Alibaba's AI Can Now Click Your Phone for You
Imagine an AI that can use apps, fill forms, and do tasks on your phone or computer automatically.
Alibaba just released Qwen-UI-Agent, an AI designed to work like a digital assistant that can actually *use* your phone, laptop, or any app—without needing special programming. Think of it like a super-smart intern who can read your screen, click buttons, fill out forms, and even handle multi-step tasks across different apps, just like a human would.
In tests, this AI outperformed leading models like GPT-5.6 and Claude Opus 4.8 when it comes to navigating real-world software, which is a big deal because most AI tools today struggle with apps that don’t have easy-to-use programming interfaces (APIs). Companies have been trying to automate repetitive tasks like data entry or report generation, but many of these tasks require clicking through clunky enterprise software—something AI hasn’t been great at until now.
DeepSeek also joined the party with a new AI model, deepseek-v4-flash-vision-exp, which adds image understanding to its existing text-based AI agents. This means the AI can now “see” screenshots, charts, or PDFs and use that visual information to make decisions—all without extra cost. For teams already using DeepSeek’s AI for automation, this upgrade makes it easier to handle tasks like reading dashboards or approving reports that include images.
Meanwhile, Tricentis is turning the spotlight on AI agents themselves, treating them like software that needs testing. Their new tools, like Aida and AgentScore, help companies evaluate how reliable these AI agents are in real-world conditions before letting them loose on important tasks. It’s like giving your AI a test drive before trusting it to drive your business.
- Alibaba’s Qwen-UI-Agent can operate phones, computers, and websites by understanding on-screen elements and clicking buttons, outperforming top AI models in real-world tests.
- DeepSeek’s new AI model can now “see” images and screenshots, making it easier to automate tasks like reading dashboards or approving reports without extra cost.
- Tricentis is testing AI agents themselves, helping companies measure reliability before trusting them with important tasks.
Why It Matters
Your next boring computer task could be done by an AI—saving time, reducing errors, and freeing you up for more important work.