News Reporter |
News Editor | Wen Shuqi
Super apps with massive traffic don’t exactly have it easy in the AI era. If anything, they’re probably even more anxious about how to reinvent themselves.
Two months back, Alipay rolled out its AI-powered super service agent, “Abao.” Back then, it wasn’t really clear where the AI version of Alipay was headed. Some folks were saying the user experience still had room to improve. Based on what we’ve seen, the AI-powered Alipay comes with two main sections—”Assets” and “Abao.” Once you hop into the app, you can chat directly with the AI assistant Abao to handle things like checking social security, booking hospital appointments, ordering food, and paying bills. From that experience, “Abao” feels more like an AI assistant that can tap into Alipay’s existing services.
Even though the beta phase saw people scrambling for invites—making it harder to get access than a golden ticket—changing the habits of hundreds of millions of users is still going to take some time.
But in the AI race among the big tech players, Alipay can’t afford to slow down. On August 17, Alipay hosted its first AI ecosystem conference and shared the latest progress on “Abao.” According to Alipay, Abao has now connected over 10,000 AI-enabled services and supports cross-device access across 5 smartphone brands and 16 major carmakers. Alipay also unveiled its full-stack agent business infrastructure and the AHA protocol system for multi-agent cross-device connectivity, teaming up with more than 20 companies including Qwen, Huawei, OPPO, Xiaomi, and Geely to build a shared agent ecosystem.
The whole point of the protocol is to pave the way for merchants to jump in. With this protocol in place, once merchants complete their AI-powered business transformation, they can plug into smart devices through “Abao” and serve users across all endpoints.
After the conference, Tulu, who leads the cross-device connectivity business for Ant Group’s Abao, told reporters that merchants are eager to cut costs when connecting to different endpoints. With this protocol, merchants will only need to build their AI service capabilities once and then connect to different devices. On the flip side, hardware makers won’t have to juggle multiple merchants—Alipay sits right in the middle layer.
When the AI version of Alipay hasn’t exactly exploded among C-end users yet, this sounds more like an ecosystem-first strategy. In fact, when Alipay starts its AI overhaul within the app, the first thing it needs to tackle is the supply side.
In 2026, tech companies are competing on AI deployment capabilities, and the Agent space has become the main battleground for big tech’s AI ambitions. In the office agent track, Tencent’s WorkBuddy is already pulling in 20.97 million monthly visits, ByteDance’s Coco has carved out a solid lead in the agent development platform space, and Baidu and others are also consolidating their product lines to push more Agent exposure and usage.
Unlike other tech giants that are cutting into scenarios like office work or content, Ant Group can’t really build its AI strategy without leaning on its existing app ecosystem.
On the standalone app front, the company has already bet on Ant Afu, a health-focused vertical, and made some headway in user numbers. Back in 2024, Ant Group tried to create another life-services AI app called “Zhi Xiaobao,” but that didn’t go so well. As a national-level life services app, Alipay’s accumulated service ecosystem and payment-centric user mindset over the years are both an advantage for AI and a tough constraint to shake off.
The launch of Abao shows that Ant Group has figured out how it wants to approach AI within Alipay. After the public beta in July, Abao has been moving fast—first Alipay rolled out its AI open platform to let merchants, institutions, and service providers plug into AI capabilities, then it announced the AI upgrade of its “Tap to Pay” feature. With this latest protocol system, the strategic positioning of the AI-powered Alipay is becoming clearer.
Ant Group CEO Han Xinyi offered a bold prediction at the AI ecosystem conference: the agent economy is set to explode within the next 6 to 12 months. He attributes this tipping point to three conditions “maturing simultaneously”—the evolution and improvement of AI technical capabilities, a shift in user demand, and the formation of a sustainable value loop.
Han believes that every technological revolution gives birth to the roads and infrastructure of its era. And Alipay’s role is to lay down that “road” for the agent economy—providing the product and tech foundation while building out trust infrastructure, AI payments, blockchain, AI databases, privacy computing, and other foundational support.
Shi Zhen, the overall head of Ant Group’s Abao business, explains the agent economy as every user having their own C-end agent that can proactively—or even autonomously—assess user needs based on their intent and requirements, then decide which agent to call in for execution. This process generates new business value beyond just traffic, transactions, and conversions.
To some extent, this overlaps with general-purpose large model apps like Qwen and Doubao, which are also starting to onboard merchants for transactions. But Alipay believes its differentiation lies in multi-device collaboration. Tulu explains that Qwen, for instance, might have original supply, and the merchants it connects to might have their own agents with different service logic than Alipay. Alipay’s focus on multi-end support is about teaming up with others to fill in the “user intent” gap and build its edge around that.
But no matter how you join the competition, user attention and screen time are always finite. Across different scenarios, agents are all fighting for the same “conversation entry point.” With so many tech companies already crowding the AI agent leaderboards, Alipay’s AI exploration is still in its first phase. On the monetization question, Tulu says everyone is still in the infrastructure-building phase right now—the immediate goal is to make the ecosystem thrive, and until that happens, it’s hard to talk about “how to charge fees or how to take a cut.”
