OPPO's ColorOS 17 Moves AI Toward Proactive, On-Device Agents

OPPO said the central shift is from AI waiting for a user command to AI understanding the user's context and deciding when to offer help. The company pointed to data already on phones, including calendars, notifications, location, travel plans, app habits and health data from wearables, as the basis for a fuller context once users grant permission.

Xiaobu Space acts as a personal affairs management entry point. It recognizes calendar arrangements, invitations in chats and upcoming movies or trips, then consolidates information scattered across apps into to-do items. Important items can be pinned at the top of Xiaobu Space.

The AI one-tap memo feature expands beyond pickup-code recognition from the previous generation. ColorOS 17 adds delivery pickup codes, contact information and concert ticket-grabbing times. The system does not only save screen content; it tries to understand the next action: a pickup code can trigger a reminder on arrival, a phone number can create a contact, and ticket-grabbing information can become an alarm accurate to the second.

OPPO placed proactive reminders in a new Xiaobu Suggestions feature. According to the company, it works with more than 40 third-party services, covering over 150 service categories and more than 700 sub-scenarios. In ordering, the system can offer a card based on past tastes and habits. For long-distance travel, it can extend prompts from pre-departure through arrival, transfers, security checks and boarding.

For location awareness, OPPO introduced a self-developed swarm sensing technology. OPPO said conventional geofencing only determines whether a user has entered an area, while ColorOS 17 seeks to distinguish micro-locations such as outside a station, passages, gates and waiting areas. OPPO claimed 95% prediction accuracy and said the feature covers metro lines in more than 30 Chinese cities. A companion fusion navigation fluid cloud connects walking, bus and metro travel into one continuous chain, reminding users of transfers and exit guidance. After getting off, Car+ can continue unfinished navigation from the car to the phone.

These features show OPPO's interpretation of proactive intelligence: AI does not always need to exist as a chat box. It can be a card that appears at the right time, an arrival reminder or navigation information flowing across devices. OPPO said the number of scenarios shown at the launch event is only a first step; usefulness will depend on recognition accuracy, third-party service coverage and whether the system remains restrained enough to avoid notification overload.

For the next stage of AgentOS, OPPO previewed Xiaobu Next. In the preview version, a complex task can be divided among multiple AI 'experts' with different abilities. Users can select agents as if forming a project team and assign tasks in a group-chat-like interface. Task progress and interim results update continuously, while key deliverables are pinned at the top of the page.

OPPO described examples including a fat-loss agent that creates a plan based on exercise and health data, accepts daily check-ins and generates a weekly review. A content creation project can have different agents handle topic selection, writing and visual design. The system can also connect to PCs, car systems, smart home devices and third-party services, turning requests in a conversation into cross-device operations. Unlike traditional voice assistants that complete one operation per command, Xiaobu Next is closer to a long-running task space that retains project context, calls personal data and reports progress.

OPPO said on-device agents can directly use files, calendars, locations, health data and device status, making cross-app and cross-device tasks easier than with cloud-only general agents. The approach also raises privacy and security concerns. OPPO listed 'on-device priority and encryption' as a safety evaluation dimension for Xiaobu Next.

ColorOS 17 also updates more conventional generative AI features. In imaging, it adds an AI retoucher based on OPPO's self-developed Da Vinci visual understanding large model. Users can make compound natural-language requests involving erasure, composition, brightening, super resolution and color adjustment; the system then breaks the request into expert models for light and shadow, image quality, composition, color, portraits and elements. It also supports uploading a reference image to replicate a visual style.

AI Notes is expanded into a personal knowledge management space. Besides text and recordings, users can import scans, video links, files and Xiaobu Memory. A note assistant can find old notes, adjust content and generate editable information summary images. As content accumulates, the system tries to extract a 'moment of inspiration' from the personal knowledge base.

AI is also placed in accessibility and security scenarios. An accessibility intelligent screen reader can identify third-party app controls that lack screen-reader labels and describe them by voice. A full-chain AI anti-fraud system combines call semantic recognition, cross-app risk judgment and AI face-swap detection to provide system-level protection for older users and other high-risk groups. These features cover imaging, office work, travel, health and accessibility. They do not all rely on the same AI technology, but they share a common goal of embedding AI into specific system flows and reducing the cost of switching between standalone apps.

AI is the most visible thread in ColorOS 17, but OPPO also updated the fluency and visual foundation. The system introduces a new 'fluid design' made up of light effects, fluid motion and flexible feedback. Light begins to carry part of the interaction cue: when music plays, a directional light follows the progress bar; when users tap or swipe, light diffuses or ripples with the finger. Fluid motion extends from the previous generation's lock screen to the whole system. A new fluid clock connects lock screen, unlock and home screen states, using the clock as a visual anchor so icons and backgrounds move around one focus. Live weather dynamically presents sunny, rainy, snowy, dawn and night states based on the user's location and time.

Flexible feedback handles the relationship between finger and interface. Buttons contract slightly when tapped, stretch with dragging and spring back when released. Notification center, control center and home screen transitions move progressively according to operation speed. OPPO said the new Aurora engine uses a fusion rendering architecture that places lock screen, home screen, wallpaper and status bar layers into a unified rendering system, reducing repeated drawing and keeping the same elements continuous across interfaces. According to OPPO, memory usage in key scenarios drops by up to 25% and rendering load by up to 30%.

The new Tide engine handles lower-level resource scheduling. The system considers app usage frequency, duration and personal preferences, applying different strategies to foreground apps, important background apps and non-important background apps. OPPO said the complete app keep-alive rate improves by 55.6% over the previous generation and set a 'six-year smooth experience' as the long-term goal for this generation. The visual and performance changes also support AI, the company said, because lighter interface forms such as fluid clouds, suggestion cards and dynamic reminders are better suited to proactive intelligence than constant pop-up dialogs and notifications.

ColorOS 17 will first be available on the OPPO Find X10 series and OnePlus 16, and will gradually roll out to more devices, according to iFanr.

Editor's Summary

OPPO's ColorOS 17 positions AI as a proactive system layer, combining context-aware suggestions, on-device multi-agent workflows and new rendering and scheduling engines. The release shows OPPO pushing AI deeper into the operating system while tying first availability to its flagship Find X10 series and OnePlus 16.