Case Studies

AI-Enhanced Meeting Room Booking System for Coworking Space

I help lean teams unlock the power of AI—through rapid prototyping, process automation, and smart, scalable solutions.

Project Overview

To enhance the productivity and coordination of a local coworking community, I designed and implemented an intelligent meeting room booking system fully integrated with Slack and capable of understanding voice commands. This replaced a loosely managed system that relied on unstructured Slack messages for same-day bookings without conflict management or validation.

Problem Statement

The community previously relied on Slack messages to reserve meeting rooms, with no formal structure, time validation, or multi-day support. This often led to:

Double bookings and scheduling confusion
No historical data or usage insights
User frustration due to unclear availability

The goal was to maintain the team's Slack-based workflow while introducing a smart, structured, and user-friendly booking system, including support for natural language voice commands.

My Solution

I developed a Slack-integrated AI-powered booking system that allows users to:

Book rooms by typing or speaking requests (e.g., “Book Room A for tomorrow at 3 PM”*
Validate bookings across multiple rooms and times
Get immediate confirmations or suggestions for alternate slots

Key Features

⚙️ Slack Socket Mode Integration

Real-time interaction with users within their preferred platform.

🗣️ Voice-Activated Booking

Users can book by voice using OpenAI Whisper and GPT-based intent mapping.

🧠 Natural Language Understanding

AI parses input like “next Tuesday morning” and maps to calendar dates and actions

📆 Persistent Booking Records

All reservations are logged in a Postgres RDS database with validation checks.

Tech Stack

Component	Technology Used
Slack Integration	Slack Socket Mode API
AI & NLP	OpenAI GPT (for command mapping), OpenAI Whisper (voice transcription)
Backend Infrastructure	AWS EC2
Database	PostgreSQL via AWS RDS
Deployment & Hosting	Ubuntu server on EC2 with HTTPS, systemd for service control

Challenges & Solutions

Challenge	Solution
Users gave vague time expressions (e.g., “later today” or “next week morning”)	Used GPT-4 to interpret human-friendly time expressions into precise datetime ranges
Voice quality and accents varied widely	Fine-tuned Whisper configuration and filtered background noise for better accuracy
Need to prevent double bookings	Added conflict-checking logic in Postgres using SQL constraints and conditional inserts
Slack latency in receiving commands	Switched to Slack's Socket Mode for lower-latency, bidirectional communication

Key Takeaways

Seamless UX starts with meeting users where they are—here, within Slack.
Natural language interfaces (text and voice) can significantly improve usability when backed by solid validation logic.
Real-world AI systems succeed not just with good models, but with thoughtful integration, UX, and infrastructure.

Screenshots & Demo