The 2026 Inclusion·Bund Summit recently kicked off under the theme "Co-creating the AI New Economy." On-site visits revealed that AI capabilities are advancing rapidly—shifting from merely assisting humans with content generation, information retrieval, and decision support to now autonomously executing tasks on behalf of users.

Complex requests now take just 10 minutes to complete. On September 11, at the summit forum, MyBank unveiled its latest progress in "AI Banking," launching Bailing 2.0, the first agent for小微 (small and micro) inclusive finance. This agent serves 42 million small and micro users, offering credit, bill, and fiscal services. Additionally, eight AI workbenches have been deployed, with AI fully integrated into core business areas such as risk control, approval, marketing, and R&D. This marks a transition for MyBank's AI services from merely "getting things done" to "doing things well."

MyBank President Feng Liang stated: "The organizational evolution of AI banking will accelerate higher-quality financial inclusion, providing services that grow alongside every small and micro business operating with integrity." A live case study demonstrated this: Mr. Wang, a restaurant owner preparing to open a second location, already had a 150,000 yuan credit line. He requested, "I'd like to increase my limit." Instead of immediately submitting an application, Bailing 2.0 engaged conversationally, gathering details on his new store investment, existing store performance, and personal capital. It then analyzed projected revenue and repayment capacity based on his current store's cash flow, brand, location, and expansion plans, assisting him in compiling materials and submitting for review. Following approval, his credit line rose from 150,000 to 400,000 yuan, with flexible drawdowns aligned to his store-opening timeline. Given the high upfront costs, he only needs to pay interest for the first six months, deferring principal repayments. The entire process—from request to finalized plan—took roughly 10 minutes; previously, such complex requests required three days.

Throughout this process, small business operators' information is accessed only within authorized limits, and key decisions like credit limits and repayment schedules remain with authorized personnel. In reality, a single user request often requires collaboration across risk control, approval, marketing, and R&D. Relying primarily on manual information relay and piecemeal coordination makes it challenging for banks to respond promptly. Du Juan, a senior researcher at Suzhou Bank Research Institute, commented: "Previously, AI was mostly used to aid humans in generating content, retrieving information, and supporting decisions. Now, AI has advanced to stage where it can independently execute tasks. As AI banking accelerates, banks' organizational structures are being reshaped—breaking down capability silos between departments and roles. When a customer presents a business need, all stages work collaboratively end-to-end, promptly addressing the complex challenges faced by small and micro enterprises." Wang Pengbo, a senior financial industry analyst at Botong Consulting, added: "AI can identify customers' genuine needs, connect banking data and services, and help operators transform vague requests into actionable financing solutions while assisting with processing. This represents a remarkable advancement in current banking AI applications."

The summit also highlighted progress beyond banking, with payment institutions achieving significant milestones. Alipay's AI payment system has undergone a new iteration, officially moving from "payable" to "trustworthy payments." Alipay introduced an AI wallet agent and upgraded its AI payment product suite, launching three AI-taking capabilities—Vibe Pay, Skill Pay, and Machine Pay—alongside innovative applications like AI subscription, AI-assisted purchasing, and AI-assisted booking. Ant Group CEO Han Xinyi noted: "AI payment is no longer just the final step of a transaction; it has become the trusted commitment engine for the agent economy."

The AI wallet agent acts like a fund manager for users within the agent ecosystem. For instance, when a user asks Qwen to order coffee, Qwen handles product selection, while the AI wallet agent maintains full context of the transaction—linking the user's original intent and budget to the agent's execution and final outcome. After completing risk assessment and confirming the transaction aligns with user authorization, it processes the payment. This evolution elevates Alipay's AI payment from merely "enabling agents to pay" to enabling agents to complete value exchanges under user authorization and trust mechanics. Alipay data shows that agent-assisted purchase tasks have grown sevenfold in the past five months, while average monthly user payment frequency has tripled.

Previously, payment was the "checkout action" at the end of a transaction. Now, it must permeate the agent's identity, intent, authorization, execution, and responsibility, serving as new infrastructure for scaling the agent economy. From supporting thought to executing actions, AI is achieving the leap from "thinking for people" to "acting for people." Establishing a comprehensive system that is identifiable, authorizable, trustworthy, and accountable is precisely the essential infrastructure needed for AI agents to break technological boundaries, land commercial applications, and sustain the thriving growth of the AI new economy.