AI Hotel Booking Automation

I want to overhaul the direct-booking flow for a group of small-to-medium hotels by inserting an always-on AI agent between guests and our current cloud-based PMS. The goal is simple: whenever a prospective guest lands on the website or opens the mobile app, the agent greets them in natural-sounding English, answers availability or amenity questions, pushes real-time rates pulled from the PMS, and walks them all the way through to a confirmed reservation—no human hand-off, no delayed follow-ups.

Scope of work
• Connect the agent to both the webfront and our iOS/Android builds through the existing booking widgets.
• Build the PMS bridge (REST/GraphQL endpoints are available) so room inventory, rate plans, and guest profiles sync in real time.
• Train the language model on our brand tone, FAQs, upsell scripts and local recommendations so replies feel personalised yet consistent.
• Implement a mini-CRM loop: if a guest abandons during checkout, the bot should nudge them by email or in-app message within minutes, quoting the same rate and holding rooms for a configurable window.
• Provide an admin dashboard where I can see chat transcripts, conversion metrics and make quick content edits without touching code.

Acceptance criteria
1. A guest can complete a booking, start-to-finish, through either channel without detecting they’re speaking to an AI.
2. Inventory, price and booking status shown by the agent match the PMS in under 5 seconds.
3. Abandoned cart follow-up messages trigger 90 % of the time within three minutes.
4. All data exchanges are logged and exportable as CSV/JSON for audit purposes.

Hand-off items: source code, deployment scripts, model fine-tuning files, brief user guide and a one-hour knowledge-transfer call.