Case Study
Revolutionizing Payment Services with Cutting-Edge Technology: VBill.cn
Under the brand Suixingfu, the company is dedicated to enhancing the operational capabilities of small and micro merchants through a strategic focus on "payment + technology". Suixingfu leverages advancements in AI, blockchain, big data, and cloud computing to foster the innovative growth of payment services
About VBill.cn:
Founded in 2011, VBill.cn is a leading smart payment platform tailored for offline transactions in China, boasting a comprehensive suite of licenses including bank card acquiring, internet payment, mobile payment, and cross-border RMB settlement. Distinguished as one of the few entities in China holding full payment licenses, VBill.cn's technology is certified by PCI and ADSS, and the company is recognized as an AAA credit and high-tech enterprise.
The company excels in four core areas: large-scale transaction processing, data-driven intelligent applications, comprehensive offline scenario development, and efficient market penetration. Its "payment + finance" model has yielded significant achievements in consumer finance, supply chain finance, new retail, and bespoke payment solutions for various industries.
Suixingfu’s business model concentrates on the payment sector, branching into supply chain finance, cross-border payments, and industrial payments. By integrating "payment + SaaS", the company executes its digital strategy, offering merchants a unified digital operations solution. As of the first half of 2021, the company collaborates with over 1,000 SaaS service providers, with its operations spanning more than 2,000 counties and towns nationwide, putting it at the industry's forefront. Its services cater to a wide range of sectors including supermarkets, community retail, smart property management, and logistics.
The Challenge: Harnessing Diverse Data for Actionable Insights
VBill.cn ingests large volumes of data from a variety of disparate sources every day. Internal transaction records, client persona profiles, service records, customer feedback, public reviews, social media mentions, consumer purchasing trends, etc. can all be corroborated to generate useful insights and uncover potential revenue opportunities. However, such analysis is impossible without data preparation and standardization.
The Solution
GPT-trainer team developed a separate, standalone instance of its core technology platform on AWS servers co-managed by VBill.cn’s IT team. This enabled all uploaded data to be physically isolated from other users and can be actively monitored by VBill.cn. This ensured that services performed by GPT-trainer conformed to relevant compliance standards.
After deployment, GPT-trainer worked with VBill.cn to categorize data of different formats and origins. Depending on the statistical analyses planned, the teams collaborated to establish a common standard for the required data format. Afterwards, GPT-trainer rapidly iterated on AI Agent setups to optimize the LLMs’ consistency in transforming select qualitative data from its original format into pre-determined target format.
Once the AI Agents have been trained, data from each category is fed into the associated Agent for transformations. When processing finished, GPT-trainer exported the data mapping and delivered it securely to VBill.cn. Since the pipeline is easily reusable and remains versatile to customization, GPT-trainer and VBill.cn’s IT team comanages it in an ongoing fashion.
Due to geopolitical constraints, the project was conducted using a combination of opensource and Chinese-native LLMs rather than OpenAI’s GPT series. GPT-trainer aims to introduce support for a greater range of LLMs in the future to accommodate the needs of a growing international client base.
GPT-trainer is SOC II and ISO 27001:2022 compliant.
