Conversation with Tencent Cloud Vice President: Enterprise intelligence is booming, and platform usage will at least double this year

Zhidongxi Author | Jiang Yu Editor | Xinyuan Zhidongxi reported on July 18 that during the 2026 WAIC, Tencent intensively released products related to embodied intelligence and enterprise-level intelligence. In terms of embodied intelligence, Zhang Zhengyou, chief scientist of Tencent, director of RoboticsX Laboratory, and director of Futian Laboratory, released three embodied models: Hy-Embodied-VLA-0.5, Hy-Embodied-VLM-1.0, and Hy-Embodied-RxBrain-1.0. At the same time, Tairos Agent embodied agent framework and Apexio agent were officially unveiled. The Tairos platform launched by RoboticsX Lab in 2025 has also been upgraded simultaneously, and will provide external open source capabilities and services such as models, agents, and development tools. In terms of enterprise-level agents, Tencent Cloud Vice President Wu Yunsheng released the overseas version of the enterprise-level agent development platform ADP 4.0 on site. At the same time, the universal desktop intelligent WorkBuddy App version was launched simultaneously. After the meeting, Tencent Cloud Vice President Wu Yunsheng communicated with Zhixixi and other media for nearly an hour and a half, talking about specific issues such as enterprise intelligence's access to old systems, Token costs, FDE delivery models and knowledge governance. Wu Yunsheng mentioned that one of the outstanding problems faced by enterprises in deploying intelligent agents is that it is difficult for intelligent agents to call existing information systems. Many old systems lack mature API or MCP interfaces. Enterprises need to first supplement the interfaces and modify the system before intelligent agents can be embedded in business processes. With the improvement of Agent Loop capabilities, some tasks that were limited by retrieval and workflow capabilities in the past are now ready for implementation. Tencent Cloud enterprise intelligence-related customer demand and platform usage have at least doubled year-on-year this year. The following are the key questions and answers compiled by Zhidongxi: 1. Old systems are still a threshold for the implementation of enterprise intelligence. Q1: When enterprises deploy intelligence, what is the most prominent resistance at present? Wu Yunsheng believes that problems such as organizational processes, data quality and ROI still exist, and are more prominent this year.