Preface:

Within a mere 24 hours, two apps of national significance have propelled the battle for AI entry points to the forefront, redefining boundaries in the Agent era.

Historically, entry points to the mobile internet were concealed within icons, search bars, grid menus, and mini-programs. Now, the focus is shifting towards natural language commands, authorized cards, and capability contracts.

Author | Fang Wensan

Image Source | Internet

Bottleneck in Existing User Base: An Entry Revolution Driven by Interaction Efficiency

The simultaneous focus on AI entry points by these two industry giants is fueled by shared industry anxieties.

Data from the Ministry of Industry and Information Technology reveals that in the third quarter of 2025, the monthly average user time spent on top apps had dwindled from a 12% growth rate in 2023 to a mere 4%, highlighting a clear limit to user attention.

For super apps, the strategy of driving usage growth by adding more features and entry points has become increasingly ineffective.

The advent of AI-driven conversational interactions presents a potential breakthrough.

This transition is akin to upgrading from a two-dimensional, hierarchical entry system to a one-dimensional, flattened command system.

Despite their differing approaches, both products share a core objective: leveraging AI technology to enhance service distribution efficiency and overcome growth bottlenecks in a mature market.

A deeper context is that the total transaction volume of third-party comprehensive payments reached RMB 428.6 trillion in 2025, with overall industry growth slowing to single digits.

In a saturated market, every incremental gain in market share demands exponentially higher costs.

The interaction efficiency improvements brought by AI have become a new lever for industry giants to tap into the existing market and unlock greater value per user.

Abao Swipe Right: Compressing Life Services into a Single-Sentence Entry Point

Alipay's redesign is a bold move, placing the AI version's entry point with a right swipe in the app and streamlining the interface into two main pages: 'Abao' and 'Assets.'

Users no longer need to navigate through a dense grid of services on the home screen. By simply stating their needs, Abao can comprehend intentions, match services, and complete tasks such as hailing a ride, ordering food, paying bills, sending parcels, scheduling repairs, checking housing provident funds, and locating charging stations.

The apparent incentive for this design is clear: as super apps offer more services, entry points risk becoming cumbersome.

When users are confronted with a multitude of buttons, channels, and mini-programs, the abundance of services transforms into cognitive overload. The AI version of Alipay aims to address the age-old problem of 'too many services, but users can't find them.'

Abao's value lies in transforming 'searching for entry points' into 'stating needs,' effectively rewriting the logic of service distribution.

Alipay has the confidence to pursue this strategy, given that China's online payment user base has surpassed 1 billion, and payment scenarios enjoy extremely high penetration rates.

Alipay's AI payment data also indicates that it has completed 300 million intelligent agent-based payments and supports 95% of general intelligent agent frameworks.

Alipay's goal is not merely to overlay an AI shell onto its existing app but to reorganize services previously scattered across different pages, mini-programs, and life scenarios into an AI-native ecosystem.

The 'Abao' page serves as the ecosystem's front desk, the 'Assets' page as its financial hub, and payment capabilities, billing abilities, and service fulfillment capabilities act as the underlying utilities.

However, constructing such an ecosystem comes with inherent challenges. Whether the ecosystem functions smoothly depends on whether every service interface is seamless, every merchant system is cooperative, and every niche demand is accurately understood.

The AI entry point has been established, but connecting the deep waters of services will take time. While it's easy for users to state a need, translating that into reliable cross-platform, cross-merchant, and cross-payment chain execution is no simple feat.

If a payment platform is reduced to merely 'finalizing the payment,' the value of its entry point will be ceded to upstream AI assistants.

In the Agent era, entry points are no longer just buttons clicked by users but the platform's ability to understand and fulfill user intentions.

AI Exclusive Cards Turn Left: Pressing the Battle to Payment Permissions

WeChat's approach appears more measured. Instead of directly launching a comprehensive AI version of WeChat for all users, it first opened up a mini-program AI development mode and then introduced AI exclusive cards through WeChat Pay.

Compared to Alipay's 'swipe right to enter the ecosystem,' WeChat seems to be laying the groundwork: first defining lanes, then setting up intersections, and finally integrating toll stations.

The core of WeChat's mini-program AI development mode is enabling developers to encapsulate mini-program capabilities into atomic interfaces and components callable by AI.

Relevant tools will focus on skill generation, validation, and evaluation, with mcp.json maintaining interface contracts, atomic components rendering return results into interactive cards, and AI calling services within the capabilities declared by developers.

AI exclusive cards fill the most sensitive gap in the consumption closed loop. WeChat Pay's AI exclusive cards, embedded within WeChat Change, have already been integrated with Tencent's desktop office efficiency agent, WorkBuddy, and will be opened up to more agent platforms in the future.

Its product logic is that users can propose consumption needs in agent conversations, with the agent completing recommendations, placing orders, and initiating payments. However, funds come from an exclusive card isolated from the main account, with the user transferring the limit and still confirming each transaction.

This is a very WeChat-like restraint. WeChat hasn't handed over the main account directly to agents or granted them boundless payment permissions. Instead, it uses an isolated card to keep risks contained.

Main account isolation, self-managed balances, and transaction-by-transaction confirmation collectively address the user's most fundamental concern: AI can handle tasks for me, but it can't spend my money recklessly.

For WeChat, with its 1.4 billion monthly active users, this infrastructure-minded approach aligns better with its ecological position.

Alipay is building a visible AI ecosystem, while WeChat is constructing an invisible AI transportation system.

Reaggregating Services vs. Safeguarding Ecological Access Rights

The differing approaches of the two companies stem from their distinct foundations.

Alipay's foundation is built on payment trust and life services. It is naturally close to money, billing, consumption, government services, and finance—scenarios that are infrequent but highly trusted, where users open Alipay with clear tasks in mind.

Past attempts to add content, social features, and life channels were essentially efforts to increase dwell time and service entry frequency.

With the advent of Abao, this strategy has shifted: instead of merely pursuing how long users stay on Alipay, the goal is to have users delegate more life tasks to Alipay.

This is particularly crucial for Alipay, as the initial AI ecosystem scenarios integrated by WeChat include e-commerce, food delivery, travel, and tourism—areas that overlap with the consumer services Alipay has long sought to dominate.

When users can say 'book me a ticket,' 'call me a ride,' or 'find nearby deals' in WeChat, WeChat gains the opportunity to stand at the task initiation point.

By launching Abao, Alipay is essentially moving the entry point battle to the Agent layer to avoid being reduced to a 'payment confirmation page' in future transaction chains.

WeChat's foundation, on the other hand, lies in its relationship chains and ecological network. Tencent's Q1 2026 earnings report shows that WeChat and WeChat combined had 1.432 billion monthly active accounts.

Its strength lies not just in user scale but in the internal circulation formed among social distribution, content consumption, mini-program operations, Channels, Search, payments, and merchant tools.

This explains why WeChat prefers to 'lay the groundwork.' It already has an ecosystem and doesn't need to tear down all buildings to rebuild; it needs to ensure AI vehicles can navigate the existing ecosystem orderly.

The mini-program AI development mode addresses capability invocation, AI exclusive cards address payment authorization, and future additions like search, recommendations, Channels, Enterprise WeChat, and merchant tools will enable WeChat to turn agents into a new form of traffic distribution within its existing ecosystem.

From an industry logic perspective, Alipay leans toward 'centralized reaggregation,' while WeChat favors 'distributed protocolization.'

Alipay hopes that once users enter Abao, the platform will uniformly understand intentions, orchestrate tasks, and dispatch services.

WeChat hopes developers will proactively make their capabilities available for AI invocation, with the platform providing dispatch rules and a payment foundation.

These two paths also entail different risks. For Alipay, constructing an ecosystem, the biggest challenge is whether the services within the ecosystem are sufficiently complete and stable.

If AI can only open entry points but cannot fulfill complex tasks, users will quickly revert to their original search and click paths.

For WeChat, laying the groundwork, the biggest challenge is if the roads are built too slowly, users won't perceive the value.

No matter how secure AI exclusive cards are, if merchant adaptations are insufficient, payment channels are incomplete, and agent scenarios are limited, users are unlikely to form high-frequency usage habits.

In the Agent era of entry point warfare, the object of contention has shifted to the delegation rights for users' next tasks.

Conclusion:

Once AI starts booking tickets, ordering food, buying coupons, paying bills, hailing rides, and renewing subscriptions for users, platforms are no longer just competing for attention but also for permissions, transactions, and responsibility.

Thus, the significance of Abao and AI exclusive cards extends beyond the product level; they represent the two companies' differing bets on AI commercial infrastructure.

In the long run, the outcome will be determined by service coverage, payment security, merchant adoption willingness, recommendation rule fairness, and responsibility boundaries.

The true dividing line lies in who can turn the permissions, fulfillment, settlement, and trust mechanisms behind AI entry points into everyday infrastructure.

The next generation of super entry points will not belong to the most talkative AI but to the system that can most safely complete tasks.

Partial sources referenced: Alipay: 'Alipay AI Version Officially Launched, Intelligent Assistant 'Abao' Brings a New Interactive Experience,' WeChat Pay: 'WeChat Pay Releases AI Exclusive Cards, Exploring Payment Solutions in the Agent Era,' 36Kr: 'Super App Entry Revolution: Alipay and WeChat's Dual AI Battles,' Huxiu APP: 'AI Reconstructs Payment Entry Points: Different Choices for Two Super Ecosystems,' Eqiu.com: '2026 Payment Industry AI Evolution: From Entry Point Competition to Infrastructure Competition'