iCHEF, a smart catering technology brand, announced today (October 21st) that it has officially launched the "First Year of AI in Taiwan's Catering Industry" in response to the government's national AI plan, "Artificial Intelligence Island." iCHEF announced that it will upgrade the POS systems of more than 10,000 partner restaurants across Taiwan through system updates, and introduce five major AI application functions, including the AI customer service module "Xuefu." Secondly, it announced a partnership with Google Cloud, WeatherRisk, and RevtelTech to jointly launch "Chunjiao," an AI data management middleware platform designed specifically for Taiwanese chain brands and based on the Gemini AI model.
Targeting the two major pain points of high import costs and scattered data
iCHEF points out that the introduction of AI into the food and beverage industry is a clear trend. According to Fortune Business Insights, the global AI market for food and beverage is projected to expand from US$109.8 billion in 2024 to US$1257 billion in 2032, at a compound annual growth rate of 35.6%.
However, the restaurant industry still faces two major pain points in implementing AI. For independent restaurants and small chains, which account for 70% of Taiwan's restaurant industry output, the hardware investment required to upgrade AI systems, such as ordering machines, can easily reach hundreds of thousands of NT dollars, creating a significant barrier to entry. For medium and large chain brands, the pain point lies in the fragmented nature of data from POS, membership, inventory, and delivery platforms, making it difficult to integrate and analyze. This leaves AI in a difficult position.
iCHEF's two major plans this time are precisely aimed at breaking through these two key obstacles.
Thousands of restaurants upgrade on-site: AI ordering recommendation function
For independent restaurants, iCHEF's biggest move this time is to directly introduce five major AI application functions without changing the system or increasing equipment investment, allowing restaurants to "upgrade on the spot" to become AI ordering platforms.
The five major AI application functions include:
• Order recommendation module:At the core of this upgrade, iCHEF AI identifies best-selling items and optimal pairings based on the store's 45 days of actual operational data, and accurately recommends them in the "golden section" during QR code ordering. According to iCHEF statistics, this recommended section has a click-through rate of up to 25%.
• Smart Menu Manager:Store owners can batch update product prices or package options simply through natural language, with AI guiding the process logic and simplifying menu management across multiple platforms.
• Multilingual menu module:AI automatically detects and provides translation prompts tailored to tourists' needs. iCHEF noted that according to Ministry of Transportation data, Southeast Asian tourists accounted for over 30% of total restaurant visitors in the first half of 2025, nearly double the proportion of Japanese tourists. Therefore, this feature allows restaurants to seamlessly serve this emerging customer base.
• AI customer service module "Xuefu":Provide 24-hour AI customer service to assist employees in real-time querying system operations, and work with a real customer service team to improve support efficiency.
• Analysis and sales module:Provides downloadable operational diagnostic reports to analyze sales revenue, average order value, number of visitors, and sales trends, and offers preliminary optimization suggestions.
iCHEF co-founder Cheng Kai-you cites a Taiwanese restaurant and pub group in Taipei's Xinyi District, where each restaurant saw a NT$5 monthly increase in sales after implementing the new functionality. Overall, all restaurants using the module saw an average monthly increase of approximately NT$3000.
Owner Lin of the renowned Fu Ge Stone Oven Chicken restaurant in Jiaoxi also shared, "I used to be afraid customers would feel like I was being sold, so I didn't dare to actively recommend dishes. With AI recommendations, the system naturally displays signature dishes and side dishes, allowing customers to enjoy their meals with ease. It's also quietly boosted our revenue, unexpectedly adding over NT$50,000 a month."
Partnering with Google and RevtelTech: Taiwan's native AI restaurant assistant "Chun Jiao"
For medium and large chain brands with large amounts of data and decentralized systems, iCHEF has teamed up with RevtelTech to jointly develop and launch an AI data middle platform called "Chun Jiao", which is expected to be officially launched during the Christmas period at the end of this year.
Chunjiao is positioned as an AI restaurant assistant, aiming to integrate cross-system data and provide accurate decision-making. Its technical architecture has three key highlights:
• Import the Google Gemini model:This middleware uses Google Cloud Platform (GCP) as its native cloud architecture and incorporates the most advanced Gemini AI model at its core.
• Integrate local "microclimate" data: WeatherRisk, a key partner in this collaboration, has imported localized microclimate data for Spring Glow.
• Cross-system data integration (MCP): To resolve data silos, Chun Jiao uses the standardized communication protocol MCP (and utilizes Google Cloud's BigQuery and Dataflow) to quickly integrate POS, membership, inventory, and other data across systems, without requiring restaurants to replace their existing ERP systems.
RevtelTech founder He Yanrui explained that they chose GCP and Gemini because of their ecosystem maturity, high stability, and flexibility, ensuring that Chunjiao possesses robust natural language understanding capabilities. iCHEF noted that this allows businesses to directly use natural language to ask questions like, "How are sales at our Taoyuan store today?" or "Will business be bad if it rains?"
Lai Zhongwei, general manager of Weather Risk, pointed out that Taiwan has a diverse terrain and significant regional climate differences. "Chun Jiao" combines high-density meteorological observations with regional high-resolution forecasts, which can enable the catering industry to upgrade its decision-making from "large-scale weather forecasts" to "accurate weather forecasts for stores," greatly improving the efficiency of food preparation and scheduling.
Matt, owner of the breakfast chain "Haochu Breakfast," shared, "Previously, language models were often unstable when analyzing data, or couldn't ask complex questions. With Chunjiao, it truly knows everything and can provide professional advice. It's like a veteran on the team who's intimately familiar with the restaurants. It knows exactly how each restaurant is performing and how much food to stock in different weather conditions."
From "Assistance" to "Collaboration": Building an AI Talent Amplifier for the Catering Industry
iCHEF emphasizes that the evolution of AI in the catering industry has progressed from being a "support tool" (saving manpower) to a "collaborator" (inspiring inspiration), and ultimately becoming a role involved in "decision-making."
"AI isn't a god, but rather an amplifier of talent," iCHEF noted. The key to accurate AI recommendations lies in integrating data from POS, sales, membership, and inventory. This time, iCHEF is integrating technologies from Google Cloud, RevtelTech, and Weather Risk, aiming to make AI the "management expert" behind every business operator.
Cheng Kaiyou concluded, "By leveraging proven AI shopping guides to increase average order value, iCHEF allows restaurants to directly recoup their monthly AI POS system fees within the same month. This makes upgrading to AI systems accessible to restaurants and empowers the entire industry. We are thrilled to witness the dawn of AI in Taiwan's restaurant industry and embrace this once-in-a-decade technological dividend."
Promoting multi-cloud architecture operations and optimistic about the development of AI Agent applications
In a subsequent interview, Cheng Kaiyou stated that this update will be free for independent restaurants, but there will be a limit on the number of AI-generated reports and other content. For medium and large chain brands, "Chun Jiao" will be the main service. At the same time, it is also expected that with the continued development of AI technology, the threshold for automating restaurant management and operations through AI Agent services will be greatly lowered in the future.
Regarding the idea of AI application, Cheng Kaiyou believes that instead of using a single model to drive all AI functions, it is more inclined to use multiple AI agent services to collaboratively complete work tasks, so as to achieve higher execution efficiency.
Regarding the decision to partner with Google Cloud, Cheng Kai-you explained that the primary considerations were its cloud services' greater suitability for localized operations and the ability to upgrade catering services through its latest Gemini AI technology. However, Cheng also explained that iCHEF will continue to maintain a multi-cloud architecture to bridge diverse technical needs. Therefore, while maintaining its partnership with Amazon and utilizing AWS services, in addition to its current collaboration with Google Cloud, the organization will evaluate the use of additional cloud and AI services over time.
Is it possible to further import the Google Gemini nano model into the POS devices used in iCHEF services, allowing small-scale AI models to run on the POS devices, thereby enabling AI services to operate more efficiently and, when necessary, to collaborate with cloud-based AI services to unleash greater computing power? Cheng Kaiyou revealed that this may become the next development direction of iCHEF, but for now, the current design will continue to be used to collect user feedback.
In addition, the name of the AI customer service module "Xuefu" is derived from the pronunciation of "CHEF", and "Chunjiao" is named to fit the local Taiwanese style. Whether it will consider launching a service called "Zhiming" in the future may depend on how iCHEF's AI technology develops in the future.
