Tencent is trying to turn WorkBuddy from a workplace assistant into a platform other companies can build around. Caixin Global reported that Tencent launched an open platform for WorkBuddy on September 2, opening its underlying agent functions to outside software developers, device manufacturers, and providers of industry applications.
The company says the first phase includes more than 100 ecosystem partners. Caixin reported that more than 30 software vendors are involved, including GF Securities, Weimob, and Beisen, while nine co-branded smart devices come from brands including Rokid and iFlytek. The launch is a concrete effort to build a network around WorkBuddy rather than keep the product inside a closed Tencent application.
Tencent’s approach reflects a broader competition over where workplace agents will live. An individual assistant can summarize a document or answer a prompt. An open platform tries to connect agents to the software, devices, and specialized data that a customer already uses. That is a more ambitious model, but it also raises harder questions about security, integration, billing, and who is responsible when an agent produces a wrong answer.
More Than 100 Partners Give WorkBuddy a Wider Reach
The partner count is the most immediate measure of Tencent’s launch. Caixin said the initial group includes software companies and hardware brands, while a Futu report based on the launch announcement said more than 30 industry applications were integrated across more than 20 sectors. The report named finance, law, health care, education, and public welfare among the sectors.
Those figures show the intended scope, not verified customer usage. An ecosystem partner may provide a tool, device, connector, or application without proving that a large number of end users are already active. Tencent has not publicly disclosed a common metric for transactions, paying business customers, completed tasks, or revenue generated through the new platform.
Still, the launch changes the strategic position of WorkBuddy. Tencent previously used the product as an AI workplace assistant. EastFrontier reported in August that WorkBuddy had a large monthly web audience. The open platform shifts the question from how many people visit a service to how many products and specialist applications can attach to it.
Tencent Cloud’s stated goal is to make agent capabilities available to three groups: developers, hardware manufacturers, and industry application providers. Each group has different needs. Developers need tools and interfaces. Hardware makers need a route to integrate AI functions into devices. Industry providers need a way to connect existing expertise and workflows without handing over the core of their own service.
Data Control Is Central to the Enterprise Design
Caixin reported that partners can keep core assets on their own servers, with WorkBuddy functioning as a forwarding layer. That detail may be as important as the headline partner count. Enterprise customers often want to use an AI agent without moving proprietary records, client information, or specialized business knowledge into a third party’s general platform.
The arrangement described by Caixin is not a complete account of the platform’s security architecture. The report does not disclose technical specifications, data-retention periods, cross-border processing rules, or the exact permissions an agent receives. But it signals that Tencent is trying to address one of the most common barriers to enterprise-agent adoption: a customer’s reluctance to expose its underlying assets.
This concern has become sharper as Chinese companies connect AI to practical tasks. Tencent’s own AI search service with Elastic is designed to help models and agents retrieve enterprise information. A tool that can retrieve data becomes more useful when it is integrated with a workflow, but it also demands clear decisions about which documents, systems, and actions are available to the agent.
Tencent is not alone in addressing this problem. Alibaba Cloud has opened Wanyou Wujie to public testing as a multi-agent collaboration workspace, while ByteDance has been combining its workplace tools under Doubao Work. The competitive issue is not only model quality. It is whether a provider can let customers connect AI to internal work without creating an unacceptable security or governance burden.
The Business Model Is Still Taking Shape
Caixin reported that revenue sharing for the WorkBuddy ecosystem is still at an early stage, and that hardware partners may bill separately for devices while WorkBuddy collects cloud-subscription fees. That leaves important questions unresolved. A platform can attract many partners, but the value distribution between Tencent, an application provider, a device maker, and a customer must be clear enough for each to keep participating.
The initial partner structure could help Tencent build a faster route into specialized use cases. GF Securities, Weimob, and Beisen bring business domains that a general AI assistant would not possess on its own. Hardware partners can put WorkBuddy into devices, creating more places where users encounter the service. Yet the company will need to prove that these connections produce a better outcome than a customer simply using a competing agent or an in-house model.
The open platform also makes WorkBuddy part of a larger market for agent operating systems. A user may not care which model sits behind a task if the service can access the right information, follow permissions, and complete work reliably. Tencent’s effort to build a partner network recognizes that model capability alone does not create that experience.
The September 2 launch is therefore an ecosystem event, not evidence that WorkBuddy has already become the default enterprise-agent platform in China. Tencent has disclosed an initial partner count and a set of target users. It has not disclosed the usage, revenue, or retention data that would show whether the network is working at scale.
What it has done is open a new front in China’s enterprise AI contest. WorkBuddy is no longer being presented only as a Tencent tool. The company now wants it to become a layer that hosts other software, devices, and specialized services. The success of that strategy will depend on whether those partners can deliver secure and useful agent experiences that customers are willing to pay for repeatedly.
