Logo

Automation Tool for a Country Estate

Digital service for booking, ordering, and guest communication

Industry:
Hospitality Countryside recreation
Type:
Telegram ecosystem (Bot + Mini App + AI Assistant)
Timeline:
6 weeks
Platform:
Telegram WebApp Server CRM

[client]

/01

The estate owner wanted to reduce routine calls and messages, increase repeat visits, and simplify booking.

At the same time, it was important to preserve the "human" communication and service approach.

[our responsibility]

/02
CJM and funnel design
Development of Telegram bot, Mini App, and AI assistant
Integration with internal CRM and check-in calendar
UX/UI design of all interfaces
Deployment, testing, and support

[client had]

/03
List of services
Booking calendar (Excel)
Service oversight by the manager
Banner

[team]

/04
1 Product Manager
2 UX/UI Designer
3 Backend Engineer
4 Frontend Engineer
5 AI Engineer
6 QA Specialist
[tech stack]
/05
React Native
React Native
React
React
TypeScript
TypeScript
Docker
Docker
Python
Python
FastAPI
FastAPI
PostgreSQL
PostgreSQL
Redis
Redis

[product features]

/06
Mini App

Mini App

  • Calendar and booking
  • Cottage selection
  • Additional services ordering
  • Payment
  • Event request
Telegram Bot

Telegram Bot

  • Reminders
  • Receipts and confirmations
  • Photo reports
  • Staff evaluation
  • Complaints and quick responses
  • Re-booking
  • Referral link
AI Assistant

AI Assistant

  • Checking available dates
  • Cottage matching
  • Answering questions
  • Ordering services via chat
  • Requests and feedback

[development stages]

/07
1 Interview with the owner and staff manager
2 Guest scenario mapping
3 Prototyping bot, Mini App, and AI dialogues
4 Integration with calendar and CRM
5 UI design of all touchpoints
6 Stage-by-stage testing by roles
7 Launch and support

[technical challenges and solutions]

/08
01
Challenge

Disparate data sources (spreadsheets, messages, documents)

Solution

Built a custom CRM integrating all orders, payments, and statuses

02
Challenge

Complex booking cases (partial overlap, rescheduling)

Solution

Custom booking logic with priorities and locks

03
Challenge

The neural network often gave overly "liberal" responses

Solution

Created a knowledge base + system of strict fallback responses and escalations

[results]

/09
  • guests who now book via Telegram
    80 %
  • admin workload
    15 min
    per day
  • Repeat visits
    38 %
  • The estate became perceived as a "smart service"

    Confirmed by post-implementation research