The Business

A family-run boutique hotel group operating three properties across the Lake District and Yorkshire Dales. Combined 47 rooms, two on-site restaurants, and a small events space. 35 employees across all sites. Annual turnover approximately £2.4 million.

The Problem

Running three properties with a small team meant everything was reactive. The owners were stretched across operations, guest relations, and financial management with no breathing room. Specific pain points:

• Pricing was static — room rates changed seasonally but didn't respond to demand. Bank holidays and local events were underpriced. Midweek was overpriced. Average occupancy sat at 62%, well below the 75-80% achievable for their segment.
• Food waste was brutal — the restaurants were ordering based on maximum covers, not predicted covers. 28% of perishable stock was going in the bin weekly. That's roughly £1,800/month straight into waste.
• Guest communications were manual — booking confirmations, pre-arrival info, review requests, and follow-ups were all handled by one receptionist. Responses were slow, inconsistent, and frequently missed entirely during busy periods.
• No visibility across properties — each hotel ran its own spreadsheets. The owners had no unified view of performance, costs, or guest patterns.

What We Built

Phase 1: Unified Dashboard (Week 1-2)

Connected all three booking systems, restaurant POS systems, and financial data into a single dashboard. Real-time visibility across all properties for the first time — occupancy, revenue per available room, restaurant covers, and cost tracking in one place.

Phase 2: Dynamic Pricing (Week 2-3)

Built an AI pricing engine trained on 2 years of booking data, incorporating:

• Local event calendars (fell running events, food festivals, bank holidays)
• Weather forecasts (Lake District bookings are heavily weather-dependent)
• Competitor pricing from OTAs (Booking.com, Expedia)
• Lead time patterns (last-minute vs advance booking behaviour)
• Day-of-week and seasonal demand curves

The system recommends price adjustments daily. The owner reviews and approves with one click. No more guessing.

Phase 3: Kitchen Demand Forecasting (Week 3-4)

Used occupancy data, booking patterns, historical restaurant usage, and local factors to predict daily covers for each restaurant. The AI generates prep lists and ordering recommendations 3 days ahead, adjusting daily as bookings update.

Phase 4: Guest Communication Automation (Week 4-6)

Built an AI-powered guest communication system that handles:

• Booking confirmations with personalised local recommendations
• Pre-arrival emails (dietary requirements, special requests, local tips)
• In-stay check-ins via WhatsApp for quick issue resolution
• Post-stay review requests timed for maximum response rate
• Repeat guest recognition with personalised offers

All messages are AI-drafted in the brand's warm, personal tone. Staff review flagged items only — about 15% of communications need human input.

Results After 4 Months

• Occupancy increased from 62% to 73% — dynamic pricing captured demand the static model was missing
• RevPAR (revenue per available room) up 22% — higher occupancy plus better pricing on high-demand periods
• Food waste reduced by 40% — from £1,800/month to £1,080/month in wasted perishables
• Guest review score improved from 8.4 to 8.9 on Booking.com — consistent, timely communication made the difference
• 15 hours per week saved on guest communications and admin
• Repeat booking rate up 28% — personalised follow-ups and recognition driving loyalty

ROI

Implementation cost: £8,500. Monthly platform and maintenance: £620.

Monthly benefit: approximately £6,800 (additional room revenue + food waste savings + staff time savings + increased repeat bookings).

Payback period: 7 weeks.

What Made It Work

• Owner buy-in from day one. The unified dashboard showed value immediately, before any AI was even running. Seeing all three properties in one place changed how they managed the business.
• AI recommends, humans decide. The pricing engine suggests, the owner approves. The kitchen system proposes prep lists, the chef adjusts. Trust built gradually.
• Hospitality tone matters. Generic AI-written guest communications would have damaged the brand. We spent real time training the system on their voice — warm, knowledgeable, never corporate.
• Seasonal patience. The pricing model improved significantly after its first full season cycle. Early results were good; post-season results were excellent.

Related: Learn more about our AI customer service and process automation services. Read about measuring ROI on AI integration, or see other case studies: retail, professional services, manufacturing.