From Smart to Finish: The Agentic Smartphone Era Arrives

An intelligent digital assistant with wide-ranging capabilities and a human-like interface? Long a favorite vision of science-fiction films, such assistants are now starting to appear in the real world — albeit within the familiar form factor of smartphones, rather than a tuxedo-clad humanoid robot butler. Welcome to the dawn of the Age of Agentic AI.

AI agents are systems capable of autonomously planning and executing multi-step tasks across apps – including interacting with third-party services on behalf of users. With agentic AI expected to become their primary interface, smartphones are set for a major evolutionary step.

The new Find X10 Series from OPPO, the world's fourth-largest smartphone maker, typifies the new era. Through a collaboration with the AI startup Manus, the phone's ColorOS 17 operating system embeds agentic AI, most prominently in a feature called AI Mind Pilot.

"We are at an important transition point — from an app-centric smartphone toward a more intent-driven experience," says Xi Chen, Product Design Director of OPPO ColorOS.

"For years, users have had to navigate between different apps and services themselves to accomplish a task. With increasingly capable AI, the interaction can start to shift from 'Which app should I open?' to 'What do I want to accomplish?'"

Agentic AI can understand an end goal, break it into steps, identify the relevant information and tools, and deliver a result. And the smartphone is the perfect place for this, because it's where people keep so much of their daily lives — photos, documents, communications — in a readily accessible place.

Combine this with modern computing power and connectivity, and it becomes the natural platform for AI to understand context and help with real-world tasks, setting the stage for a more personal and proactive mobile experience.

"Apps remain the underlying infrastructure, while AI becomes the intelligent layer that connects information, tools and services around the user’s intent."

OPPO's pursuit of what it calls Personalized Proactive Intelligence has been years in the making. The smartphone maker has been exploring and implementing self-trained language models since 2020, and introduced a range of AI-powered features – AI Editor, Summary, Translate and more — in its recent model launches. 

Now, with ColorOS 17 in the Find X10 Series, AI is core to the experience. “We are moving from individual AI features toward deeper AIOS integration, agentic capabilities and more personalized experiences,” says Xi.

This view — that the next stage of smartphone AI should move beyond answering questions or generating content toward understanding users’ goals and helping them complete tasks — is one shared by AI startup Manus, which bills its agentic AI as an “action engine”.

With capabilities in creative image generation and editing, research and learning, and document and presentation creation, Manus AI helps turn ideas and information into tangible results. This made it a natural fit for OPPO’s vision of a smartphone that can understand user intent and help turn it into action. 

The two companies brought together Manus’s agentic capabilities and OPPO’s device-level context and OS integration, to create Agent Mode within ColorOS 17’s AI Mind Pilot. 

"The collaboration is about combining agentic intelligence with the smartphone’s context and capabilities, making it possible for AI to work more directly around the user’s intent,” says Tao Zhang, Co-Founder and Chief Product Officer of Manus AI.

“Manus can, with user authorization, directly find, organize and work with existing information within the OPPO smartphone itself, including Photos and Files, rather than requiring users to manually gather and upload it again."

Agent Mode enables the AI to work across information and tools to complete multi-step tasks. What makes the experience seamless is device integration: OPPO’s proprietary on-device AI models power the file and photo recognition and organization, while the Manus agent searches for and understands relevant online information and is able to combine it with the authorized on-device content.

So, if a user wants to create a travel itinerary, they can simply state the goal and let the agent handle the steps, incorporating relevant information from online sources with photos and files already stored on the phone.

"This allows users to move from simply accessing information to getting work done directly on their phone." 

Turning a powerful agentic capability into an experience that feels natural, reliable and useful on a smartphone was no easy task. The different systems had to communicate reliably and ensure that information could be retrieved and organized accurately.

Multiple layers of safeguards help ensure the user is firmly in control of agentic AI. 

Users must explicitly authorize access to personal information, and access is limited to fulfilling a stated goal. When working with files and photos, processing takes place on-device, meaning that no photos or files leave the phone. 

When AI Mind Pilot needs to work with online sources, user confirmation is required if any personal information needs to be shared with an external service. Indeed, if there is any ambiguity in a use case, or meaningful consequence, the AI system will seek clarification and confirmation before proceeding.

"Our principle is simple,” says Xi. “Personalized Proactive Intelligence means that agentic AI will understand the goal behind the request and help users move toward the outcome, not remove the user from the decision-making process."

The new era of smartphones is thus less a collection of separate apps and menus, and more of a conversational intelligent personal interface — AI joining the dots between the device, apps and context, to help users accomplish what they want.

Robot butlers might still be a while away. But the rest of your digital life might soon surprise even the sci-fi writers.