Ahorra un 25 % (o incluso más) en tus costes de Kafka | Acepta el reto del ahorro con Kafka de Confluent

Featured Partner: iLink Digital

Revolutionizing Air Travel Support Operations with Generative AI

In the fast-paced world of air travel, delivering exceptional customer experiences is non-negotiable. To meet this demand for excellence, iLink Digital partnered with Confluent, a global leader in real-time data streaming and event-driven architecture. This collaboration has led to the development of a demo showcasing generative AI with data in motion for customer support operations within the airline industry.

Partner use case - Build with Confluent badge

Revolutionizing Air Travel Support Operations Using Confluent and Generative AI



Disclaimer: The name "ConFLYent Airlines" is used for demonstration & understanding purposes only.

Air travel has always been an exciting yet occasionally daunting experience for passengers. From booking flights and managing baggage to dealing with unexpected flight delays, passengers are often faced with numerous questions and concerns.

To enhance the passenger experience and streamline communication, we collaborated with Confluent to develop an intelligent chat application for ConFLYent Airlines. Our intelligent chat application builds the customer context by pulling the most up-to-date information and integrating it with LLMs (large language models) to provide human-like responses.

Challenges in the Airline Industry

The airline industry operates in a high-stakes environment, where every minute counts and customer satisfaction is paramount. However, it faces several critical challenges including:

Real-Time Data

Timely updates on flight information, gate changes, and policy updates are vital for passengers. Traditional systems struggle to deliver real-time data.

Scalability

Airlines must handle sudden surges in passenger queries during peak travel times, which can overwhelm traditional customer service channels.

Personalization

Passengers expect personalized services, such as meal plan changes or flight rescheduling. Traditional bots often lack the ability to provide this level of customization.

Data Integration

Integrating disparate data sources and AI models into a cohesive system can be complex and costly.

Build with Confluent

This use case leverages the following building blocks in Confluent Cloud.

Reference Architecture





A Closer Look at the ConFLYent Airlines Chat Application

Before we explore the scenarios, let’s understand the magic behind ConFLYent Airlines’ real-time chat application. At its core, this application is a perfect blend of cutting-edge technology and seamless architecture.

On the front end, our Real-time WebSocket-based Chat Application, hosted on Azure App Service, offers passengers a user-friendly interface for engaging with our services. It’s tightly integrated with the “Questions and Answers” Confluent topics, ensuring seamless communication between passengers and our AI-driven assistance.

Here’s how it works:

Data Sourcing

Various Airline data sources such as Flight operation systems, CRM systems, Check-in & boarding systems, and Flight security & policy systems act as data sources, supplying critical information about flights, passengers, gate information, weather-related information, and airline policies. This data is streamed into Confluent topics within the Confluent environment.

Consumer Group & LLM Integration

Enrichment of data takes place here. An Azure App Service communicates with a Language Model (LLM), enhancing the raw data. The enriched data is stored in the “Flight Data Embeddings” Confluent topic. Then the enriched data from the “Flight Data Embeddings” Confluent topic is retrieved and stored in Vector DB.

Chat Application

The Chat Application Backend Azure App Service orchestrates communication between the user interface and the data sources. It retrieves enriched data and interacts with the Language Model like Chat GPT for natural language processing.

Resources

Contact Us

Contact iLink Digital to learn more about this use case and get started.