The Cloudy AI
The Cloudy AI: Experience air travel reimagined, with AI that anticipates your needs and creates a seamless experience.
Created on 2nd October 2024
•
The Cloudy AI
The Cloudy AI: Experience air travel reimagined, with AI that anticipates your needs and creates a seamless experience.
Describe your project
1 In-Scope: Personalized Travel Recommendations: The Cloudy AI analyzes passenger data and preferences to provide tailored flight suggestions, baggage recommendations, and even in-flight amenity choices. Predictive Disruption Management: By leveraging real-time data and predictive analytics, The Cloudy AI anticipates flight changes, cancellations, and other disruptions, proactively adjusting itineraries and offering alternative solutions. Seamless Communication and Support: The Cloudy AI communicates with passengers in a natural and conversational way, providing real-time updates, answering questions, and addressing concerns. Optimized Resource Allocation: The Cloudy AI analyzes data to optimize resource allocation within Akasa Air, anticipating demand fluctuations and improving operational efficiency.
2. Out-of-Scope: Physical Travel Arrangements: The Cloudy AI focuses on the digital aspect of travel, not directly managing physical aspects like ground transportation, hotels, or airport services. Personalized Content Creation: While the AI can personalize recommendations, it is not designed to create personalized content like travel blogs or itinerary suggestions. Direct Integration with Airlines: While the AI can integrate with existing booking systems, it does not aim to directly integrate with airline databases for real-time flight availability or pricing. 3. Future Opportunities: a) Expanded Travel Services: The Cloudy AI could expand its offerings to include personalized recommendations for travel insurance, airport lounges, and other travel-related services. b) Integration with Smart Devices: Connecting The Cloudy AI with smart devices like smartwatches and voice assistants can further enhance the travel experience. c) Advanced Personalization d) Cross-Industry Applications: The technology behind The Cloudy AI can be applied to other sectors like ground transportation, hospitality, and healthcare, offering personalized and seamless experiences for customers.
Challenges I ran into
The "Lost in Translation" Bug: A Tale of GenAI and Language Understanding One of the biggest hurdles we encountered while building The Cloudy AI was the "Lost in Translation" bug. This arose when the AI struggled to understand nuances in natural language during user interactions. The Problem: GenAI models rely on vast amounts of data to learn language patterns and generate responses. However, travel-specific terminology, slang, and user-specific phrasing can be tricky for the AI to interpret accurately. For example, a user might say "I'm flexible with my flight times" meaning they are open to different departure and arrival times. However, the AI might interpret this literally, meaning the user is willing to fly on any day, even with significant time differences. The Solution: We addressed this bug with a two-pronged approach: Fine-tuning the AI: We trained The Cloudy AI on a massive dataset of travel-specific text and conversations, including airline communications, travel blogs, and user reviews. This exposed the AI to a wider range of travel-related language, improving its understanding of nuances and slang. Contextualizing the Conversation: We implemented a "conversation context" feature. This means the AI remembers previous interactions with the user, allowing it to better interpret subsequent queries. For example, if a user asks about flights to a specific city, the AI will understand follow-up questions about baggage allowances or in-flight services in the context of that destination. Outcome By addressing the "Lost in Translation" bug, The Cloudy AI significantly improved its ability to understand user requests and provide accurate and relevant responses. This resulted in a more intuitive and personalized user experience, making travel planning and management smoother and more enjoyable. Lesson Learned: This experience highlighted the importance of addressing language nuances when developing AI-powered solutions. Fine-tuning the model with domain.
Tracks Applied (1)