Wukong centres Alibaba’s enterprise AI push on agent orchestration

Alibaba Group has introduced Wukong, an enterprise AI platform built around coordinating multiple agents within business workflows. The platform is positioned as a productivity layer that brings task execution, rather than standalone AI outputs, into everyday enterprise operations.

Wukong operates through a single interface that can orchestrate multiple agents across local systems, browsers and cloud environments. It supports tasks such as document editing, spreadsheet updates, approvals, meeting transcription and research, focusing on execution across common enterprise processes.

Agent workflows tied to enterprise tools

The platform is tightly integrated with DingTalk, Alibaba’s enterprise collaboration platform with more than 20 million corporate users. Wukong can be accessed either as a standalone desktop application or embedded within DingTalk, allowing users to invoke agent workflows directly within existing communication environments.

Alibaba is extending this approach beyond its own ecosystem. Wukong is expected to connect with messaging platforms including Slack, Microsoft Teams and WeChat, enabling access across mobile and desktop environments without requiring a shift in collaboration tools.

This integration model places Wukong within the operational layer of enterprise software rather than as a separate AI interface, with agents triggered from within existing workflows.

Internal reorganisation shapes platform direction

The launch follows Alibaba Group’s reorganisation under the Alibaba Token Hub Business Group, with Wukong positioned as the flagship enterprise AI platform under this structure. The platform is developed by the Wukong Business Unit, indicating a dedicated focus on agent-driven enterprise systems.

The shift reflects a more centralised effort to build AI capabilities into enterprise-facing products. Wukong’s architecture is built on a reworked version of DingTalk’s interface, rebuilt as a command-line interface and open API layer, giving the platform direct access to enterprise functions and enabling autonomous task planning and execution.

This design allows Wukong to generate instructions and coordinate workflows across multiple agents without requiring continuous human input.

Security and execution in business environments

Security features are embedded into the platform’s core design, with identity authentication, access controls and enterprise sandboxes intended to address operational risks in business environments. These controls are positioned as necessary for enabling agents to act across systems while maintaining governance.

Wukong is currently available through invitation-only beta testing, signalling a controlled rollout as Alibaba refines deployment in enterprise settings.

Ecosystem integration and industry use cases

Alibaba is extending Wukong’s capabilities through integrations with its broader ecosystem, including Taobao, Tmall, 1688, Alipay and Alibaba Cloud. These services will be exposed as modular agent skills, alongside support for third-party integrations, enabling the platform to handle functions such as storefront design, supplier management, payments and infrastructure scheduling.

The company has also introduced One-Person Team solutions across ten sectors, including e-commerce, manufacturing, legal services, finance and software development. These packages combine industry-specific agent skills to allow individuals or small teams to manage workflows typically handled by larger organisations.

The approach positions Wukong as both an enterprise tool and a scaling mechanism for smaller operators, using agents to extend operational capacity without increasing headcount.