Anyone who has waited too long for a table at a busy hotpot restaurant knows the problem from the customer side. Restaurant operators know it from the other side: not enough staff, high turnover, and rising wages. WiXtar thinks an AI store manager can take on a big share of that work, and it brought a working demo to Computex 2026 to prove it.
Editor
Kai T chevron_right
Table of Contents
What the AI Store Manager actually does
WiXtar, the food and beverage technology brand under Taiwan's BenQ Qisda Group, has teamed up with chain operator 6owl Door to run what it calls the world's first Agentic AI hotpot store. The centrepiece is a "24/7 AI Store Manager" that ties together the dining floor (front of house), the kitchen and stockroom (back of house), and head office decisions into one system.
Instead of a dashboard that simply reports numbers, the platform is built to act on them. It gathers data, reads the scene through cameras, and turns that into instructions for staff, from prep lists before opening to safety checks after closing.
Four AI tools under one roof
The system breaks down into four parts, each aimed at a job that usually leans on an experienced manager:
- AI store patrol robot: uses 360-degree cameras and vision-language models to run round-the-clock inspections, shifting focus by mealtime, from hygiene and grooming checks before service to food-prep monitoring during service and safety checks after.
- AI self-ordering kiosk: detects when a customer hesitates at the screen and nudges them with set-meal suggestions, which lifts the average order value and eases the load on counter staff.
- AI decision platform: pulls together sales, membership, inventory and ERP data to recommend ingredient prep, staffing and restocking, built on D8AI's AI Agent Builder.
- AI ordering and receiving: predicts orders from past sales, weather and holidays, then uses image recognition to verify deliveries and trim food waste.
Why restaurants are reaching for this
The pressure is financial. WiXtar points to Taiwan's hotpot sector, worth around NT$40 billion a year, where a labour crunch keeps frontline stores short-handed. A 2025 labour report there showed F&B salaries climbing 7.5 percent, yet vacancies and turnover did not ease, a sign that paying more is not solving the problem on its own.
That story will sound familiar to anyone running a restaurant in Malaysia, where steamboat and hotpot chains are a mall-floor fixture and self-ordering kiosks are already common at fast-food counters. The pitch here is to push automation past ordering and into the manager's daily judgement calls.
May Kang, chief executive of WiXtar, framed it as a shift in what software is for: "In the age of Agentic AI, global SaaS is transitioning from licensing sales modules to actively driving business outcomes."
Where and when
For now this is a showcase rather than a shelf product. WiXtar demonstrated the system at the BenQ Qisda Group booth (M0104) at the Taipei Nangang Exhibition Center during Computex, which ran from 2 to 5 June 2026. The company says it is scaling its Agentic AI model across more than 47,000 chain outlets in the Asia-Pacific region. A Malaysian rollout and pricing have not been confirmed.
If the demo holds up outside a trade show, the interesting question for local diners is not whether a robot takes your order, that already happens, but whether the person who used to run the floor is quietly replaced by software that never clocks off.
Cover image courtesy of Yihan Wang on Unsplash.