All case studies

AI & Automation

Booking engine and AI concierge for nine properties

Direct bookings taken back from OTAs, with a multilingual concierge handling requests before and during stays.

Client

Summit Stay · Goa

Industry

Hospitality

Timeline

10 weeks

Team

5 (PM, 2 full-stack, AI engineer, QA)

Summit Stay · Goa — Booking engine and AI concierge for nine properties

Overview

Nine properties, most of their occupancy coming from aggregators, and 18-22% of every room night going out as commission. Their own website could not show live availability, so a guest who wanted to book direct usually gave up and went back to the OTA. Guest requests, meanwhile, arrived on five channels and belonged to nobody in particular.

The challenge

  • Direct bookings under 15% of room nights.
  • Rates and availability updated by hand, per channel, usually late.
  • Requests coming in on phone, email, Instagram, WhatsApp and the front desk with no owner.
  • No pre-arrival offer of transfers, dining or experiences.

How we approached it

01

Make the direct site the accurate one

A channel-manager integration keeps inventory and rates in sync, so the direct site is never stale and never undercut by an aggregator.

02

A concierge that can actually do things

It answers in English, Hindi and the Konkani-inflected phrasing guests actually use, and it creates real bookings and service requests rather than reciting an FAQ. Anything sensitive — a complaint, a refund — goes to the property manager with the full conversation attached.

03

Ask at the right moment

Transfers, dining and experiences are offered two days before arrival and on the first evening, which is where the conversions sit. We tested three other timings; these two won.

04

A human is always one message away

The agent says up front that a person can take over, and does so instantly when asked. Guests trusted it more, not less, for saying it.

What we built

Direct booking engine with live availability
Channel manager and rate parity sync
Multilingual AI concierge across web and WhatsApp
Pre-arrival and in-stay upsell flows
Payments, deposits and refund handling
Per-property performance dashboard

Technology stack

Frontend

  • Next.js
  • React
  • Tailwind CSS

Backend

  • Node.js
  • PostgreSQL
  • Payment gateway

AI

  • LLM concierge
  • WhatsApp Business API

The results

+29%

direct bookings

₹92 L

annual OTA commission saved

71%

guest queries resolved by the agent

+₹1,850

average upsell per stay

"A guest got an answer at 1 a.m. about an early check-in and booked a transfer in the same conversation. We own that relationship now instead of renting it from an aggregator every season."
Neha FernandesGeneral Manager, Summit Stay