AI System for Turning Hotel Event Notes into Work Instructions

The problem

Room layouts, service times and equipment changes accumulate in email attachments and WhatsApp messages.

The banquet manager transfers each change separately to kitchen, room setup and technical team lists.

Staff have to confirm separately that teams received the same plan version and saw the changed preparation tasks.

What we set out to improve

• Whether changes match the correct event and approved starting plan
• Whether departmental instructions use the same approved plan version
• Whether missing receipt confirmations and open tasks affected by changes are visible

How the system works

1. 01Link messages and documents received by email or WhatsApp to the event code and approved starting plan. Use only the official WhatsApp Business Platform for WhatsApp.
2. 02OCR converts writing in scanned documents into text. Extract times, rooms, layouts, attendee counts and equipment requests from messages and plans.
3. 03Compare the new request with the previous plan. Flag conflicting instructions and missing details.
4. 04Use the hotel’s work rules to prepare the effect of each change on kitchen, room setup and technical service tasks.
5. 05The banquet manager approves the scope and latest version. Create departmental work instructions linked to that same version.
6. 06Track departmental receipt confirmations and open preparation tasks. Reopen affected tasks for review when another change arrives.

Methods we used

• Match messages, documents and the starting plan through the event code.
• Use text extraction and document reading to show changed fields alongside their sources.
• Compare plan versions to prepare task drafts for affected departments.
• Keep the approved version, departmental instructions and receipt confirmations together in the hotel’s event table.

Where people stay involved

Data and security

Attendee lists and special requests have separate access permissions. Health and allergy notes are kept off the general event board and restricted to the authorised service manager. Sensitive content is separated before reading, and processing uses only fields needed for the task. Links between source documents, versions and authorised corrections are preserved; retention and deletion rules are defined with the hotel. WhatsApp communication respects permissions, opt-out preferences and required templates; notifications exclude sensitive content.

Who this suits