On August 27th, the 19th Frost & Sullivan Global Growth, Innovation and Leadership Summit and the 4th New Investment Conference (hereinafter referred to as the '2025 Frost & Sullivan New Investment Conference') ESG and New Productivity Summit Forum, hosted by the world-leading growth consulting firm Frost & Sullivan (Frost & Sullivan, abbreviated as 'Frost & Sullivan'), was grandly held at the Shanghai Jing'an Shangri-La Hotel. The forum, themed 'Sustainable Development and New Productivity', brought together industry leaders, leading enterprises, investment institutions, and professional service providers. It focused on new investment and financing opportunities in the fields of sustainable development and new productivity, and jointly explored the capital and industrial forces that enable businesses to navigate cycles.
At this forum, Zhou Yuan, Chief Product Officer of Yiqiying and General Manager of the Fiscal and Tax Value Development Center under Shouyous Group, shared a keynote speech titled "Exploration of Agent Product Commercialization under the Artificial Intelligence Wave." The speech mainly covered three aspects: the three major trends in AI development by 2025, the introduction of "Digital Intelligent Workplace 1.0," and the development and prospects of "Digital Intelligent Workplace."
Zhou Yuan, Chief Product Officer of Yiyiying and General Manager of the Fiscal and Tax Value Development Center
The following are the key points from Zhou Yuan's speech:
I. AI Development Trends in 2025: Model Capability Enhancement and Deepening Agent Product Cognition
Zhou Yuan pointed out that against the backdrop of continuous iterations of large language models, performance improvement and cost reduction have become the norm. Although the release of new models such as GPT-5 has attracted attention, its influence is not as great as before. AI is increasingly becoming a basic tool, and model manufacturers still have room for deepening and specialized development.
This trend provides a mature market foundation for the large-scale application of AI, which is favorable for the emergence of large-scale and blockbuster AI applications. At the same time, users' understanding of Agent products is becoming more mature, and their choices are becoming more rational. Products that relied solely on demonstration effects in the early stages will gradually be phased out. The competitive threshold for products has significantly increased, and users now value maturity, brand, delivery capability, and actual value more than just price or pure technical performance.
In addition, the evolution speed of Agent products does not mainly depend on AI technology, but rather relies on the company's existing customer insights, SaaS capabilities, solution design, and delivery implementation strength. Among these, understanding the industry and the maturity of the company's SaaS solutions are key supports. Therefore, the conditions for the emergence of current B-side Agent hit products are as follows: understanding the industry, understanding customers, understanding business, possessing mature SaaS and delivery capabilities, and finally, having a grasp of AI technology and its boundaries.
II. Agent Product Practice: An Intelligent Fiscal and Tax Business Engine Driven by Large Models
Digital Smart Workshop 1.0, centered around digital smart advisors, digital smart accountants, and digital smart invoicers, achieves a comprehensive insight into AI workflow processes and intelligent reconstruction. Its concept is based on the AI Agent architecture, with large language models serving as the 'brain' to support reasoning and task execution.
In the field of financial and tax services, AI has achieved mature applications in multiple aspects, including: AI + marketing lead acquisition, intelligent outbound call lead screening, and knowledge graph content generation; AI + financial and tax delivery, realizing full-process automated operations and enhancing standardization through intelligent review; AI + advisory services, intelligent basic communication, policy interpretation and case matching to assist in consulting; AI + value-added conversion, customer profiling and precise business screening; AI + expert consulting, real-time risk monitoring, compliance strategy and case recommendation.
In addition, the core engine of Digital Smart Factory 1.0, 'Caiying Zhi Xin', constructs intelligent solutions covering multiple business scenarios such as digital accounting, digital consulting, and digital invoicing through multi-model fusion, knowledge base, memory mechanism, and process orchestration. It achieves intelligent processing across the entire link and is the core engine for the enterprise's fiscal and tax digital transformation. The system solves 50% of problems through the knowledge base, decomposes complex tasks into decision-making chains, strengthens constraints to reduce interference, and constructs high-quality test sets to ensure capability verification.
III. Commercialization Achievements and Scenario Implementation: Employee Efficiency Improvement and Cost Optimization
Zhou Yuan introduced in detail the business value brought by Digital Smart Workplace 1.0. In the bookkeeping scenario, 'Digital Smart Accounting' has increased the industry average efficiency from 80 sets of accounts per month to 800 sets per month, an efficiency improvement of 10 times. At the same time, it helps institutions reduce accounting labor costs by 42.5%, with an accuracy rate of bookkeeping as high as 90%, covering over 500 sub-scenarios across 10 major industries.
In the invoicing scenario, the 'Digital Intelligent Invoicer' has achieved a 330% improvement in invoicing efficiency, reduced labor costs by 40%, and provided 7×24-hour uninterrupted service, greatly enhancing invoicing accuracy and customer service experience.
At the same time, the market performance of Digital Smart Factory 1.0 is remarkable. Due to the highly measurable value brought by AI, the Digital Smart Factory Agent product has become the fastest-growing product in trial sales history at Yiqingying. The AI invoicing trial sale achieved 4 million revenue in its first month, and the AI accounting trial sale generated 1 million revenue in just 10 days, fully demonstrating the market's demand for AI products that can effectively reduce costs and increase efficiency.
IV. Future Outlook: Building an Evolving AI Business Ecosystem
Looking ahead, Yiyiying plans to build an Agent business system to build long-term comprehensive business competitiveness. It is currently constructing an evolving Agent business system that will go beyond single-point tool intelligence, forming a full path from terminal customer touchpoint demand to terminal customer delivery results. It aims to promote the AIization of the entire sales-to-delivery-to-market chain from marketing customer acquisition, financial and tax delivery, advisory services, value-added conversion.
The goal is ultimately to form an AI business ecosystem for the continuous evolution of the fiscal and tax service industry. Through systematic capability openness and empowerment, we aim to initiate a '10x' product evolution cycle, continuously broaden the ceiling of enterprise services, provide complete digital financial department solutions, and jointly drive the overall digital transformation of China's fiscal and tax service industry.
