Alibaba Launches 'UniFuncs': World's First Enterprise-Native AI Work Platform
Alibaba launched 'Wukong', the world's first enterprise-native AI work platform, deeply embedded in DingTalk. DingTalk underwent CLI-level reconstruction enabling AI Agents to directly operate thousands of built-in enterprise functions. Alibaba simultaneously established the ATH business group under CEO Eddie Wu.
Alibaba Launches 'Wukong' AI-Native Work Platform: DingTalk CLI Reconstruction Enables AI Agents to Operate Enterprise Functions
The Core Event
On March 16, 2026, Alibaba Group announced the establishment of the Alibaba Token Hub (ATH) business group, directly led by CEO Eddie Wu. The following day, at the AI DingTalk 2.0 annual product launch event, DingTalk officially unveiled the 'Wukong' AI-native work platform.
Wukong Platform Positioning
Wukong is not a chatbot layered on top of DingTalk — it represents a fundamental architectural transformation. DingTalk underwent **CLI-level reconstruction**, with core code rewritten to enable AI Agents to directly operate DingTalk's thousands of built-in enterprise functions. This transforms the work paradigm from "humans using software" to "AI using software."
Users can issue natural language commands through DingTalk, and Wukong coordinates multiple AI Agents to automate daily tasks including document editing, spreadsheet updates, meeting transcription, and research — enabling 24/7 continuous operations.
Core Capabilities
Multi-Agent Coordination: Wukong orchestrates multiple AI Agents working in concert, each handling specialized tasks — more like an AI team than a single model doing everything.
Security Sandbox: All AI operations run within security sandbox environments, inheriting the organization's permission structure. Token usage and costs are transparently tracked.
Skills Marketplace: Wukong integrates B2B capabilities from across Alibaba's ecosystem as "skills," aiming to create "the world's largest B2B skills marketplace."
Cross-Platform Expansion: Future plans include connecting with Slack, Microsoft Teams, and WeChat.
Organizational Signal
The ATH business group reports directly to CEO Eddie Wu — not a divisional president — signaling Alibaba's highest-level commitment to its AI Agent strategy. Wukong is not just a product launch; it's a landmark organizational restructuring for the AI era.
Competitive Landscape
| Platform | Parent | Entry Point | User Base |
|----------|--------|-------------|-----------|
| Wukong | Alibaba | DingTalk (20M orgs / 800M users) | China #1 |
| Copilot | Microsoft | Teams/Office 365 | Global #1 |
| Agentforce | Salesforce | CRM | Enterprise Sales |
| SnowWork | Snowflake | Data Cloud | Data-intensive |
DingTalk's massive China user base of 20 million enterprise organizations and 800 million users is Wukong's core moat. However, international expansion faces strong competition from Slack and Teams.
Industry Impact
1. **CLI reconstruction as key innovation**: Transforming software into an API/CLI that AI can directly call may become the standard architecture for enterprise AI platforms
2. **"AI uses software" era**: The paradigm shift from AI assisting humans to AI directly operating software
3. **China enterprise AI market reshaping**: Alibaba, ByteDance, and Tencent will compete intensely in the enterprise AI Agent space
4. **Hardware-software integration**: DingTalk simultaneously launched AI hardware (A1Pro, AI earbuds H1)
Potential Risks
- Wukong is currently in invite-only testing; real-world effectiveness remains to be validated
- CLI reconstruction stability and compatibility need extensive enterprise validation
- Enterprise trust in AI automation takes time to establish
- International expansion faces data compliance and cultural adaptation challenges