AI for Everyone: How a Unified Agent Is Bringing Intelligent Services to Millions

AI has progressed rapidly in recent years, reshaping expectations about how people interact with technology. However, the benefits aren’t experienced evenly. Many people still struggle to access or use AI services effectively because digital tasks are too complex, interfaces are too inconsistent, or support isn’t designed for their needs. In practice, AI capabilities are advancing faster than its ability to serve the full breadth of users.

In China, for example, 1.34 billion phones and more than 780 million TVs don’t have built-in AI features according to China Unicom. Elderly users, people with disabilities and households in rural areas are the most exposed to this gap, often finding digital services too fragmented or frustrating.

China Unicom and Huawei propose a solution. By delivering AI as a cloud-based service across devices, it aims to make advanced capabilities more consistent, more intuitive and more accessible.

Where Today’s AI Falls Short for Users

Modern AI experiences tend to depend on your device. High-end hardware supports integrated assistants and streamlined interfaces, but lower-cost devices often run older operating systems with limited support for advanced features. This creates a fragmented experience where tasks must be completed over several apps or screens, each with its own layout, login process or navigation requirements. For less-confident users, these steps can discourage engagement.

As AI becomes more central to digital interaction, disparities in access will only grow unless a more unified approach benefits a broader group of users.

What Changes When Intelligence Sits in the Cloud

If the underlying ecosystem remains fragmented, improving individual devices won’t solve the problem. China Unicom and Huawei claim that their AI Service Function (AISF) addresses this by shifting intelligence to the cloud and delivering it as a single service on phones, TVs and home equipment.

Unlike on-device approaches, this centralized model avoids uneven capability tied to hardware updates and helps maintain a consistent experience. It relies on stable connectivity to provide a scalable way to deliver a unified service.

This removes dependency on device performance. People with older or lower-cost hardware can access the same AI capabilities because the processing takes place centrally rather than on the device. This broadens the reach of advanced capabilities to a much wider base of users, where connectivity allows.

A unified agent also reduces the fragmentation that characterizes many digital tasks. Preferences, routines and past interactions travel with the user, reducing the effort required to move between screens.

The practical impact is evident in real-world examples, as highlighted below:

Scenario	AISF solution
87-year-old requesting a travel route	Converts a simple natural-language request into a complete task flow, identifying the destination, selecting the route and executing all navigation steps automatically.
Elderly user asking for information or media	Detects user intentions from short prompts, selects the appropriate app, retrieves the requested content and runs the necessary actions autonomously.
Visually impaired user asking, “What’s in front of me?”	Provides real-time environmental intelligence by describing objects, colours and obstacles and offering safe directional guidance.
Families operating shared devices	Ensures continuity across screens by applying shared routines, preferences and past interactions, reducing the need for device-specific actions.

Consolidating tasks into a single service reduces the effort needed. Users who struggle with app navigation or multistep workflows can initiate tasks through simple requests, with the AI agent handling the underlying processes. Consistency across devices supports households with varied levels of digital confidence, reducing the need to manage separate interfaces or settings.

By acting as a single point of interaction in the home, solutions like AISF adapt the technology to the user rather than expecting users to adapt to different interfaces. Shifting from the burden of device-specific interactions helps create a more consistent and inclusive way to manage digital tasks.

Supporting Cross-Device AI

The AISF approach is focused on creating a consistent user experience, powered by:

• Unified memory structure: Maintains a single user profile on multiple devices, avoiding repetitive set-up and ensuring continuity.
• Cloud-based processing: Enables fast responses by interpreting intent centrally, reducing delays that can make voice interaction and complex interfaces difficult.
• Better understanding of intentions: Interprets layered or imprecise prompts, lowering the effort required during multistep tasks.
• Autonomous execution of tasks: Completes actions in applications, even those without formal integration, reducing manual navigation and risk.

These features don’t remove all complexity, but they support a more intuitive, “muscle memory” style of use that encourages people to broaden how they engage with AI services.

Security Is a Core Requirement, Not an Add-On

Security and data protection underpin the design of AISF. The model is built on clear identity assurance, tightly controlled access to personal information and robust safeguards on transmission, processing and storage.

A SIM-based digital identity provides the foundation for verification. By using operator-level real-person authentication, the AI agent’s identity is bound to the verified user, reducing risks of impersonation and preventing duplicate or fake identities. Credentials are stored in the SIM card’s encrypted chip and transmitted through operator-level encrypted channels. Information shared between devices and the cloud is protected throughout, with only an encrypted identifier exposed externally.

Control over personal memory follows the same principle. Users can review, amend or delete stored information, and the system can only access specific memories with explicit permission. Memory data is encrypted from end to end and held in isolated storage, preventing service providers from accessing raw content and strengthening the boundary between user information and system operation.

Additional safeguards govern how generative outputs are produced and how contextual information is applied, supporting predictable and appropriate behaviour for different interactions. This combination of identity binding, encrypted transmission, data isolation and permission controls provides a security posture while enabling consistent operation across devices.

Shared devices like TVs raise unique concerns about user authentication. Simple approaches, such as profile selection or household controls, ensure the agent operates under the correct identity. Clarity will be increasingly important as these services expand.

The Large-Scale Impact So Far

The initial roll-out of AISF shows the potential of a unified, cloud-based model. Since its commercial launch in early 2025, the service has reached 13 million users of phones that rely on cloud-based AI capabilities and about 2 million home users, indicating strong demand for a model capable of working for a wide range of devices.

Performance metrics show notable improvements. According to China Unicom and Huawei, the service’s understanding of user requests has reached 98% accuracy, response times have shifted from seconds to milliseconds and multistep task execution now achieves a 90% success rate. These gains translate into smoother and more reliable interactions.

User sentiment reflects the same trend. The net promoter score for cloud AI phone users has increased from 70 to 77, user approval ratings on China-based app platforms have risen by 8% and overall satisfaction has reached 89%.

These outcomes show how removing device dependence can broaden access. Crucially, longer-term performance will depend on how well the model expands. As adoption grows, multiuser environments may also raise questions about distinguishing between users during shared sessions.

A Wider Vision for Inclusive AI

Solutions like AISF reflect a wider shift toward more-consistent digital ecosystems, where intelligence is accessed through a service layer rather than embedded separately in each device. The model will spread to other devices, including smartwatches, tablets, PCs and in-vehicle systems, allowing users to interact with the same assistant in a broader range of contexts.

Centralizing capability in the cloud means that updates can be applied uniformly, reducing fragmentation and limiting dependency on device-specific features. For users, this creates implications for digital inclusion: if advanced AI services can be accessed through existing devices, barriers associated with hardware cost or varying levels of digital literacy become less restrictive.

In this context, AISF points to an alternative model for accessing AI, one defined by consistency and reach rather than device specifications. It offers a practical route to making advanced capabilities available to diverse segments of the population.

Why Delivery Will Shape the Future of AI

The way AI is delivered matters as much as the models behind it. Moving capability into the cloud and providing a consistent service to various devices changes who can benefit and at what scale. As organizations consider how to make advanced services more widely available, the question becomes one of distribution and stable connectivity rather than raw capability.

A cloud-based, cross-device model offers a clear route to making AI accessible through the technology people already use. It provides a practical example of how intelligent services can support a broader population, not just those with the latest technology. The broader question is how far this approach can be generalized, and whether other providers will adopt similar models as AI becomes table stakes for devices and services.