๐ŸผRecentcollected in 74m

Alipay Launches Radical AI-Powered Conversational Revamp

Alipay Launches Radical AI-Powered Conversational Revamp
PostLinkedIn
๐ŸผRead original on Pandaily

๐Ÿ’กSee how a global payment giant is replacing traditional UI with AI agents to redefine the super-app experience.

โšก 30-Second TL;DR

What Changed

Transition from static app menus to conversational AI interaction

Why It Matters

This move validates the 'Agentic UI' trend, where conversational interfaces replace traditional navigation. It sets a new standard for super-apps to integrate LLMs directly into user workflows.

What To Do Next

Explore implementing a conversational intent-matching layer if your product has a complex menu structure that could be simplified by an AI agent.

Who should care:Developers & AI Engineers

๐Ÿง  Deep Insight

Web-grounded analysis with 26 cited sources.

๐Ÿ”‘ Enhanced Key Takeaways

  • โ€ขThe conversational AI revamp, internally codenamed "Bao Plan," introduces an AI assistant named "Ah Bao" (or "Abao") that allows users to switch to a completely new, conversation-driven interface with a single click, moving beyond simply embedding an AI assistant.
  • โ€ขAlipay employs a dual-track approach for integrating its vast ecosystem of services: encouraging willing merchants to actively integrate their services as AI-callable functions (MCP interfaces) and simultaneously using AI-powered visual interface comprehension ("screen-reading") to operate unmodified mini-programs under user authorization.
  • โ€ขAnt Group has been developing its own large language models (LLMs), including the Ling, Ring, and Ming series, with trillion-parameter models like Ling-1T, Ring-1T, Ling-2.5-1T, and Ring-2.5-1T being open-sourced and designed for efficient reasoning and agent interaction.
  • โ€ขAlipay AI Pay, an AI-native payment solution introduced in 2025, had already surpassed 100 million users by February 2026 and processed over 300 million AI payment transactions by May 2026, demonstrating significant early adoption of AI-driven payments.
  • โ€ขThe shift to an AI-first super app is a strategic response by Ant Group following regulatory interventions in 2020 that halted its IPO and placed restrictions on its lending operations, pushing the company to innovate and expand into new business areas like AI.
๐Ÿ“Š Competitor Analysisโ–ธ Show
Feature/PlatformAlipay (Ah Bao)WeChat (AI Agent)ByteDance (Douyin/Doubao)Alibaba (Qwen)
Core ApproachRadical conversational AI interface overhaul, optional one-click switch to AI-native UI.Embedded AI agent within existing social contexts (chats, official accounts), no separate interface.AI deeply integrated across services, Doubao as an AI-native app for e-commerce.LLM app integrated with ecosystem agents, full integration with Taobao for AI shopping.
Service ScopeCoordinates 10,000+ services (ride-hailing, food delivery, mutual funds, etc.) via natural language.Expected to invoke all ecosystem capabilities (mini-programs, payments, official accounts) via natural language.Direct orders and payments within Doubao app for e-commerce, general AI services.Handles 100M+ conversational requests daily across maps, ride-hailing, shopping, etc.
Payment IntegrationAI agents can complete payments with explicit user authorization; launched Token Pay and AI Wallet.AI agent can execute investment operations and other financial tasks with authorization.Users can place orders and complete payments directly within the Doubao app.AI-powered shopping journey includes order placement and fulfillment.
Developer IntegrationDual-track: encourages merchant integration as AI-callable functions (MCP) or uses AI screen-reading for unmodified mini-programs.Two modes: developers authorize source code access or encapsulate services into "Skills."(Details not specified for developer integration with AI agents)(Details not specified for developer integration with AI agents)
User Base (approx.)~1 billion users.~1.432 billion monthly active users.(Not specified, but Douyin is widely used)(Qwen serves 100M+ conversational requests daily)

๐Ÿ› ๏ธ Technical Deep Dive

  • Ant Group's AI models (Ling, Ring, Ming series) are built on an advanced Mixture-of-Experts (MoE) architecture.
  • Ling-2.5-1T and Ring-2.5-1T are trillion-parameter models, with Ling-2.5-1T designed for efficient reasoning and agent interaction, supporting context lengths up to one million tokens.
  • Ling-2.5-1T demonstrates high reasoning efficiency, matching frontier models on the AIME 2026 benchmark while using significantly fewer tokens (approx. 5,890 tokens compared to 15,000-23,000 for comparable systems).
  • Ring-2.5-1T utilizes a hybrid linear architecture to improve structured problem-solving, achieving gold-medal standards on academic mathematics benchmarks like the International Mathematical Olympiad 2025.
  • Ling-1T, a trillion-parameter model, was pretrained on 20 trillion tokens, with over 40% of the final pretraining data consisting of chain-of-thought data.
  • Fine-tuning for Ling-1T involved CoT-Evo (generating and evolving chains of thought) and an internal reinforcement learning algorithm called Linguistic-Unit Policy Optimization (LPO), which treats sentences as semantic action units for precise reward alignment.
  • Alipay's AI system can use "screen-reading" operations to understand and interact with unmodified mini-programs under user authorization, enabling compatibility without requiring developers to rewrite their services.
  • The Alipay AI Pay solution supports 95% of general intelligent agent frameworks.
  • Alipay has introduced "Token Pay" to handle extremely small, high-frequency payments and "AI Wallet" for autonomous agent spending, forming key infrastructure for the Agent Economy.

๐Ÿ”ฎ Future ImplicationsAI analysis grounded in cited sources

The conversational AI revamp will significantly increase user engagement and transaction volume within Alipay.
By simplifying access to over 10,000 services through natural language, users will find it easier and quicker to complete tasks, leading to more frequent use of the platform for various daily needs and financial transactions.
Ant Group's open-sourcing of its advanced LLMs (Ling, Ring, Ming series) will accelerate the development of AI agents and financial AI applications across the industry.
Providing access to trillion-parameter models designed for efficient reasoning and agent interaction will enable developers to build more sophisticated AI solutions, potentially fostering a broader ecosystem of AI-powered financial and service applications.
The competition between Alipay and WeChat in integrating AI agents will lead to a rapid evolution of "super app" functionalities and user expectations in China.
Both platforms, with over a billion users each, are investing heavily in AI agents to retain and expand their user base, pushing innovation in conversational commerce, personalized services, and autonomous task completion.

โณ Timeline

2020-11
Ant Group's planned IPO halted by Chinese regulators, leading to a strategic pivot towards AI development.
2025
Alipay AI Pay, an AI-native payment solution, was introduced.
2025-10
Ant Group released Ling-1T, a trillion-parameter general-purpose large language model, and the Ling 2.0 series, demonstrating its foundational AI capabilities.
2026-02
Alipay AI Pay surpassed 100 million users and processed over 120 million transactions in a week, marking significant adoption of its AI-native payment solution.
2026-05
Alipay announced over 300 million AI payment transactions and launched Token Pay and AI Wallet, further building out its AI payment ecosystem.
2026-06-16
Alipay officially unveiled its AI-powered conversational revamp with the "Ah Bao" assistant, representing its biggest platform overhaul in two decades.
๐Ÿ“ฐ

Weekly AI Recap

Read this week's curated digest of top AI events โ†’

๐Ÿ‘‰Related Updates

AI-curated news aggregator. All content rights belong to original publishers.
Original source: Pandaily โ†—