Alipay’s Bold Three‑Step AI Play

Alipay announced the launch of its AI Open Platform, positioning it as a foundational infrastructure for AI‑driven commerce. The author argues that AI payment can only succeed when transaction scenarios are fully connected, allowing AI agents to act as an aggregation and routing layer rather than merely presenting transaction data.

Three‑Step Roadmap

Step 1 (Sep 2025 – May 2026): AI payment infrastructure. Alipay aims to complete 300 million AI‑agent payments by May 2026, having already achieved 3 billion AI‑smart payments since the previous September. The AI payment ecosystem conference introduced Token Pay and AI wallet, solving the core infrastructure challenge.

Step 2 (June 2026): AI version of Alipay ("阿宝"). The new AI‑centric app embeds transaction services within a conversational interface, replacing traditional menu‑driven service discovery. Users speak a request and the AI automatically matches and executes the payment, shifting from "people find services" to "services find people".

Step 3 (July 2026): Open the AI platform to all merchants. Merchants can upgrade their mini‑program APIs to MCP plugins or Skills, enabling AI to call their services without rewriting code. A single API grants access to 1 billion Alipay users and can be invoked from AI assistants, smart car‑heads, AI glasses, IoT devices, and third‑party large‑model applications.

No code rewrite: merchants only need to adapt their APIs.

One API, cross‑platform reach: services are available on phones, AI assistants, car systems, glasses, and IoT.

Integrated settlement: hosting, distribution, transaction, and settlement are fully managed by the platform.

Reduced development cost: unified protocol conversion means a single integration serves multiple endpoints.

The author likens this approach to SWIFT, acting as a universal translator that routes services from any device to the Alipay ecosystem.

Despite the technical promise, challenges remain. Alipay’s 1 billion users currently use less than 10 % of its functions because users must manually locate services (e.g., ordering food via Ele.me). AI‑driven automatic matching is presented as the solution. However, shifting user habits from traditional payment flows to conversational AI interactions is identified as a major hurdle.

Competitive pressure is highlighted: WeChat Pay has already released an AI‑specific card, leveraging its dominant social network; other players such as Douyin, Meituan, Huawei, and global AI giants are also pursuing AI‑payment standards and ecosystem links. This competition is expected to accelerate ecosystem evolution.

In the broader context, the author frames AI as the new "transaction operating system" for smart agents (car heads‑up displays, AI glasses, large‑model assistants). By aggregating services, routing them across diverse hardware, and closing the payment loop with Token Pay and the ACT protocol, Alipay aims to become the central middleware that connects AI with the real world.

Ultimately, the success of this AI commerce ecosystem depends on merchant adoption and user willingness to engage with AI agents for everyday transactions.