Two problems, two audiences
Pamper Me’s online presence was limited. Customers had no good way to browse treatments and book, and the site did little to turn visitors into customers.
Behind the scenes, the information staff needed to run the business — customers, staff schedules, inventory and finance — lived in separate systems. A quick decision meant logging into several of them and assembling the answer by hand, and what should have been a thirty-second question routinely took fifteen minutes.
For staff: fifteen minutes down to thirty seconds
An operational question that took about fifteen minutes of report-gathering now gets an answer in under thirty seconds, from real-time data rather than outdated spreadsheets.
Staff ask a chat-based AI assistant, which answers questions and helps them carry out actions across the business. It runs on a master-agent architecture orchestrated in n8n: a master agent receives each request, classifies its intent and calls the specialised agents it needs — inventory, customers, staff and finance — running them in parallel and merging their results into one answer.
Conversation context carries across follow-up questions, so a manager can ask about this month and then simply ask "what about last month?". The agents connect to the existing scheduling, accounting and customer-review systems through their APIs, with error recovery so one unavailable source does not sink the whole response.
For customers: browse and book on any screen
The customer site is a responsive React 19 application with Bootstrap 5 components, designed for mobile and desktop. Visitors browse a catalogue of treatments with detailed descriptions and look at provider profiles and availability, with data loaded from the backend APIs through Axios. It shows off Pamper Me’s services properly, and has brought higher customer engagement and more bookings.