August 7, Hangzhou. At the award ceremony of the 2026 Zhejiang Entrepreneurs Youth Sci-Tech Cooperation Week, Yang Yinfeng, Chairman of Zhejiang FANJIA Holding Co., Ltd. (hereinafter referred to as FANJIA), received the trophy for the “2026 Zhejiang Entrepreneurs Youth Role Model” with a calm smile.

At the helm of FANJIA, he is advancing an ambitious master plan: launching an AI embodied robot with autonomous payment capability. He states plainly that “the traditional SaaS model will largely be phased out in the future.” Before the audience, his path of breaking through industry boundaries unfolds from these two statements.

Perceive, Comprehend, Decide

“Perceive, comprehend, decide — these nine words are the core priority,” Yang Yinfeng summarizes the technical essence of FANJIA’s AI embodied robot.

“Perceive” refers to multimodal perception, enabling the robot to identify objects and people within the physical world. “Comprehend” relies on natural language processing to help the robot understand instructions. The real challenge lies in “decide”: under what conditions a robot may execute a payment, when it cannot, how spending caps are defined, and who holds authorization. These matters require multi-layered coupling of perception, rule engines and security mechanisms.

He uses a simple scenario: sending the robot downstairs to buy a cup of coffee. “You cannot take goods without paying. Payment is the most fundamental capability for robots operating in the physical world.” In Yang Yinfeng’s view, payment methods are undergoing their third major evolution. “The older generation carried wallets, our generation uses mobile phones, and in the future, intelligent agents may manage wallets on our behalf.” The agent can not only spend money but also earn and manage it, and payment merely serves as its entry point into the physical world.

FANJIA’s newly launched AI embodied robot has attracted widespread attention for its autonomous payment function. While the claim of being “the world’s first” awaits further third-party verification, its pioneering direction is clear: enabling AI to move beyond virtual dialogue and perform tasks in the physical world. Asked about payment security and privacy risks, Yang Yinfeng cites multiple solutions including Web3.0 distributed technology, hardware-level encryption and algorithmic encryption. He also reveals that the team is exploring the application of quantum encryption.

This strategic layout is not conceived out of thin air. As large-model technologies extend AI capabilities into the physical world, automated and intelligent payment is becoming an infrastructure-level demand. FANJIA aims to become one node of this infrastructure, venturing into the uncharted territory of “collaborative work between carbon-based and silicon-based entities.”

Seventeen Years Focused on One Mission

Behind this bold vision lies seventeen years of steady, down-to-earth accumulation.

In 2009, Yang Yinfeng founded the company with seven team members in a rented residential apartment, using pooled seed funding. Today, FANJIA has grown into a leading Zhejiang Entrepreneurs Pre-Unicorn serving more than 10,000 large-and-medium-sized enterprises, with a 50% coverage rate among main-board listed companies. Valued at USD 500 million, FANJIA has helped enterprises cut costs by over RMB 270 million cumulatively.

“Over the past 17 years, we have always taken user demands as the starting point of innovation.” Yang Yinfeng sums up the company’s product logic in one sentence: FANJIA builds whatever clients need.

The first large-scale validation of this logic took place between 2012 and 2015. Back then, corporate travel management still operated on the primitive model where employees paid upfront and submitted invoices for reimbursement. FANJIA rolled out FANJIA Travel, its intelligent travel management platform, shifting applications, approvals, bookings and settlements fully online and creating a closed-loop workflow with “no advance payment, no reimbursement, no physical invoices.” For enterprises, travel expenses are monitored from the moment they are incurred; for employees, the tedious task of pasting invoices is eliminated.

After 2018, FANJIA expanded its product boundaries. The company built a service-oriented enterprise supply chain platform, extending its business from corporate travel to enterprise centralized procurement and employee consumption, and launched the WorkBee point system. Employees earn points for cost-saving or low-carbon travel choices during business trips, which can be redeemed for benefits. In Yang Yinfeng’s words, this delivers “two-way empowerment: enterprises reduce costs and boost efficiency while employee satisfaction rises.”

During the interview, Yang Yinfeng repeatedly emphasizes eight words: improve efficiency, elevate well-being. He believes B2B business can only sustain long-term success by achieving both.

Extending AI Capabilities into Physical Space

If FANJIA’s first decade reshaped enterprise service workflows through digitalization, Yang Yinfeng has placed his bet on artificial intelligence over the last five years. His judgment on the industry’s future is sharp and direct. The AI era demands “live data” — data that machines can call in real time, analyze independently and use to generate decisions. When he saw AI make autonomous judgments through convergent logic, he said: “I knew we had to invest heavily in this field.”

Based on this insight, FANJIA developed its Enterprise Service AI Agent and launched the AI embodied robot from 2020 to 2025, gradually extending capabilities from the digital realm into physical space.

When asked about the competitive landscape where major tech firms are racing to build general intelligent agents, Yang Yinfeng remains clear-headed. “Any general model needs industry-specific data for training. Whoever has the deepest roots in the sector and the richest experience will gain a share of the market.” FANJIA’s 17 years of enterprise services and accumulated real transaction data form its inherent moat.

He summarizes the two core pain points of the B2B market: corporate efficiency and employee productivity. “We only need to serve enterprises well around these two priorities, and that is sufficient.” Looking ahead, he returns to a simple truth: the B2B market cannot survive on concepts alone. “If the product performs well, enterprises will adopt it. Good word of mouth keeps the business alive; without it, products rarely survive.”

Outside work, he boxes and plays table tennis. In boxing, you cannot flinch from incoming punches; you must watch your opponent and spot weaknesses. He says entrepreneurship is much the same. In Yang Yinfeng’s view, the rigid data model of traditional SaaS represents a major flaw within enterprise services. And his AI embodied robot may well be that long-prepared decisive punch.