The Business Challenge
The business situation the organisation faced before the project began.
Small accommodation businesses spend significant time answering repetitive WhatsApp questions, checking room availability, and managing bookings manually. This slows response times and reduces customer satisfaction.
Why AI Was the Right Solution
Why AI was the right approach — and what alternatives were considered.
Customer messages are often written in different ways and cannot always be handled effectively with fixed menus or simple keyword rules. AI was suitable because it can understand natural-language questions, maintain conversational context, and generate relevant responses while following defined business instructions.
How the Solution Was Delivered
Discovery, design, development, testing and rollout — the journey, not the tooling.
- 1Business workflow and common customer questions were identified
- 2Conversation and booking inquiry flows were designed
- 3AI prompts and response rules were developed
- 4The assistant interface and supporting web components were built
- 5Typical hospitality scenarios were tested and refined
- 6The prototype was prepared for future WhatsApp API integration
Key Technical & Architecture Decisions
Architecture, model selection, workflow and trade-offs.
The solution was designed with separate layers for the user interface, conversation logic, AI prompting, and business data. JavaScript was used for the application flow, while Supabase was selected as the planned data layer for rooms, bookings, guests, and settings. The architecture was kept modular so that a WhatsApp provider, AI model, or booking system could be replaced later without rebuilding the entire application.
Challenges & How They Were Solved
Obstacles hit along the way and how they were overcome.
The main challenge was preventing inconsistent or unsupported AI responses. This was addressed by defining clear prompt instructions, limiting the assistant to hospitality-related tasks, and using structured booking flows. Another challenge was keeping the experience simple on mobile devices, so the interface and conversation steps were designed to be responsive and easy to test.
User Adoption & Change Management
How users responded — training, change management and feedback loops.
Because this is currently a portfolio prototype, no live staff rollout has been completed. The interface was designed to require minimal training, with familiar WhatsApp-style conversations and clear booking steps. A future rollout would begin with assisted mode, allowing staff to review AI responses before enabling full automation.
Business Outcomes & Impact
The measurable outcomes from the underlying AI Deployment — with the story behind them.
Prototype demonstrates automated handling of customer inquiries with significantly reduced manual responses and improved response consistency.
Business Outcome Categories
The story behind the numbers
The prototype demonstrates how hospitality businesses could reduce repetitive manual responses, provide faster and more consistent customer support, and handle booking inquiries outside normal business hours. It also provides a reusable foundation for future integration with live room availability, customer records, and booking systems.
Lessons Learned
What surprised the team, what worked well, and what would be done differently.
- Clear business rules improve AI response consistency
- Structured conversation flows reduce user confusion
- AI should provide an easy handover to a human
- Mobile usability is essential for WhatsApp-based services
- Prototype testing should cover unclear and unexpected questions
- Modular architecture makes future integrations easier
Future Opportunities
Where this solution could go next.
Future versions could include live WhatsApp API integration, real-time room availability, automated booking confirmation, payment links, multilingual support, conversation analytics, staff handover, and a dashboard for monitoring AI performance. The assistant could also be adapted for restaurants, clinics, property rentals, and other service businesses.
Professional Reflections
Guidance for another builder tackling a similar problem.
This project showed me that effective AI automation requires more than connecting an AI model. The business workflow, response boundaries, fallback behaviour, data structure, and user experience must all be designed carefully. Starting with a focused use case made the solution easier to test and improve.