-
-
Lil Gouda Loving TalkTuahAirline
-
Call Log
-
Flight Dashboard
Inspiration
Airports can be overwhelming, especially for travelers unfamiliar with their surroundings or dealing with long lines and awkward tasks like seat-swapping. I wanted to create a solution that reduces stress, eliminates confrontation, and empowers passengers with a proactive, human-like assistant. TalkTuahAirlines is designed to provide a seamless experience, combining practical features like airport navigation and TSA guidance with the ability to perform tasks, such as calling passengers for seat swaps or notifying users about flight updates.
What it does
TalkTuahAirlines is a multilingual, task-oriented AI assistant that sounds as natural as a human. Key features include:
Airport Navigation: • Guides users to gates, restrooms, restaurants, and lounges using preloaded airport maps. • Answers specific questions like, "Where's the nearest coffee shop near Gate A12?"
Seat Swapping: • Handles confrontation-free seat swaps by calling the current seat holder, negotiating the change, and notifying the original requester of the outcome.
Flight Information: • Provides real-time flight status, gate changes, and boarding times. • Answers questions like, "How full is my flight?"
TSA Guidelines: • Provides packing advice and explains TSA rules, e.g., "Can I bring a laptop in my carry-on?" • Helps users prepare for security checks by providing step-by-step tips.
Real-Time Alerts: • Notifies users about delays, gate changes, or emergencies with actionable recommendations.
Multilingual Support: • Available in English and Spanish, with natural, human-like voice responses powered by Eleven Labs.
Always-On Accessibility: • Available 24/7 via phone---no apps or internet required.
How I built it
• Backend: Built using Node.js to handle API calls, manage workflows, and facilitate seat swaps.
• Frontend: Developed a responsive interface with React and styled it using Tailwind CSS.
• Voice AI: Used Retell AI, a wrapper for:
• GPT API for natural language understanding and task processing.
• Twilio API for inbound/outbound call management.
• Eleven Labs API for lifelike voice synthesis in English and Spanish.
Data: My real American Airlines Flight Next Week :)
Challenges I faced
• Task-Oriented AI: Creating a flow where the AI could manage multi-step tasks like seat-swapping while maintaining smooth, natural interactions.
• Voice Integration: Fine-tuning Eleven Labs and Twilio to ensure the AI sounded lifelike and approachable.
• Time Constraints: Balancing ambitious features like seat swapping, airport knowledge base etc.
What I learned
• How to integrate multiple APIs to create a cohesive system that feels human-like.
• The importance of prioritizing user-focused features like multilingual support and proactive notifications.
• Building AI that does tasks, not just answers questions, adds real-world usability and value.
Log in or sign up for Devpost to join the conversation.