Cabby Logo
Architecture Diagram

Cabby – The AI Solution to Expensive Office Commutes

Demo Video

🎥 Watch Cabby in Action:
YouTube Video :

Inspiration

During my internships in Noida and Gurugram, I experienced the same struggle every day. Cabs were comfortable but too expensive to book alone, so I often chose the metro. Like thousands of other employees, I found myself standing in overcrowded coaches during peak hours, while outside I could see hundreds of cabs carrying just a single passenger.

That made me wonder:

What if people leaving the same office, at the same time, and heading in the same direction could simply share one cab?

That's how this idea was born.

What it does

Cabby is an AI-powered Slack agent that intelligently matches employees traveling on similar routes at similar times, enabling them to share a cab instead of booking separate rides. By splitting the fare, employees enjoy a cheaper, safer, and more comfortable commute, while organizations benefit from reduced traffic, lower carbon emissions, and better utilization of existing transportation resources.

Employees simply provide:

📍 Pickup location
🏠 Destination
🕒 Preferred pickup time
⚙️ Additional commuting preferences

The AI analyzes route similarity and departure times to create optimal ride-sharing groups. Instead of three or four people booking separate cabs, Cabby recommends traveling together in a single cab.

User Journey

📌 User Flow Diagram

Cabby User Flow

The workflow consists of:

Employee submits commute details in Slack.
AI analyzes routes and timings.
Similar commuters are grouped together.
Employees approve or decline the ride.
Shared cab is booked.
Trip details are stored for safety and accountability.

The result is:

💰 Lower travel costs through fare splitting
🚖 Fewer vehicles on the road
🌱 Reduced carbon emissions
😊 Comfortable door-to-door transportation
♻️ Better utilization of available cab seats

For organizational safety, Cabby also maintains a record of ride groups, including who traveled together and the trip details, providing accountability whenever required.

How we built it

We built Cabby as a Slack-native AI agent that integrates seamlessly into employees' daily workflow, eliminating the need for a separate application.

Using Slack Bolt, Block Kit, and Socket Mode, we created an interactive experience where employees can submit their commute details, receive ride matches, approve or decline shared rides, and get real-time updates—all within Slack.

The backend is powered by Node.js and TypeScript, with SQLite handling user profiles, ride requests, and trip records. We integrated the Google Places API for accurate location autocomplete, Google Maps for route navigation, and the Gemini API to intelligently analyze routes, group employees with similar destinations and timings, and estimate shared fares.

To make Cabby future-ready, we exposed its capabilities through an MCP (Model Context Protocol) Server. Instead of limiting Cabby to Slack, the MCP server allows any compatible AI agent or productivity tool to securely access Cabby's functionality—creating ride requests, retrieving employee profiles, analyzing routes, monitoring system health, and managing trips through standardized tools. This transforms Cabby from a standalone Slack bot into an AI-powered transportation platform that can seamlessly integrate with future enterprise AI assistants and workplace automation systems.

System Architecture

📌 Architecture Diagram

The result is an intelligent, scalable, and collaborative solution that automates employee ride sharing while making daily commuting cheaper, safer, and more sustainable.

Challenges we ran into

Building Cabby wasn't just about matching people—it was about finding the right people to travel together.

One of our biggest challenges was accurately identifying employees whose routes genuinely overlap, rather than simply having nearby destinations. We also had to design a matching system that maximizes the chances of forming ride groups while keeping detours and waiting times minimal.

Another major challenge was ensuring rider safety and trust. Since employees are sharing rides with colleagues, we built features to maintain trip records and participant details, providing accountability and an added layer of security for every journey.

Balancing accuracy, efficiency, and safety was the key challenge in building Cabby.

Accomplishments that we're proud of

We're proud that Cabby solves a real-world problem with a practical, production-ready solution rather than just being a hackathon prototype.

Our AI accurately identifies employees traveling along similar routes, enabling meaningful ride groups that significantly reduce the cost of daily office commutes while minimizing unnecessary detours.

We also focused heavily on the user experience. By integrating directly into Slack with a simple, conversational workflow, employees can find and manage shared rides effortlessly without learning a new platform.

Most importantly, Cabby feels like a production-grade AI agent, not a hackathon demo. From intelligent route matching and safety features to a polished Slack experience and MCP integration, we built a solution that organizations could realistically adopt and scale.

What's next for Cabby

Our next goal is to take Cabby beyond the workplace and make it available to everyone. We envision a platform where anyone can find trusted ride-sharing partners traveling along the same route—without needing to belong to an organization.

To achieve this, we'll focus on building robust identity verification, safety measures, trust scoring, and secure ride history, ensuring users can share rides with confidence. We also plan to integrate with cab providers and real-time traffic data to make ride matching even smarter.

Ultimately, we want Cabby to become the go-to AI platform for safe, affordable, and sustainable shared commuting for everyone.