Get ready to revolutionize your mobile experience with Honor’s groundbreaking AI-Powered Screen Companion, a GUI-based mobile AI agent that reads and understands your screen.
Honor has unveiled its new UI Agent, a ‘GUI-based mobile AI agent‘ that can read and understand your screen. This innovative technology is powered by Google‘s ‘Gemini 2 model’ and aims to simplify tasks on your behalf by understanding the graphical user interface (GUI). The primary demo showcased the agent booking a table at a restaurant through OpenTable.
The Honor UI Agent is a virtual assistant developed by Huawei, integrated into their smart devices.
It offers features like voice control, personalized recommendations, and device management.
The agent uses natural language processing to understand user commands and respond accordingly.
With the ability to learn user habits, it provides tailored suggestions for apps, music, and more.
Honor UI Agent is designed to enhance user experience, making it easier to interact with devices and access various services.
The Honor UI Agent uses multimodal screen context recognition to perform tasks autonomously, without requiring users to train the assistant manually. This approach is more cost-effective than relying on external APIs, as seen in competitors like Apple, Samsung, and Google. Instead of following strict steps, the agent can understand the semantic elements of the user interface and execute a multi-step process to complete your request.
How it Works
In the demo, an Honor spokesperson asked the UI Agent to book a table for four people, specifying local food. The agent successfully chose a restaurant but encountered an issue when it required a credit card to confirm the reservation. At this point, the user had to take over. However, Honor assures that future versions of the UI Agent won’t require users to see the process unfold.
The agent can also be flexible in its queries, allowing users to ask for highly rated restaurants or other specific preferences. While it doesn’t cross-reference data from multiple sources, it does use context recognition to understand user preferences over time.
Collaboration with Google and Qualcomm
Honor’s UI Agent leverages Google‘s ‘Gemini 2 large language model’, which powers intent recognition and enhanced semantic understanding of the screen. The company has also partnered with Qualcomm to keep data on-device and develop a personal knowledge base that learns user preferences. This collaboration aims to provide a seamless AI-powered experience.
Other Announcements
At its keynote at Mobile World Congress 2025, Honor announced several new products, including the Honor Earbuds Open, Honor Watch 5 Ultra smartwatch, Honor Pad V9 tablet, and Honor MagicBook Pro 14 laptop. The company also promised seven years of software updates for its flagship Magic 7 Pro and upcoming devices, matching the update policies of Google and Samsung.
The Future of Artificial Intelligence
Honor’s UI Agent is just one example of the growing trend towards agentic artificial intelligence. Other companies like Rabbit are working on similar technologies, such as ‘Teach Mode’, which allows users to train their assistants manually. The Honor UI Agent‘s self-reliant AI execution model and multimodal screen context recognition capabilities make it a significant player in this space.
Agentic artificial intelligence refers to AI systems that exhibit human-like 'agency, autonomy, and decision-making capabilities.'
These systems can perceive their environment, set goals, and take actions to achieve them.
Agentic AI is often associated with advanced machine learning techniques, such as reinforcement learning and cognitive architectures.
According to a report by Gartner, 40% of enterprises will have deployed agentic AI by 2023.
This technology has the potential to revolutionize industries like healthcare, finance, and transportation.
As technology continues to evolve, we can expect to see more innovative applications of artificial intelligence in our daily lives. Honor’s new AI agent is just the beginning of what promises to be an exciting future for agentic AI.
