Connecting the world and beyond

AI Agents

3.0.jpg This category focuses on AI Agent systems equipped with autonomous environment perception, memory management, task planning, and tool execution capabilities. Its standards regulate the core capability evaluation of agents, platform frameworks, interoperability protocols, and specific applications in vertical scenarios.

3.1 Agent Framework & Interoperability

Defines the general architectural capability metrics of AI agents based on pre-trained foundation models, and proposes communication collaboration and standardized interface interoperability specifications between different agent systems.

ITU-T F.748.46: Requirements and evaluation methods of artificial intelligence agents based on large scale pre-trained models (2025)
Specifies the functional requirements and testing/rating methods for foundation model-based agents in dimensions such as visual perception, multimodal content generation, memory, and planning.

F.748.46 Framework of requirement and evaluation methods for artificial intelligence agents based on large scale pre-trained models 

ITU-T F.AAI: Technical Requirements for AI Agent Interoperability (2025)
Standardizes agent interoperability and API integration standards, enabling task discovery and collaborative communication between different agents through unified task and message objects.

ITU-T F.RF-AIAC-FM: Requirements and Framework for AI Agent Collaboration for Foundation Models (2025)
Relying on foundation model platforms, provides the control system and technical architecture for Multi-Agent Collaboration (AIAC) in processing complex tasks.

ITU-T F.AIAP: Framework and requirements for AI agents platform (2025)
Standardizes the unified framework of AI agent platforms, covering multimodal model integration, tool component management, and user interaction interfaces.

3.2 Scenario-Specific Agents

Details the dedicated agent service architectures and scenario-specific functional evaluation requirements for vertical application scenarios such as mobile smart terminals, customer service, and software development programming.

ITU-T F.SPAIAS: Requirements and evaluation methods of Smart phone artificial intelligence agents (2025)
Specifies the service scope and evaluation methods for perception, text/image/video creation of terminal agents deployed on smartphones with LLMs as the core decision-making center.

ITU-T F.RF-CSAIA: Requirements and framework for Customer Service AI Agents (2025)
Standardizes the customer service agent system architecture and external application interaction API requirements applied in full-process customer service scenarios (e.g., pre-sales consultation, after-sales processing).

ITU-T F.748.70: Technical requirements of intelligent development tools for multimedia applications (2025)
Regulates the application requirements of utilizing AI agents to assist the multimedia software development lifecycle, realizing intelligent code generation/completion, development documentation generation, and test design.

F.748.70 Framework of intelligent development tools for multimedia applications