HotelFlow: hotels don’t need another dashboard — they need the next action

HotelFlow is a hotel operations engine designed to streamline the tasks that arise before a guest's checkout, such as late arrivals, dirty rooms, broken AC units, extra towels, and VIP moves. Typically, these operations are handled through WhatsApp and phone calls, leaving managers to determine their next course of action. HotelFlow tackles this challenge by reading from one Excel workbook containing room data, staff information, reservations, rules, and chaotic events, then generating a live queue of the next 5, 15, and 30-minute tasks.

Each job is scored based on urgency, guest impact, arrival time, VIP status or order issues, and assigned to staff with the appropriate skills and location. If no staff is available, the system escalates the issue. Users can simulate hotel chaos to see the plan automatically adapt to new events.

The project was not built as another property management system (PMS); instead, it is the layer that decides what should be done next. HotelFlow is developed using Next.js 14, TypeScript, Tailwind, Vercel, and xlsx for importing hotel data. The decision engine is located in src/lib/engine.ts and is not based on an LLM chatbot. Vercel's serverless architecture allows the system to send the current hotel state when processing chaos or resolution incidents.

HotelFlow was a solo project undertaken by the author, who aimed to create an operations control room for hotel staff facing multiple issues simultaneously. The author chose the chaos button as the most memorable feature; by injecting a simulated chaotic event (checkouts, AC failure, leak, VIP move, unavailable staff) and observing how the NOW, 15, and 30-minute queues reorganize, the user gains a clear understanding of the platform's capabilities.