China’s Manus AI signals the next evolution in autonomous AI

The news: Manus AI, developed by Chinese startup Monica, is an autonomous AI agent designed to think, plan, and execute real-world tasks—surpassing traditional chatbots and delving into autonomous AI, per The Register. It’s currently free and available by invite-only, but interested users can apply for access.

This is the latest example of China’s AI startups releasing next-gen competitive models—intensifying pressure on US counterparts to ramp up features or make existing AI more affordable.

What does Manus AI offer? Unlike OpenAI’s GPT-4 or Google’s Gemini, which provide text-based responses, Manus completes entire workflows.

Demos show it generating stock analysis, planning travel itineraries, and building websites with minimal human input. It’s hinting at operators, the next phase of AI evolution. 

Here are some examples of what Manus AI can do:

• Autonomous task execution: Manus completes multistep workflows independently, from financial analysis to making decisions and solving problems.
• Multi-agent system: It deploys specialized sub-agents, guided by an “executor agent,” to handle complex tasks and manage workflows like legal document drafting, supply chain optimization, and personalized education planning.
• Cloud-based automation: Manus runs tasks in the background, notifying users only when complete. Marketers can automate campaign analysis, ad-performance tracking, and content generation without constant oversight.

The AI operator uses reinforcement learning and innovates through trial and error for continuous improvement, which can refine its decision-making over time. One potential benefit of this method is that it can adapt to complex scenarios faster than traditional AI solutions.

Some caveats: Companies testing complex automated AI systems like Manus need to come to terms with their corporate data and internal system running on remote and foreign AI servers, a red flag for security and regulatory compliance. In fact, 38% of business leaders worldwide say the latter is a barrier to developing and deploying generative AI (genAI), per Deloitte.

Our take: Limited access could impede interest and adoption, but if Manus AI is everything it claims to be, it could eclipse DeepSeek R1 as the next AI breakthrough. 

The frenetic rate of AI innovation makes it difficult and confusing for businesses to evaluate and decide on the right tools for their needs. Trial and error, especially for free models, could help companies determine their ideal AI.