Alibaba readies enterprise AI agent built on Qwen

Alibaba plans to roll out an agentic AI service for enterprises based on its Qwen model and may announce the product as soon as this week, people familiar with the matter said.

The tool was developed by the team behind DingTalk, and is said to be able to operate computers, while including features to protect data.

The company intends to integrate services, but has not disclosed pricing or the initial scope of integration.

Eddie Wu, CEO, promised over US$53 billion of investment in AI last year, and Alibaba has named artificial general intelligence as a strategic goal.

The launch comes as Alibaba faces scrutiny over its AI strategy.

🔗 Source: Bloomberg

🧠 Food for thought

Implications, context, and why it matters.

Alibaba’s enterprise AI agent may build on consumer agent features, though commercialization remains unproven

The enterprise offering draws on systems already running at large scale in consumer products.
Alibaba’s Qwen App uses agentic AI for tasks such as ordering food through Taobao Instant Commerce and booking travel on Fliggy, reaching over 100 million monthly active users within two months of launch 1.
DingTalk, the team behind the workplace product, has rolled out a Tongyi Qianwen-powered AI agent to 700 million users for work tasks such as document summaries and business trip bookings on third-party platforms 2.
The enterprise release remains planned, while pricing and early integration coverage have not been disclosed.

The enterprise release lands in a price squeeze and a wide revenue gap versus US rivals

It enters a market where Alibaba has already cut AI model prices by as much as 97% 3.
Those cuts leave some Chinese models up to 50 times cheaper than US counterparts, which suggests pricing will favor broad adoption rather than fast margin capture 3.
Alibaba has reported triple-digit AI revenue growth for six straight quarters, yet its estimated $1 billion AI revenue run-rate sits roughly 13 times below Microsoft Azure’s 3.
The gap ties back to China’s smaller enterprise software market, and the agentic product targets that shortfall while progress will take time 3.