An AI concierge that guides travellers to exactly three perfect boutique hotels through natural conversation, like chatting with a knowledgeable local friend.
Before
200+Generic results from a booking site, mixing chain hotels with authentic boutique properties
After
3 picksHandpicked in under two minutes, each with a personalised reason it fits, at 99.9% uptime
Overview
A conversation instead of a search.
Inndira helps travellers discover boutique hotels in New Orleans through natural conversation. Unlike traditional search engines that overwhelm users with hundreds of generic results, Inndira guides travellers through a personalised discovery, one question at a time.
The system combines multiple AI models and real-time search, and can email curated picks directly to travellers with insider tips and booking guidance.
The problem
Traditional hotel search is broken for boutique travellers.
Filters are binary, results are endless, and nothing in a booking site can hear "historic charm with modern amenities" and know what you mean.
01
Information overload
Major booking sites return 200+ results, mixing chain hotels with authentic boutique properties.
02
No personalisation
Filters are binary and cannot grasp nuance like "historic charm with modern amenities".
03
Decision fatigue
Travellers spend hours comparing options, reading reviews and second-guessing choices.
The solution
A conversation that feels like a knowledgeable local friend.
Inndira gathers preferences one question at a time, understands context, and accepts vague answers like "surprise me", adapting its recommendations accordingly.
Rather than hundreds of results, Inndira delivers exactly three handpicked properties, each with a personalised explanation of why it matches the traveller.
Inndira emails the curated picks directly to travellers, with hotel images, booking links, neighbourhood insights and insider tips, all validated for deliverability before sending.
LangGraph agent architectureTen-model fallback across Anthropic, OpenAI, Google AI and GroqProprietary curated hotel knowledge baseLive web lookupEmail delivery of curated picksCRM webhook lead handoff
3Curated picks, instead of 200+ generic search results
<2 minFrom first message to personalised recommendations
85%Reduction in operational costs versus traditional concierge models
99.9%Uptime, with fallback across four model providers
Questions
About this project.
It asks one question at a time, searches a curated knowledge base of 79 hand-enriched boutique hotels with more than 480 room types and 100+ structured fields per property, blends in live web lookup, and returns three ranked picks with a reason for each and booking links.
Travellers get three curated picks instead of 200+ generic results in under two minutes, operational costs fell 85% against traditional concierge models, and the service holds 99.9% uptime with fallback across four model providers.
A consent-first contact flow validates the traveller's email, sends the curated picks with insider tips and booking guidance, and delivers the qualified lead to the client's CRM by webhook.
