUE

Baidu founder Robin Li: The next phase of AI is the intelligent organization

Read the article    English (automatic translation)

Summary

Baidu co-founder Robin Li argued in his Baidu World 2026 speech that the next phase of AI adoption requires companies to restructure around cross-functional tasks, documented expertise, and measurable workflows. Li cited the increase in Baidu's daily active AI agents from approximately 5,000 to 37,000 as evidence supporting this thesis.

Why it matters

This statement illustrates how a major Chinese tech company frames the bottleneck in AI deployment not just as a technical issue, but as an organizational one. This is precisely where most organizations are currently stuck. The classification of 'producers and sellers' and task-based work distribution could influence how Chinese companies restructure their headcount and KPIs by 2027. Additionally, it offers a rare internal perspective on Baidu's agent adoption metrics, which serves as a useful benchmark for the pace of enterprise AI rollout.

Highlights

  • Three principles: Task-based (not role-based) work, documented expertise for human and agent reuse, and workflow-based measurement
  • Baidu internal AI agents: Daily active 5,000 → 37,000; weekly tasks: 84,000 → 1 million+ (within one year)
  • Work is divided into 'producers' (creating products/services) and 'sellers' (communication/trust)

Critical take

The three 'principles' are essentially standard business process management advice (cross-functional teams, knowledge management, KPIs) repackaged with AI terminology. The impressive agent and task counts are self-reported by Baidu without independent verification. No information is provided regarding failure rates, costs, or which tasks were truly automated versus merely routed through an agent interface.

This critique was written by the local AI model Qwen3.8-27B. I chose the story; I do not review each critique before it is published.
Back to the issue