AI-Driven NDC Search: A PoC for the Next Era of Flight Search

Location

Global

Industry

Services

Challenge

Traditional flight search engines expect users to manually fill out multiple fields and filters. While logical, this design assumes users not only already know which filters they need but also where to find and how to interact with them appropriately. Many of these filters are hidden in side panels or embedded in advanced menus, making the experience cumbersome for less experienced travelers, especially when working with more dynamic content such as the NDC.

At the same time, the growing use of AI-enabled search agents and assistants has changed how people search for information. Travelers are becoming more comfortable typing or speaking natural language requests rather than seeking out filters and filling out search parameters. As a result, users are beginning to type or ask questions like “Find me a round-trip flight from New York to Los Angeles next weekend” rather than modifying their search parameters manually. This shift in user behavior has redefined what convenience means in the booking experience.

Solution Description

Our PoC introduces a natural language input interface that builds on the functionality of our existing NDC Flight Search PoC by greeting users with an LLM-powered text box where they can simply describe their travel plans in plain language. The LLM interprets user inputs, maps them to specific search parameters, and returns matching flight options.

Once initial input is gathered, users can refine their search through the same natural language format, getting updated results. Standard search fields, such as dates, locations, or passenger numbers, remain available, but the PoC allows for a fully conversational experience from start to finish.

This application connects to Sabre’s Bargain Finder Max (BFM) API, which supports a broad set of filters. By mapping natural language inputs to different BFM parameters, the PoC enables users to access a deep level of customization without technical effort; no complex commands or search for specific filters, just direct, intuitive requests.

Users can also filter existing results, revisit previous searches from history, and easily remove or adjust filters without leaving the conversational flow.

Process Flow

The user opens the flight search interface and types a request “Flight from New York to Los Angeles”

The application interprets text and returns flight options that match the request
The user types in further specifications
- “Roundtrip, return August 8th, one adult, one child, business class”
The application updates the search with new specifications
The user decides to remove the "business class" filter, clicks the “x”, and refines the search
The application returns options with the removed filter
The user decides to revisit a previous search and clicks on one of the history items
The application restores results from the past search

Business Benefits

Natural language input simplifies the search experience by providing users with an intuitive way to translate their needs into actions
Continued support for natural language refinement removes the need to navigate complex filters or search parameters
Search history allows for easy navigation between previous and current searches
“Stackable” filters ensure refinements can be added in any desired configuration
Barrier to entry for less confident/experienced users is lowered
Significantly simplified interaction with dynamic NDC parameters and data
Straightforward way of engaging with filters which might be “hidden” or “unavailable” in regular flight search contexts
Improved flexibility and customization of flight offers

Technologies Used

Java 17
Python
Spring Boot 3
Lombok
Mongo (Embedded Mongo)
Thymeleaf
OpenAPI 3
OAuth2
Docker
Docker Compose
Maven
AWS (ECR, ECS (Fargate), DocumentDB, CloudWatch, Bedrock)
Spring AI

Learn More

Interested in learning more about DataArt’s Policy-Compliant Chatbot and how it can enhance your travel platform? We would love to hear from you!

Contact our Travel Agency Solutions team to discuss how these technologies can streamline your business operations, assist your agents, integrate seamlessly with your NDC ecosystem, and more.

Please provide your contact details, and we will get back to you promptly.

First name*

Last name*

Email*

Job title

Company*

Phone

What would you like to discuss?*

How did you learn about DataArt?

What is your role?

I agree to receive DataArt's monthly marketing newsletter.

I have read and agree to the Privacy Policy*