On March 6, 2025, the Chinese AI team launched the general-purpose AI agent Manus. Users only need to say “help me design a 3D game”, and Manus can independently call the modeling software, write code, deploy services, and finally deliver a playable demo. As can be seen from the currently released demonstration video, Manus not only encapsulates and applies complex projects, but also has the ability to think independently and can achieve multi-agent collaboration, which has attracted widespread attention. At the same time, the implementation of enterprise-level AI Agents has also become a key direction for industrial transformation – Yonyou BIP Enterprise AI Core Product “Zhiyou” is the culmination of this change.
If Manus is a “personal intelligent assistant”, then Zhiyou is a “big manager of enterprise intelligence”. The two share the technical logic of agents, but Zhiyou focuses more on enterprise-level tasks. Yonyou defines Zhiyou as a super entrance to work, and users only need to issue instructions, and Zhiyou can mobilize hundreds of agents to automatically build business processes and execute tasks. Users can efficiently control the overall situation by only approving at key decision-making links.
Whether it is Manus scheduling tools in personal office scenarios, or Zhiyou coordinating financial AI, human AI and other agents in enterprise-level scenarios, both adopt the agent collaboration framework of “requirements understanding, task disassembly, tool calling, and result delivery”. Users do not need to organize complex prompts, just use natural language instructions like daily communication, such as “Activate the East China 618 Promotion Plan”, and Zhiyou can automatically analyze requirements and call resources to achieve business intelligence.
Zhiyou is not only the dispatcher of enterprise-level agents, but also the innovator of enterprise intelligent collaboration models. Its core capabilities are mainly reflected in two aspects: the innovation of intelligent scheduling engines and intelligent collaboration models.
First of all, Zhiyou has changed the traditional model of “people looking for tools” to “tools looking for people”. When users put forward requirements, such as “generating quarterly sales analysis reports”, Zhiyou will automatically plan tasks, call data analysis agents to collect sales data, count sales in each region, and generate visual reports. At the same time, it will dispatch the market analysis agent to compare historical data, identify sales trends, and call the intelligent writing agent to generate report summaries. If abnormal sales are found in a certain area, Zhiyou will automatically alert the relevant person in charge and provide optimization suggestions to ensure that the company responds quickly to market changes.
Secondly, Zhiyou creates a new enterprise intelligent collaboration model through “supergroup”. When the project is launched, Zhiyou will automatically create a supergroup to centralize employees, agents, approval processes, and knowledge bases on one platform for one-stop collaboration. If a link is stagnant for too long, Zhiyou will automatically remind the person in charge, and if the task is urgent, it will automatically notify the management. At the same time, the results of the meeting discussion will automatically generate a wiki document archive to ensure information precipitation and sharing. In addition, Zhiyou can also analyze project data, provide optimization suggestions, and improve business operational efficiency. Zhiyou is not only the dispatcher of enterprise-level agents, but also the innovator of enterprise intelligent collaboration models. Its core capabilities are mainly reflected in two aspects: the innovation of intelligent scheduling engines and intelligent collaboration models.
Zhiyou has promoted the deep transformation of enterprise intelligence, bringing about three major evolutions of interaction mode, functional boundaries and data value.
01 Interaction mode upgrade
While traditional office operations require manual operation of multiple systems, Zhiyou makes intelligent assistants the entrance to work, allowing users to describe their needs, and the system can automatically match processes and execute tasks, greatly improving work efficiency.
02 Functional boundaries expanded
Zhiyou expands the AI toolset into an autonomous and collaborative agent network, which can call market analysis, supply chain management, intelligent approval and other functions across systems to ensure smooth task execution. AI is no longer a single tool, but has become the core support for enterprise intelligent operations.
03 Data value leap
Zhiyou has evolved enterprise data from “static storage” to “dynamic intelligence”. Relying on the BIP data base, Zhiyou can not only analyze historical data, but also continuously learn and optimize algorithms, allowing enterprise processes to evolve continuously. For example, in the manufacturing industry, Zhiyou can monitor production data, optimize quality inspection processes, and improve yield rates.
IDC research shows that in the next year and a half, the application of generative AI by enterprises will become a core strategy, and nearly 40% of core business will be affected by AI. Corporate investment in AI technology is expected to grow by more than 25%.
The goal of Yonyou BIP Enterprise AI is to deeply integrate AI technology and AI large models with the business process and operation management of enterprises, and can fully mine and utilize core business data within the enterprise and the industry, so as to fully release the value and potential of data, make enterprise decision-making and operation smarter and more excellent, and make the business society more efficient. In this AI Agent revolution, Yonyou has laid out the core areas of enterprise-level intelligent collaboration, and will continue to promote the deep integration of AI and business in the future, helping enterprises move towards a new era of digital intelligence.
