Case StudyAutomationAI

A Restaurant in Vake Replaced its Hostess with AI, Saving 4,800 GEL/month. Here is What They Do Now.

A Restaurant in Vake Replaced its Hostess with AI, Saving 4,800 GEL/month. Here is What They Do Now.
Antonio Araujo / unsplash

In Tbilisi's restaurant industry, guests choose speed. If someone wants to book a table for Friday evening in a cozy place in Vake, they won't wait half an hour for a reply. They will write to three establishments on Instagram and go to the one that replies first.

Today, we will analyze a real case study of a small family restaurant serving European cuisine in Tbilisi (Vake district, 6 tables, about 60–80 orders per day including delivery) that fully automated its incoming request processing using AI.

The Problem: 25% of Bookings Down the Drain

Before automation, calls, Instagram Direct, and WhatsApp messages were handled by a hostess-administrator. Their salary, including taxes (20% Income Tax + pension contributions) and meals, was around 1,600 GEL net (about 2,100 GEL in company expenses). Taking into account the shift schedule (two administrators), the restaurant spent 4,200 GEL per month on phone support payroll, excluding the costs of setting up their workspace.

At the same time, an audit of the chat logs revealed shocking results:

  • Loss of Night Bookings: About 25% of table reservation requests came after 11:00 PM. Administrators were sleeping at this time. By 10:00 AM, customers had already changed their plans.
  • Misses During Rush Hours: During peak hours (Friday/Saturday evening), the administrator physically could not keep up with Instagram messages because they were welcoming guests in the hall. Responses were delayed by 40–50 minutes.
  • Delivery Chaos: Takeaway orders were accepted manually in chats. Administrators mixed up addresses, forgot to inform the kitchen about guest allergies (e.g., "nut-free"), and often forgot to call back for confirmation.

AI Implementation: Digitizing the Restaurant in 3 Days

The restaurant owner approached aiNOW to solve this problem. We didn't build complex apps; instead, we integrated an AI assistant directly into existing communication channels:

  • Integration: WhatsApp Business API + Instagram API + website widget.
  • Knowledge Base: The AI was trained on the seating chart (table layout), the menu (ingredients, calories, allergens), delivery terms, and booking rules.
  • Synchronization: Real-time integration with the internal table booking sheet.

Results: The Implementation Timeline

Month 1: Recovering Lost Guests

The AI assistant took over 100% of incoming text requests. It responded in Georgian, English, and Russian within 2 seconds.

  • The number of missed inquiries dropped to 0%.
  • During the first month, the number of successful table bookings grew by 12%, solely due to the prompt processing of night leads.
  • The kitchen stopped receiving orders with mixed-up ingredients: the AI clearly recorded comments for the order and duplicated them in the CRM card.

Month 3: Saving 4,800 GEL/month and Shifting Focus

Three months later, the restaurant completely eliminated the services of shift phone administrators. The duties of welcoming guests in the hall were taken over by the senior waiter (with a salary increase), while the AI handled all correspondence, booking, and delivery orders.

Financial Result:

  • AI assistant expenses from aiNOW (monthly fee + API traffic): 600 GEL/month.
  • Previous phone support administrator payroll expenses: 4,200 GEL/month.
  • Additional profit from booking growth: about 1,200 GEL/month.
  • Net savings for the restaurant: 4,800 GEL per month.

What Does the Restaurant Owner Do Now?

The most common fear among restaurateurs is: "If you remove a human from the chat, the restaurant will lose its soul."

In practice, the opposite happened. Having freed up 4,800 GEL monthly and got rid of the endless quality control of chats, the restaurant owner finally focused on development:

  • Menu Updates: Personally found artisanal cheese and wine suppliers in Kakheti, improving the quality of the dishes.
  • Local Marketing: Launched targeted advertising in Vake and Saburtalo (the AI easily handles the influx of new leads from ads, as it has no limit on simultaneous dialogues).
  • Improving Service in the Dining Hall: Invested part of the saved money in training waiters, raising the establishment's average order value by 15%.

The AI assistant does not replace hospitality in the dining hall. It eliminates the routine and chaos in chats that prevent your business from making money.

Want to test an AI administrator for your cafe or restaurant in Tbilisi? Write to the aiNOW team at `ainow.ge`. We will build a demo version of the bot with your menu in 2 days.

FAQ

How does the AI employee learn the specifics of our business?

We integrate the AI with the company's knowledge base (catalog, guidelines, CRM) in 3-5 days. It learns from your real data.

What are the requirements for integrating the automation system?

The integration occurs seamlessly using the official APIs of messaging apps (WhatsApp Business API, Instagram Direct) and your CRM system.