AI & Automation Lead · Inhouse Digitalisation

A German accommodation operator in a turnaround. Bookings, invoices, payments, cleaning and guest messages each lived in a separate tool, and none of them talked to the others. Quotes for a commercial property-management system came to about €60,000 a year — for software that still would not have matched how the business actually worked.

An in-house platform for web and Telegram, instead of the licence. It joins eight external systems — channel manager, accounting, banking, telephony, messaging, Google Workspace — behind a single login. I own the spec, the architecture and the deployment, and it runs in the EU region.

• Cleaning and crew logistics: about two hours down to ten minutes. Scheduling a day of cleanings and routing the crews was manual work across two tools. Bookings and cleanings now live on one calendar, assigned and rescheduled by drag and drop.
• A large group booking with its invoices: about an hour down to three to five minutes. Dates and units are entered once, the client gets a branded offer page, and one click creates the booking and raises the draft invoices.
• A single quote: about ten minutes down to one or two. The offer page is generated from the same data, so nothing is retyped.
• Two people run what once took a much larger team. Bookings, invoicing, payment control, cleaning, maintenance and quotes all run on this platform every working day.

An AI assistant operates the booking system through an integration server I built: preview and confirmation before every write, typed confirmation for the risky ones, an audit log underneath. Anything that touches money or a guest is drafted by the machine and signed off by a person.