HGDC 2026：智能体装进OS 荣耀重构AI时代人机交互新范式

Key highlights from Honor’s 2026 Global Developer Conference include the fully upgraded YOYO NEXT agent and the system-level AgenticOS, designed to greatly streamline smartphone operations and boost daily use efficiency.

1. Core experience upgrades reshape traditional interaction logic: Unlike the previous YOYO, which required step-by-step voice commands, the new YOYO NEXT leverages system-level capabilities to complete complex, cross-app task chains with a single prompt, eliminating the need for manual navigation across apps. It can also proactively detect scenarios to deliver relevant services without requiring users to open apps and search for functions.

2. Practical features cover a wide range of daily scenarios: Upgraded AI creation tools offer improved object removal, ID photo generation, and virtual outfit try-on effects; task management features support phone use, work, and travel scenarios; human-like real-time assistance is available during video calls; and proactive reminders cover use cases such as commuting, ride-hailing, and ticket collection. To date, these proactive services cover more than 100 scenarios and have served over 100 million users.

3. Continuously evolving personalized experience: Equipped with long-term memory capabilities, the system learns users’ habits over time. It also supports custom integration of third-party large models and control of third-party apps, with a built-in skill store for ongoing feature expansion.

Honor’s launch of a new system-level AI agent interaction paradigm signals the direction of experience upgrades and new user service reach logic in the consumer electronics sector in the AI era, providing clear references for brands’ product development, user operations, and marketing conversion.

1. Clear shift in consumer trends: AI is evolving from conversational assistants focused on Q&A and chat to proactive agents capable of completing complex real-world tasks. User demand for AI is shifting from novelty to efficiency, and seamless, non-intrusive services across apps and devices represent the core experience direction going forward.

2. Fundamental change in user reach logic: The traditional model of relying on users to actively open apps and search for services is becoming less efficient. Proactive service recommendations based on system-level scenario awareness and user habit memory enable precise reach at the exact moment user needs arise. Currently, Honor’s YOYO Recommendations cover 7 system-level entry points, serve over 100 million users, and support more than 100 proactive service scenarios, delivering both improved user experience and commercial value growth.

3. New brand partnership pathways are now open: Honor offers multiple integration paths, including MCP, A2A, and Skill, enabling brands to reach target users precisely through system-level intelligent distribution, effectively improving conversion efficiency and reducing technical costs for service integration.

The system-level AI agent ecosystem launched by Honor points to new traffic growth pathways and low-barrier collaboration opportunities for service sellers in the AI era, and warrants early attention and strategic layout.

1. Changing consumer demand creates new opportunities: Most smartphone users find the manual, cross-app, step-by-step process of completing tasks cumbersome, creating strong demand for intelligent tools that can proactively identify needs and deliver end-to-end connected services. The commercial conversion potential of system-level proactive service scenarios is expanding rapidly.

2. Platform support policies lower entry barriers: Honor provides partners with multiple integration options, including MCP, A2A, and Skill, allowing sellers to choose based on their own capabilities and minimize entry costs. The platform also offers full-link support from integration to distribution, leveraging system-level intelligent distribution to accurately match sellers with target users, significantly improve conversion efficiency, and turn service capabilities into tangible commercial returns.

3. Shifting traffic logic signals new directions: The traditional traffic model relying on active user search and store visits is changing. Services embedded in the system-level intelligent service chain, which proactively appear when user needs arise, will gain greater exposure. Sellers can move early to integrate and capture initial traffic dividends.

Honor’s announced direction for system-level AI agent technology for AI phones provides clear action guidance for consumer electronics and smart hardware factories in product design, digital transformation, and opportunity capture.

1. New requirements for product design are clear: The core competitiveness of future smart hardware will no longer depend solely on standalone hardware specifications. Instead, hardware must adapt to system-level, cross-device, and cross-app AI agent scheduling, supporting technical features such as real-time sensing, end-cloud collaboration, millisecond-level response, and local data security protection to meet the intelligent experience standards for interconnected devices including phones, PCs, and tablets.

2. Ecosystem collaboration creates new business opportunities: Honor has opened third-party ecosystem and Robot Phone hardware collaboration channels. Factories with hardware R&D and manufacturing capabilities can develop compatible peripheral hardware around agent interaction scenarios, integrate into the Honor ecosystem to reach massive users directly, and unlock new revenue growth.

3. Relevant insights can inform digital transformation: Factories can draw on the logic of system-level intelligent services to build cross-process automated execution links in production and operations, reducing communication and execution costs from manual step-by-step operations and improving overall operational efficiency. The model of processing data locally on the end side also offers reference for building factory data security protection systems.

Honor’s demonstrated implementation results for system-level AgenticOS and YOYO NEXT agents clearly outline development trends, existing user pain points, and actionable technical solution directions in the AI service industry, offering strong reference value for service providers’ business planning.

1. Industry development trends are clear: AI applications are accelerating their transition from shallow, in-app feature-add-on assistants to system-level agents capable of proactively completing tasks across apps and devices. The inflection point for large-scale deployment has arrived; AI services that truly solve users’ real complex tasks are the products the future market demands.

2. Core customer pain points are evident: Key friction points for users using AI services include cumbersome cross-app operation steps, the need for users to actively find and trigger services, AI’s lack of understanding of user habits, high development costs for cross-device scheduling functions, and difficulty ensuring end-side data security.

3. Actionable solution directions are defined: Four core capabilities—real-time perception, long-term memory, reasoning and planning, and automated execution—can be built on a system-level framework. An end-cloud collaborative architecture can deliver millisecond-level response and local data protection, while diversified integration methods can aggregate ecosystem partner capabilities, reduce developers’ cross-scenario scheduling costs, and improve service conversion through system-level distribution.

Honor’s AI ecosystem layout and system-level agent implementation path provide a proven reference model for internet platforms and ecosystem operation platforms in functional iteration, merchant recruitment and operations, and service upgrades.

1. Innovative platform operation practices are worth referencing: Platforms should break down service silos between traditional apps and build an intelligent service framework at the underlying level that supports cross-app, cross-device, and cross-modal invocation, upgrading services from “users finding services” to “services finding users” through precise, proactive recommendations based on scenario awareness and user habit memory. Honor’s practice has already covered over 100 million users and launched more than 100 service scenarios, fully validating the model’s feasibility.

2. Ecosystem merchant recruitment support systems can be optimized: Platforms should provide settled partners with diversified, low-barrier integration options, such as flexible access via MCP, A2A, and Skill, and build a full-link support system from integration to intelligent distribution, helping partners improve conversion efficiency, generate tangible commercial returns, and continuously increase platform attractiveness to partners.

3. Key operational risk mitigation points are clear: For end-side AI services, platforms must enforce robust local data protection, balance intelligent experiences with user privacy and security to avoid data leaks, and reserve space for users to customize third-party service access, meeting personalized needs and improving overall experience fluency.

The system-level AI agent implementation progress disclosed at Honor’s Global Developer Conference provides cutting-edge practical samples and new research directions for studies on AI industry development and business model evolution.

1. Clear new trends in industry development: The AI industry is reaching a critical inflection point in 2026. AI technology is accelerating its transition from “articulate” responsive assistants to proactive “action-delivering” agents, shifting from feature add-ons within individual apps to penetration into underlying system capabilities. The industry’s first mobile operating system to achieve commercial deployment of a system-level Agent Harness has been officially released, marking the entry of agents into large-scale application that will deeply empower industries across the board.

2. A new ecosystem co-creation business model has taken shape: System players open underlying AI capabilities and system entries, providing foundational support including end-side and cloud large models, cross-app scheduling, and intelligent distribution. Developers integrate through diversified, low-barrier approaches and reach users precisely via system-level proactive recommendations, achieving dual growth in user experience and commercial value. This model has so far been validated by practice across 100 million users and more than 100 service scenarios.

3. New research topics are emerging: Issues including local data protection mechanisms under end-cloud collaborative architectures, the definition of user privacy boundaries brought by long-term memory functions, and the allocation of service rights and responsibilities for system-level agents hold high value for follow-up research.