IDC released a general agent evaluation report: NanoWork received three perfect scores and a Top rating for security and controllability
NanoWork performed well in IDC's "Technical Assessment of China's Enterprise-Level General Agent Products" and became the only top-level product that is safe and controllable.
Quick Look
- IDC released the "Technical Assessment of China's Enterprise-Level General Agent Products" report.
- NanoWork achieved three perfect scores in nine dimensions, and became the only product among 12 participating products to receive a Top rating in the "security and controllability" dimension, highlighting its enterprise-level security and implementation capabilities.
AI-generated summary
Why It Matters
IDC released the "Technical Assessment of China's Enterprise-Level General Agent Products" report, which conducted horizontal actual tests on 12 mainstream products.
Recently, IDC, an international authoritative research institution, officially released the "Technical Assessment of Chinese Enterprise-Level General Agent Products" report, which conducted horizontal actual measurements of nearly a hundred real-scenario tasks on 12 mainstream general-purpose Agent products. The evaluation results show that NanoWork achieved full scores in three of the nine evaluation dimensions, among which the "safety and controllability" dimension became the only product among participating manufacturers to receive a Top rating.
This IDC evaluation uses a unified model, a unified test environment and a unified tool interface. 12 products completed nearly a hundred real scene tasks under the same conditions, eliminating model differences and environmental interference to the greatest extent, and the conclusions are horizontally comparable.
The evaluation results show that NanoWork received full scores in the three dimensions of multi-agent orchestration, security and controllability, ecology and openness. What is particularly noteworthy is that in the "safety and controllability" dimension, Nano Work is the only product among the 12 participating companies that reaches the Top level.
NanoWork product technology evaluation results (IDC "China Enterprise-level General Agent Product Technology Evaluation", 2026)
IDC pointed out in the report that for enterprise-level scenario-based intelligent agents, in addition to task completion capabilities, enterprise-level capabilities such as security mechanisms, authority control, operation auditing, execution stability, and enterprise system connection capabilities are the prerequisites for general-purpose agents to truly enter enterprise-level scenarios. At present, most products can already identify prompt word injection and automatically release or block based on risk levels. However, enterprise-level products still need to establish an independent identity and permission system for the Agent, configure temporary credentials and separately set operation scope.
IDC survey data also confirms this trend: 53.6% of enterprises list "completing end-to-end tasks across systems" as the primary driving factor for evaluating general agents, and 52.0% of enterprises believe that ROI quantification and cost management are the key to the implementation of Agent scale. This means that the Agent must go deep into the core system and data of the enterprise, and every operation step must be observable and attributable. Therefore, independent identity permissions, operational audits and security boundaries have become a necessity. When agents begin to read and write corporate documents, operate business systems, and call third-party tools for employees, security and controllability are no longer a bonus, but the basic threshold for enterprise selection.
The relevant person in charge of Nano Work said that as an intelligent office platform under 360, Nano Work has regarded safety and reliability as the core gene of the product since its inception. "Many products add security as a functional module, but we have written security into every layer of the product from the architectural level." The person in charge said, "IDC's evaluation results verify our direction - in open multi-agent collaboration and ecological connections, we can still maintain the highest level of security and controllability. Security does not mean closed. Real enterprise-level capabilities allow enterprises to enjoy the efficiency of agents while incorporating every step of the operation into a manageable, controllable, and traceable governance system."
As general agents move from "early trials" to "large-scale applications," enterprises' requirements for security and reliability will continue to increase. NanoWork's performance in this IDC evaluation provides an authoritative basis for enterprise selection and also sets a benchmark for enterprise-level intelligent security construction for the industry.
Open Questions
- Who are the other 12 specific products participating in the review?
- What detailed indicators are included in the nine specific evaluation dimensions?



