Ant Group's Abao AI Assistant Opens Public Beta: Can Conversational Transactions Replace App Operations?

📖 Glossary

AI Box (also known as Agent Computer / Agent PC), is a dedicated local hardware device that runs AI Agents. Pre-installed with an AI agent management system, plug-and-play, running 24/7. Users can remotely command AI to work via Discord, Slack, Telegram, WhatsApp, and more.

Abstract: On July 2, Alipay's AI assistant "Abao" officially entered public beta. Users can swipe right to enter a conversational interface and complete 72 daily services—from checking housing fund balances to claiming discount coupons—with a single sentence. The shift from "display-based" to "conversation-based" interaction represents a paradigm change in app design. But for everyday users, a true AI assistant shouldn't live only inside a phone—products like Kaihe AIBOX, a local AI agent computer that runs 24/7, let you direct AI through WeChat voice messages without even opening your computer, filling the gap that phone apps can't cover.

On July 2, Alipay officially announced that its AI-powered version, "Abao," has entered public beta. iOS and Android users can search "Abao" or "Ant Abao" in app stores or within the Alipay app to experience it—no invitation code needed. Users swipe right on the Alipay page to enter the AI conversation interface and instruct Abao to handle tasks in natural language.

This marks a critical step for consumer AI assistants shifting from "Q&A type" to "transaction type." But can conversational transactions really replace traditional app operations? Let's examine what Abao does and the industry trends behind it.

What Is Abao: Not a Chatbot, a Transactional AI

Many people's first reaction is to dismiss Abao as yet another chatbot. But Ant Group's positioning is clear: Abao doesn't just answer questions. It uses conversation to match users' intent to specific services and complete queries, transactions, coupon claims, shopping, and planning.

Simply put, instead of navigating menus to find the right entry point, you say one sentence and Abao takes you directly to the service.

The public beta launches with 72 smart service skills across six life scenarios:

Home Management: Cleaning services, repair appointments
Transportation: Ride-hailing, route planning, ticket booking
Shopping Deals: Discount coupons, price comparison, ordering
Government Services: Housing fund queries, document processing
Wallet & Bills: Bill inquiries, expense tracking
Companion Interaction: Casual chat, lifestyle suggestions

For example, when a user says "show me nearby food coupons," Abao recommends coffee, tea, fast food, and dessert deals within the conversation, guiding the user through ordering and discount redemption. Say "I need to buy an umbrella," and Abao filters by specifications and use cases, supporting browsing, selection, and checkout within the dialog.

From "Display-Based" to "Conversational": A Paradigm Shift

Traditional apps follow a "display-based" interaction logic—homepages packed with icons and entry points where users find, click, and navigate themselves. Alipay's homepage has dozens of function entries; finding a housing fund query might require three or four taps.

Abao transforms this into a "conversational" model—users state their needs, AI matches services, and jumps directly to the processing page. This is powered by a domain-specific life services model that improves recognition of vague, colloquial, and multi-condition requests. For instance, a user saying "I want to spend a weekend in Guizhou, nothing too tiring, good for photos" prompts Abao to generate travel recommendations based on destination, timing, pace, and photography preferences.

However, for fund-related operations (transfers, payments), final confirmation remains with the user. AI handles workflow routing and service interface deployment, while payment actions require manual user confirmation—a design that balances efficiency with security.

Abao AI assistant conversational service scenarios

The Ceiling of Conversational Transactions

Abao represents one direction for consumer AI assistants: wrapping app-based services in a conversational layer to reduce the cost of finding entry points. But it has structural limitations:

First, it depends on phone and app. You must open the Alipay app, swipe right to enter the conversation interface. If your phone isn't nearby or the app isn't running, Abao can't help.

Second, service scope is ecosystem-bound. 72 skills sound substantial, but compared to thousands of services within Alipay, it's a fraction. The public beta covers high-frequency scenarios; long-tail needs still rely on traditional entry points.

Third, it only works within Alipay. Abao can invoke services within the Alipay ecosystem—you can't ask it to check WeChat files, reply to Feishu messages, or monitor smart home devices.

This is exactly where local AI box products offer complementary value. Not all AI assistants need to be locked inside a phone—Kaihe AIBOX is a smart agent computer that runs 24/7 at home, with OpenClaw agent management system pre-installed. Users direct AI through WeChat voice messages without opening any app or turning on a computer.

Consumer AI assistant evolution from single device to multi-terminal

Two approaches serve different scenarios: Abao optimizes in-app service calling efficiency, while an AI box solves "how AI helps you when your phone isn't around." One subtracts within apps; the other adds beyond apps.

Three Trends in Consumer AI Assistants

From Abao's public beta, three trends emerge:

Trend 1: From "Can Chat" to "Can Act." Early AI assistants mostly stayed at the Q&A level. Abao turns conversation into a service invocation entry point—AI doesn't just answer "how to check housing fund" but directly opens the query page. This is a qualitative shift from information tool to transactional tool.

Trend 2: From "General" to "Domain-Specific." General-purpose large models can chat about anything but lack precision in specific life scenarios. Abao uses a life services domain model optimized for transactional intent like "help me claim coupons" or "I need an umbrella." Domain models significantly outperform general models in targeted scenarios.

Trend 3: From "Single Device" to "Multi-Terminal." Phones are just one载体 for AI assistants. Increasingly, users expect AI to run across computers, smart home devices, and standalone hardware. Like Kaihe AIBOX, which lets you direct AI remotely via WeChat—the AI monitors scheduled tasks, organizes information, and auto-replies even when your phone is off.

How Should Regular Users Choose?

If your primary need is convenience within Alipay, Abao's public beta is worth trying. Swipe right, say a sentence, and handle housing fund queries, coupon claims, and ride-hailing faster than navigating menus.

If you want an AI assistant that doesn't depend on a phone—one you can direct via WeChat voice, that runs tasks while you sleep, and connects across platforms—an AI box like Kaihe AIBOX is more suitable. Plug it in, and your AI assistant is on duty 24/7.

Conversational transactions are becoming reality, but a true AI assistant shouldn't be locked inside any single app. When AI works across platforms, devices, and around the clock, "conversational transactions" truly replaces app operations.

Ant Group's Abao AI Assistant Opens Public Beta: Can Conversational Transactions Replace App Operations?