In 2024, Yang Chengzhi opened a coffee shop, homestay, small supermarket, and bakery in a 4A scenic area.

In order to facilitate the management of various store operations, he has added a set of financial SaaS (Software as a Service) software to each store.

To his surprise, because of the incompatibility between the financial SaaS software and the collection POS machine, Yang Chengzhi had to manually import the revenue data of the day into the financial SaaS software every day, so that the latter could combine the business data of the past week to show which products had exceeded the expected revenue and which products had not generated enough revenue.

Yang Chengzhi said that repeatedly manually importing revenue data every day not only takes a lot of time, but also often leads to errors in business decisions due to omissions in the manually entered data.

This entrepreneurship has transformed me from a specialist in roasting coffee production to a versatile student who needs to be proficient in business analysis. But I can't do this versatile student well, "he said.

Zhou Ye, Chairman and CEO of Huifu Tianxia, told reporters that Huifu Tianxia has served over 10 million small and medium-sized merchants. In the process of interacting with them, he found that many small and medium-sized merchants hope to use digital management to significantly improve management efficiency and business performance, but at the same time do not know how to correctly and efficiently use SaaS software digital tools such as product management, business analysis, membership management, and finance.

Many small and medium-sized merchants say that their digital operation is just a QR code for receiving payments, but they are at a loss on how to use the cash flow statements behind the QR code to explore new business opportunities and increase revenue, "said Zhou Ye.

Starting from this year, a hidden battle around small and medium-sized merchants using AI is underway among payment institutions. Many payment institutions have taken the initiative to provide AI tools covering enterprise finance, membership management, and business analysis to small and medium-sized merchants, enhancing their digital and intelligent operational capabilities.

What are these payment institutions up to?

Small and medium-sized merchants face AI challenges

Chen Bin is the head of a restaurant in Guangdong province, with an annual revenue of approximately 4 million yuan.

In order to "maintain" this revenue in the fiercely competitive catering market, he has been using multiple SaaS software since last year.

At the beginning of last year, he purchased membership management SaaS software and periodically sent dining coupons to loyal customers; Six months later, he purchased an intelligent ordering software that allows customers to quickly order and also speeds up the process of serving dishes in the kitchen.

For Chen Bin, who only has a high school education, it is quite difficult to proficiently operate these two SaaS software.

Over the past year, Chen Bin has repeatedly requested that smart ordering software service providers send people to adjust dish prices, as he has not understood the specific operating steps of this software.

When using membership management SaaS software to send dining coupons to old customers, Chen Bin also needs to spend a lot of time and effort. He said that it takes 1-2 hours each time, and the user interface of the membership management software is complex, making it easy to make mistakes and causing the operation to repeat from scratch.

Zhou Ye told reporters that this situation is quite common among small and medium-sized merchants. Through three months of market research, he found that small and medium-sized merchants mainly encounter difficulties in delivery, use, and operation in digital operations.

The person in charge of a leisure sports stadium told Zhou Ye that the staff spent an hour and a half repeatedly debugging just one rule of time based charging and employee sharing, because they didn't know how to use this SaaS software correctly.

Among small and medium-sized merchants who purchase marketing and membership maintenance SaaS software, the actual usage rate of such digital tools is less than 40%, because more than half of small and medium-sized merchants do not know how to combine their own business characteristics to carry out online customer acquisition, activate members who have not come to consume for a period of time, and improve repurchase rates through private domain marketing. ”Zhou Ye said.

This led Zhou Ye to develop an AI operating system for intelligent management of small and medium-sized merchants, highly integrated with various SaaS software capabilities and data, and ready to use out of the box - allowing small and medium-sized merchants to independently and correctly call various SaaS software with just one or two sentences, and complete their required revenue data import analysis, product management, membership management, and other business operations.

This job is easy to say, but it presents many challenges to do. AI operating systems need to have the ability to connect various SaaS software with various data, understand a wide range of industry management knowledge, intelligently schedule various SaaS software to accurately complete business instructions issued by small and medium-sized merchants, and convert dispersed heterogeneous data into a unified business language.

Recently, Huifu Tianxia launched the AI native intelligent agent operating system AIDA for merchant operations. Zhou Ye told reporters that some small and medium-sized merchants have been trying AIDA for nearly three months. Through the feedback from these merchants, Huifu Tianxia has been continuously improving its capabilities in unified management scheduling, refined permission control, and end-to-end intelligent orchestration.

He believes that there is still room for further expansion in the service boundary of this AI native intelligent operating system. For example, the pressure of fresh food loss in restaurants is quite high. If the system can proactively suggest price discounts and promotions for certain dishes based on the inventory status of fresh food in restaurants, it can help restaurants avoid food loss in advance.

As the head of the product innovation department of a payment institution, Liu Wei is also promoting AI tools for business analysis, membership management, and marketing promotion to small and medium-sized merchants. The approach of this payment institution is to first acquire a SaaS software service provider that provides business services for small and medium-sized merchants, and have the latter develop AI tools that load their own payment functions; Then rent and sell this AI tool to small and medium-sized merchants through the network channels of payment institutions.

Many payment institutions are following this practice, "Liu Wei said. By enabling small and medium-sized merchants to use AI, payment institutions can not only gain new revenue growth points from renting and selling AI tools, but also drive a corresponding increase in payment business revenue.

In the first three quarters of this year, AI tools increased the overall revenue of Liu Wei's payment institution by over 6% compared to the same period last year.

At the same time, Liu Wei revealed that the number of small and medium-sized merchants using AI tools has started to decrease recently, as the space for developing "easy to do" customers is limited, and other small and medium-sized merchants particularly value the input-output ratio.

Paving the way for Agent to Agent payments?

Liu Wei admitted that another reason why payment institutions promote AI tools to small and medium-sized merchants is to lay the foundation for the "Agent to Agent" payment business layout.

The so-called Agent to Agent payment refers to consumers completing various purchases and payments for goods and services through intelligent agents, while merchants also provide sales and receipts for goods and services through intelligent agents. So, the future payment entities will shift from "people and merchants" to two intelligent agents.

With the rapid development of AI technology, Agent to Agent payments have entered application scenarios from the concept demonstration stage.

Since the beginning of this year, multiple payment institutions have successively launched intelligent agent payment services and focused on expanding their application scenarios.

Liu Wei told reporters that once small and medium-sized merchants use AI, it will accelerate their entry into the intelligent payment system and further expand the market coverage of Agent to Agent payments.

In the past two months, a payment institution has recommended its Agent to Agent payment product service to Yang Chengzhi. Yang Chengzhi found that if he wants to introduce this new payment model, he needs to first invest in the AI transformation of the overall operation system of four stores, including authorizing intelligent agents to automatically retrieve merchant payment QR codes, enterprise accounts, and financial SaaS software.

If Agent to Agent payments could increase my coffee shop, homestay, and bakery business by over 30%, I would consider spending this money. ”Yang Chengzhi said.

At present, Yang Chengzhi has not taken any action. Because his friend suggested that he wait until the new payment system becomes more mature and reliable before choosing an opportunity to join.

Third party payment institutions believe that the development of this new payment technology still needs to overcome multiple obstacles.

On a technical level, in order for Agent to Agent payments to be widely implemented, it is necessary to address issues such as trustworthiness and security. This requires payment institutions to do a lot of work, including collecting all data within the ecosystem, completing data cleaning and risk modeling, and summarizing and integrating all operational knowledge of vertical industries. In addition, payment institutions need to efficiently integrate APIs (Application Programming Interfaces) with various resources to enable intelligent agents to autonomously and accurately call resources, data, and tools, ensuring that intelligent agent payments do not make errors.

More importantly, in the era of Agent to Agent payments, various payment institutions are still exploring what the corresponding payment carrier is.

Zhou Ye told reporters that as of now, there are still a few institutions worldwide that can implement "Agent to Agent" payments, and they mainly focus on specific vertical industries because vertical industries have concentrated knowledge, easy data processing, and require relatively clear intelligent agent payment capabilities, making it easier to implement.

He admitted that as AI increasingly reconstructs the payment relationships between merchants, financial institutions, and consumers, he does not know the ultimate form of future payments. But one thing is certain, that is, future payments will definitely be automated and intelligent.

In Zhou Ye's view, the payment industry is undergoing a new round of transformation due to the popularization of AI technology, but whether this round of transformation will completely change the current payment format depends on whether the interaction between the payee and the payer has undergone substantial changes.

In summary, it is whether AI technology truly turns the interaction between the payee and the payer into natural language. Its development speed largely depends on the digitalization and AI process of the majority of merchants. ”Zhou Ye said that in this wave of technological change, it is extremely important for the payment industry to maintain high accuracy.

(Author Chen Zhi)