Enterprise AI Agents Move From Experiments to Real Deployment
Enterprise AI agents are software entities built on language models that can connect to data, tools, and cloud services so they can plan and complete multi-step tasks with minimal human input. As enterprises move from pilots to production, they are discovering that powerful Qwen AI models or similar systems are not enough: agents must run inside secure environments, connect to existing infrastructure, and fit into familiar workflows. This is driving a new race in AI agent development where cloud providers compete not just on model benchmarks, but on how well they support deployment, automation, and training. Alibaba and Tencent now present two contrasting answers to the same problem: one strategy optimises for scale and deep integration with cloud operations, while the other emphasises smaller language models, desktop tools, and lower compute demands.
Alibaba’s Qwen Cloud Aims at Full-Stack Enterprise AI Agents
Alibaba Cloud is expanding Qwen from a strong model into a full enterprise AI agents platform. Its latest model, Qwen3.7-Max, is available on Model Studio and, according to Artificial Analysis’s Intelligence Index, “ranked fifth globally and first among Chinese models” with a score of 56.6 points. The company is pairing this capability with Qwen Cloud, an AI-native cloud platform with three entry points: Skills for agents, a command line interface for workflow integration, and a website for human use. Core cloud products such as databases, big data, operations, and security now expose Skill-based and MCP-compatible tools so agents can call cloud services like functions. This design makes AI agent development part of everyday cloud operations, reinforcing Alibaba’s bet that large enterprises will choose tightly integrated, scalable infrastructure when rolling out production agents across teams and business units.
Tencent’s WorkBuddy Pushes Smaller Language Models and Desktop Agents
Tencent Cloud is taking a different route with WorkBuddy, an OpenClaw-compatible desktop AI agent for workplace automation that is now available to global users after an initial launch in its home market. Instead of centering only on flagship models, Tencent is “betting cheaper deployment and agent software can win users while rivals still emphasize chips and model scale.” WorkBuddy focuses on multi-step productivity tasks and lower compute demands, using smaller models where possible so enterprises can deploy agents on existing hardware and networks. Tencent supports this with investment in AI, including revenue growth of 9% year over year and capital expenditure rising 16% to RMB31.9 billion, as reported in recent filings. The Hy3 Mixture-of-Experts model is framed around practical utility and cost efficiency, underscoring Tencent’s view that accessible tools can matter more than headline parameter counts.
Different Market Segments: Cloud-Scale Integration vs Everyday Accessibility
Although both companies compete in enterprise AI agents, their strategies target different needs. Alibaba focuses on large organisations that want AI tied into complex cloud estates. Its Qwen Cloud platform, Skill-based access to more than 60 cloud products, and upgraded agent runtime environments show a priority on scalability, security, and centralised management. The JVS Agent Suite and JVS Mobile extend this further, promising autonomous agents and multi-agent collaboration across enterprise and mobile workflows. Tencent, by contrast, optimises for accessibility and rapid adoption. WorkBuddy brings AI agents to the desktop, where employees already spend their time, and smaller language models help lower cost and latency. For enterprises, the choice may come down to whether they need deep cloud integration for mission-critical systems, or lighter-weight tools that improve everyday productivity without large infrastructure changes.
OpenClaw and Interoperability: A New Battleground for Enterprise AI
A shared theme in both strategies is the push toward interoperable standards for AI agents, with OpenClaw emerging as an important signal. Alibaba’s JVS Claw Teams is built on the OpenClaw framework with cloud-native security, enabling round-the-clock operations and centralised distribution of proprietary Skills. JVS Mobile is also described as native OpenClaw, suggesting that its enterprise AI agents can coordinate across applications using a consistent format. Tencent’s WorkBuddy is likewise OpenClaw-compatible, giving enterprises a path to reuse tools and patterns across vendors instead of locking themselves into one stack. As enterprise AI agents mature, this kind of compatibility could become a key differentiator: Qwen AI models and WorkBuddy agents may compete directly, but OpenClaw support hints at a future where enterprises assemble mixed ecosystems of large-scale cloud agents and lighter desktop agents that still work together.