As an enterprise level AI application development platform meticulously built by SIE Information for enterprise efficiency, SIE AI platform (Shanmou GPT) has fully supported MCP (Model Context Protocol) and A2A (Agent to Agent, Automated Agent Cooperation Protocol) protocols, respectively from the dimensions of "vertical resource connection" and "horizontal collaboration network", providing a technical base for industrial intelligence, realizing the reuse, flow and cooperation of multi model and multi agent capabilities among different systems, promoting enterprise level applications from "functional islands" to "intelligent collaboration", and accelerating the landing of AI scenario applications "the last mile". Its core values are as follows:

With the rapid development of artificial intelligence technology, various AI models continue to emerge, covering multiple fields such as natural language processing, image recognition, and data analysis, providing powerful tools for enterprises to solve various complex problems. However, whether it is a large model or various small models, they often exist independently and lack effective interconnection mechanisms, forming a "model island", which leads to many difficulties for enterprises in the application process, such as repeated development, resource waste, and inability to share data, severely restricting the large-scale application of AI technology in enterprises.

Shanmou GPT has completed the integration of MCP services commonly used by mainstream enterprises in the market, such as maps, search, express delivery, office communication, etc. At the same time, it is gradually developing MCP services for SIE's independent and agent products, providing basic capabilities for the smooth implementation of enterprise AI application scenarios and achieving collaborative intelligence between different tools/services under model coordination. At that time, developers do not need to repeatedly develop the same or similar functions. They can directly use existing models that support MCP protocol as components to quickly build and combine AI applications that meet specific enterprise needs, achieving seamless integration and collaborative work between multiple models and tools, greatly shortening the development cycle of AI applications, and reducing the workload and cost of enterprise AI application development.

At the same time, Shanmou GPT supports custom addition and introduction of more third-party MCP protocols, providing two modes of remote server docking and local server installation/docking to support different scenario requirements.

（SIE MCP Market）

The A2A protocol is like an intelligent link, breaking the previous situation of isolation and difficulty in collaboration between various business systems and agents. This protocol is based on advanced communication and data exchange technologies, establishing a standardized, efficient, and flexible connection framework that enables seamless integration and data sharing of agents with different functions and application scenarios, ensuring the rapid flow and precise interaction of business information throughout the entire enterprise chain.

Based on the knowledge and best practices of SIE Information in the fields of finance and taxation, human resources, marketing, supply chain, research and development, production and manufacturing, SIE Shanmou GPT will sort out various agents involved in the whole scenario of "research production supply sales service management" and encapsulate them into a standardized A2A protocol to achieve efficient collaboration across applications and agents. For example, in the production process, the production equipment agent, quality inspection agent, and logistics scheduling agent can share information in real time, work together, automatically adjust equipment operating parameters, optimize quality inspection processes, and arrange logistics distribution plans based on production progress, thereby reducing friction and delays between links, optimizing the entire business process, and improving enterprise operational efficiency.

（The integration of industry know-how and AI capabilities）

In addition, by combining the Agent routing function, tasks can be intelligently assigned to the most suitable target Agent based on business requirements and data characteristics, further optimizing collaboration efficiency. For example, in the process of enterprise order processing, the order agent assigns order tasks to corresponding production agents, procurement agents, or logistics agents through agent routing based on factors such as order type and urgency, ensuring that orders can be processed quickly and accurately, and improving the efficiency of enterprise order processing and customer response speed.

Looking forward to the future, SIE Information will continue to deepen the technological innovation and application expansion of MCP and A2A protocols, further optimize the collaboration capability of the full scene agent, strengthen cooperation and sharing with upstream and downstream enterprises in the industrial chain, commit to creating a more open, collaborative and intelligent industrial ecosystem, contribute more to promoting the upgrading of industrial intelligence, lead the industrial full link intelligence to a higher development stage, and jointly meet the brilliant future of the intelligent industrial era.