SKYFARE ASSISTANT

As travel resumed after the COVID-19 pandemic, airlines faced a shift in customer behavior and expectations. The cost distribution across different services changed, leading to a greater need for automation in customer support. The rise of digital transformation in the airline industry highlighted the importance of efficient, AI-powered chatbots to handle customer inquiries seamlessly.

The Airline Chatbot is designed to enhance the customer experience by providing instant responses to common queries such as:

Booking flights Checking flight status Understanding cancellation policies Providing baggage and meal information Collecting customer feedback This chatbot streamlines operations by handling repetitive inquiries, allowing airline staff to focus on more complex customer needs.

Platform: Developed using Dify.ai for an intuitive chatbot interface.

Natural Language Processing (NLP): Implemented LLM-powered intent detection to classify customer queries. Knowledge Base Integration: Connected to an airline-specific knowledge database for accurate responses. Custom Chatflow: Designed a modular workflow, allowing flexible updates for new routes, policies, and promotions. Deployment: Integrated into the airline’s website, app, and customer service channels.

Challenges we ran into

Handling Multiple Intents: Some users ask multiple questions in one message, requiring advanced query classification. Balancing Automation & Human Support: Ensuring the chatbot seamlessly transfers users to a human agent when needed. Updating Real-Time Data: Integrating flight status APIs while keeping response times efficient.

Accomplishments that we're proud of: Reduced Customer Support Load: Automated 60%+ of

common queries, freeing up human agents.

Improved Response Accuracy: Achieved 90%+ intent detection accuracy using AI-powered classification. Seamless User Experience: Designed an intuitive, multi-channel chatbot for smooth customer interactions. Enhanced Customer Satisfaction: Reduced average response time from minutes to seconds.

What we learned

User Behavior Insights: Customers prefer quick, direct responses over long explanations. Iterative Improvement: Continuous testing & feedback helped refine the chatbot’s language model and decision tree. Data Integration Matters: Connecting real-time flight and booking systems enhances chatbot reliability.

What's next for “SKYFARE ASSISTANT”

Multi-Language Support: Expanding to serve international travelers in multiple languages. Personalized Assistance: Using AI to provide customized travel recommendations based on user preferences. Voice Integration: Implementing voice-based interactions for a hands-free experience. Advanced Fraud Detection: Enhancing security to prevent fraudulent booking and refund requests.

Built With

• dfiy
• englsih
• llm
• openai
• python
• taiwanese