Artificial intelligence (AI) is evolving fast. But the next phase may not just be about smarter models -- it could be about systems that actually take action.
This new category of software, called agentic AI, is designed to take basic instructions about a task, then complete the steps to handle that task, make decisions, and operate with minimal additional human input.
And while much of the attention in this sector is still on U.S. tech giants, China-based Alibaba Group (BABA +1.75%) is quietly building tools that fit directly into this new theme.
Image source: Getty Images.
The next phase of AI is about AI doing things
So far, most artificial intelligence systems have focused on generating outputs -- text, images, or predictions. But agentic AI goes further. Instead of just answering questions, these systems can:
- Complete multistep workflows
- Interact with software tools and other digital agents
- Execute tasks across systems
In simple terms, they are intended to act more like digital employees than chatbots. That shift matters since the real value of AI isn't just what answers it can generate -- it's what it can do on our behalf.
As companies look to automate operations and reduce manual work, demand for these systems could grow quickly. To put the opportunity into perspective, according to a study by the researchers at Markets.us, the market for agentic AI, valued at just $5.2 billion in 2024, is expected to grow to $197 billion by 2034, a massive prize for companies to capture.
NYSE: BABA
Key Data Points
Alibaba is already turning this technology into real products
That's where Alibaba stands out. While many companies are still talking about AI potential, it is already building applied, agent-driven tools.
Two examples highlight this strategy:
1. Wukong: An enterprise AI control center
Alibaba's Wukong platform is designed to coordinate multiple AI agents within a single system for enterprises.
It can:
- Edit documents
- Update spreadsheets
- Transcribe meetings
- Conduct research
All that within one interface. Think of it as a multiagent operating system to help manage business workflows, rather than a single AI tool.
2. Accio Work: AI to run business operations
Alibaba's newer platform, Accio Work, is another example of how agentic AI could work in real life.
The new AI tools work as a plug-and-play "task force" that can autonomously handle complex business operations, especially for small and medium-sized enterprises.
It can:
- Manage supplier sourcing and negotiations
- Handle sales tax and customs processes across markets
- Automate workflows like inventory tracking and analysis
In effect, these systems function as virtual employees, helping businesses run operations with minimal manual input.
So imagine a seller running an online store on Shopify has to source products, pay bills, and run online marketing campaigns. With the help of Accio Work's digital staff, they can now delegate most of those tasks to AI agents, freeing up employees' time for more important stuff.
This could lead to even bigger things
Put together, these products suggest something important: Alibaba isn't just building AI models -- it's building AI systems that can run real businesses.
That's a very different positioning. If agentic AI becomes the next major wave, companies that deliver real-world automation tools -- not just models -- could have a significant advantage. Alibaba has demonstrated a glimpse of what it's trying to achieve in the long run.
While the opportunity is huge, execution will be key
Alibaba's positioning is promising. As the leading cloud infrastructure provider in China, it has a natural platform to deploy these tools at scale.
But there are real risks. Competition in this arena is intense, with global players like Palantir, Microsoft, and Amazon pushing aggressively into enterprise AI.
There are also company-specific challenges such as execution risk, regulatory and geopolitical risks, and weak investor sentiment toward the Chinese tech sector.
Even the broader agentic AI trend comes with concerns, particularly around security and control when AI systems execute real-world tasks.
What investors should watch
There is no question that the adoption of AI will grow. The key question is whether agentic AI will become widely adopted, and whether Alibaba can take a leading position in that space.
Here are a few points to watch in the coming quarters:
- Are platforms like Wukong and Accio gaining real enterprise adoption?
- Is Alibaba Cloud seeing stronger demand tied to AI workloads?
- Can Alibaba turn these tools into scalable revenue streams?
These will determine whether Alibaba's agentic AI strategy becomes a meaningful growth driver. If it does, Alibaba could be one of the most underappreciated AI stocks in the market today.
