en.Wedoany.com Reported - By 2026, the evolution of AI Agents has crossed the threshold of "being able to chat" and entered a new phase of "being able to work." This forces a fundamental shift in the core proposition of enterprise software: software is no longer just a tool for recording business operations but has evolved into a "digital employee" capable of participating in business execution and even making autonomous decisions. Liu Zhongwen, Vice President of Kingdee China and General Manager of the R&D Center, pointed out in an interview with a reporter from *China Information Weekly* that the industry is bound to undergo a complete restructuring, and the opportunity lies in who can first rethink the future form of enterprise software from the ground up.
AI Agents Force a Paradigm Shift at the Core of Enterprise Software
Addressing the current state of the enterprise software industry, Liu Zhongwen stated: "Over the past few decades, enterprise software has primarily focused on solving the problems of business digitization and process digitization, essentially helping businesses move their operations into systems. In the Agent era, software begins to possess capabilities for understanding, decision-making, and execution. It is no longer just a tool but becomes a participant in the enterprise's operational system."
This change forces the entire industry to rethink the essence of software. In the past, the core tasks of enterprise software were "recording" and "workflow," moving offline operations online to make processes run more smoothly. However, the emergence of AI Agents gives software, for the first time, the ability to complete tasks autonomously. It can understand a business goal, call upon systems, analyze data, and execute operations on its own.
Liu Zhongwen admitted that Kingdee feels both "a sense of crisis and excitement" about this. The crisis comes from the inevitable restructuring the industry will undergo, while the excitement stems from the opportunities Kingdee sees. The enterprise management software field possesses a wealth of real business scenarios, process rules, organizational knowledge, and operational data—precisely the foundation that Agents need to create value. "Therefore, Kingdee has chosen not to simply overlay AI features on existing products but to fundamentally rethink the future form of enterprise software from the ground up."
Kingdee's cloud transformation over the past decade addressed changes in software delivery methods, shifting from on-premise deployment to cloud subscriptions. However, Liu Zhongwen clearly stated that the depth of this transformation far exceeds that of the cloud shift. The cloud transformation changed delivery and deployment methods, whereas the AI-native transformation changes enterprise operations themselves and the human-machine relationship. Based on this judgment, Kingdee has been steadily advancing its AI-native transformation over the past two years: from embedding AI capabilities, to building an enterprise management AI system, and finally to the official launch of the enterprise AI operating system "Lingji" in May 2026. This path is not about stacking features but about rebuilding the underlying architecture for the era of intelligent agents.
Why Is Large-Scale Enterprise AI Implementation Still a Challenge?
In Liu Zhongwen's view, enterprise AI implementation faces three specific challenges. First, the complexity of enterprise scenarios. Enterprises are not personal tool scenarios. They involve organizational structures, permission boundaries, business rules, financial constraints, audit requirements, and cross-departmental collaboration. For AI to truly complete tasks, it must understand business semantics and access processes and data. Second, results must be measurable. Customers won't pay just because "AI is smart"; they need to see concrete improvements in areas like review efficiency, reimbursement efficiency, and decision-making efficiency. Third, improving individual efficiency does not equal improving organizational efficiency. What enterprises truly need is not more AI tools but a mechanism that allows AI to operate in synergy with the organization, enabling the continuous accumulation and reuse of knowledge, experience, and capabilities.
The key to solving these problems lies not in launching AI assistants or toolkits with richer features, but in building a set of foundational capabilities that allow AI to truly enter the business closed loop. This includes understanding business semantics, complying with organizational permissions and process rules, and ensuring the traceability and auditability of execution results. This is fundamentally different from the logic of "building systems first, then running processes" in enterprise digitization over the past few years. The era of intelligent agents demands a new paradigm: "AI understands business, AI is embedded in processes, and AI operates within a governance framework."
Liu Zhongwen introduced that Kingdee launched the enterprise AI operating system "Lingji" in May this year. The core idea of its technical architecture is precisely centered around the industry challenges mentioned above, attempting to solve issues of semantic understanding, permission adaptation, process integration, and governance compliance between AI and enterprise business through a unified operating system layer. However, he also emphasized that this is not something Kingdee can accomplish alone. "Lingji" incorporates the Skill Marketplace, Agent Marketplace, and open protocols as part of its architecture, essentially attempting to build a multi-party capability ecosystem. "What enterprises will need in the future is not more point AI tools, but a set of underlying operating systems that allow AI to understand business, execute business, and continuously grow," Liu Zhongwen said. "This requires the entire industry to explore together; it's impossible for one company to achieve it in isolation."
From Selling Tools to Co-building Enterprise Intelligence: An Industrial Revolution Far Beyond Cloud Transformation
The launch of Lingji is not just the release of a single product by Kingdee; it reflects Kingdee's judgment on the direction of the entire industry. Liu Zhongwen believes that enterprise software in the era of intelligent agents will undergo changes at three levels.
First, the entry point for enterprise software will change. In the past, users entered software by clicking menus, filling out forms, and initiating processes; in the future, they will more often express their intentions through natural language, with AI calling upon systems, data, tools, and agents to complete tasks.
Second, the value of enterprise software will change. In the past, software mainly improved process efficiency; in the future, AI will further enhance decision-making quality, organizational collaboration efficiency, and business outcomes. Software will no longer be just a tool but a part of enterprise intelligence.
Third, Kingdee's own role will change. During the cloud transformation period, Kingdee helped enterprises move from on-premise deployment to cloud subscriptions; in the era of intelligent agents, Kingdee aims to help enterprises transition from digital systems to AI-native organizations. Liu Zhongwen summarized it in one sentence: "In the past, we delivered systems and processes; in the future, we will deliver operable, governable, and evolvable enterprise intelligence."
These changes also impose new requirements on the business models of enterprise software. As AI gradually participates in enterprise operations, customers are paying more attention to the operational value, efficiency improvements, and business outcomes brought by AI, rather than just the software features themselves. This means that enterprise software vendors need to continuously explore new business models that better match the AI era, centered around customer value. When "usage volume" is no longer the sole measure of value, the business models of enterprise software will continue to evolve. Kingdee is exploring a transition from "selling tools" to "delivering results." Liu Zhongwen stated that the core challenge is not just technological breakthroughs but also making AI truly embedded in the business closed loop and taking direct responsibility for efficiency, quality, and governance outcomes.
The transformation of business models also calls for an open ecosystem. In Lingji's six-layer architecture, the "Marketplace and Ecosystem Layer," through the Skill Marketplace, Agent Marketplace, and open protocols, enables enterprises, individuals, and ecosystem partners to participate in value creation and capability evolution. Liu Zhongwen concluded: "What enterprises will need in the future is not more point AI tools, but a set of underlying operating systems that allow AI to understand business, execute business, and continuously grow."
At the end of the interview, Liu Zhongwen stated: "Enterprise software in the era of intelligent agents will evolve from 'process digitization' to 'organizational intelligence.' Kingdee's role is gradually evolving from a 'tool provider' to a 'co-builder of enterprise intelligence.'"
