AI Agents Need a "Safety Certificate" to Go Live: DingTalk Wukong Passes World's First International AI Management System Certification

Recently, Wukong successfully obtained the ISO/IEC 42001:2023 Artificial Intelligence Management System certification issued by China Quality Certification Center (CQC). After rigorous third-party auditing, Wukong's AI management framework, together with DingTalk’s, has met international standards in terms of security, transparency, and compliance.

DingTalk actually became the first in China to obtain this certification back in 2023, establishing a solid foundation in the field of AI safety and compliance. Now, Wukong’s successful attainment of ISO/IEC 42001 recognition further extends and elevates DingTalk’s commitment to AI safety—demonstrating consistent dedication to security and trustworthiness, from platform to intelligent agents. The enterprise AI capabilities represented by Wukong have now undergone comprehensive validation—from product technology to safety and compliance—under the world’s most authoritative AI governance standard framework.

Why does this matter?

With the EU’s AI Act officially taking effect and China’s Interim Measures for the Administration of Generative Artificial Intelligence Services being progressively implemented, AI compliance has shifted from a “nice-to-have” to a mandatory “entry requirement.” ISO/IEC 42001 is the world’s first international standard specifically designed for artificial intelligence management systems, jointly published by the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC). It establishes global benchmarks for risk management, ethical compliance, data security, and accountability across the entire AI lifecycle.

In this audit, Wukong covered all core areas including AI ethics and fairness, data quality and bias control, human oversight mechanisms, transparency, and explainability. It also achieved breakthroughs in security—effectively defending against AI-specific attacks such as prompt injection and model poisoning, ensuring enterprise data and business operations remain secure.

How did Wukong achieve this?

Unlike the common industry practice of “getting functionality working first and addressing security later,” Wukong embedded security into its foundational architecture from day one. It has built a layered security system covering runtime environments, permission management, data flows, and operational auditing. Every skill deployed on the platform must pass a security scan, and during execution, a Policy Engine performs real-time evaluation, immediately blocking any non-compliant actions instead of patching issues after the fact.

Transparency

Transparency and traceability have always been focal points for both the industry and the public. Wukong built an AI-native file system from scratch, automatically saving a complete snapshot at every step. If any issue arises, it can roll back to any historical version within seconds. All requests and operations are fully logged—clearly showing who initiated them, which device was used, and under what environment. This is not a "black box," but an AI that can clearly explain why each decision was made.

Data Security

For data security, Wukong implements multi-layered protection measures: For highly regulated industries such as finance and government, Wukong supports dedicated cloud deployment, ensuring core data never leaves the enterprise boundary. It also employs container-level sandbox isolation, confining risks within the smallest possible "cage." Additionally, all operations support full-chain audit and traceability, truly achieving the principle of "never touching unauthorized data, and allowing immediate reversal of accidental operations."

Permission Management

In terms of permission governance, Wukong established a dual-layer system combining baseline security rules with enterprise-customizable policies. It introduces the "principle of permission intersection"—meaning the data accessible to AI will never exceed the intersection of the user’s own permissions and the requester’s permissions—thereby preventing unauthorized access to sensitive data at the architectural level.

Real enterprise scenarios validate practical security capabilities

As DingTalk’s enterprise-grade AI agent platform, Wukong is already widely deployed across e-commerce, retail, and manufacturing sectors, supporting key use cases such as content creation, operational analytics, customer follow-up, order management, and AI application development. Suzhou Light Energy uses Wukong to analyze nearly one million charging station order records—building a BI dashboard with a single sentence that previously took half a day. Yiwu Ucclea reduced HR payroll processing time from two days to just ten minutes using Wukong, while increasing new product launch success rates from 60% to 92%.

These real-world cases underscore a crucial truth: when AI becomes deeply integrated into enterprise workflows, security and trustworthiness are the essential prerequisites for any efficiency gains. Wukong is not merely a capable AI assistant, but a trustworthy AI agent enterprises can confidently adopt.

Against the backdrop of increasingly stringent global AI regulation, achieving ISO/IEC 42001 certification ahead of peers not only fulfills commitments to user trust but also sets an industry benchmark for "security-first" development.

Going forward, Wukong will release an AI Security White Paper, comprehensively disclosing its security design philosophy and implementation details across dimensions such as compliance, end-to-end protection, data security, privacy preservation, and ecosystem safety.

In today’s era of rapidly advancing AI capabilities, Wukong has chosen a more difficult yet fundamentally correct path—building a solid foundation of security before enabling AI to go further.

We dedicated to serving clients with professional DingTalk solutions. If you'd like to learn more about DingTalk platform applications, feel free to contact our online customer service or email at This email address is being protected from spambots. You need JavaScript enabled to view it.. With a skilled development and operations team and extensive market experience, we’re ready to deliver expert DingTalk services and solutions tailored to your needs!