Cleartrip connects travelers in India to flights through a network of global distribution systems, such as travel agents and booking sites, as well as direct connections from airlines themselves. Yet several vital data points were found to be missing or unsupported across airline suppliers, which led to an inconsistent experience for customers using the platform. Cleartrip brought in Thoughtworks to rebuild the booking platform on a foundation that could seamlessly support their nine existing airline suppliers and offer a scalable solution for future growth.
Challenge: How do you scale airline coverage on a world-class booking platform without destabilizing the experience?
Cleartrip holds a vast quantity of information fed from nine airline suppliers on its travel technology platform, each of which incorporates data from potentially dozens of airlines, which thousands of customers rely on every day to make their flight bookings. This makes any gaps in that information, due to incompatibility issues between the airline suppliers and the Cleartrip platform, a customer-facing problem. This could be anything from a fare that can't be displayed, to a seat that can't be selected, or a cancellation that takes days to process because it has to be done manually.
Closing these gaps across nine airline suppliers, each with its own contract, datasets, and quirks, meant more than patching individual bugs. Cleartrip needed to reverse-engineer the legacy connections with its current suppliers to bring every supplier's data up to a consistent standard, while building a platform that could scale to bring in more airline data down the line.
Solution: Redesigning a powerful airline booking platform with modular systems and microservices
Thoughtworks started by developing a plan of action. Rather than the standard legacy system modernization of lifting old code into new containers, the team documented design specs for each integration before writing implementation code, an approach that minimized technical difficulties throughout the process. Booking logic for the platform was broken into independent, reusable flows, including search, hold, book, cancel and ancillary handling, all as part of a microservices migration that replaced a single block of code tied to one of the nine airline suppliers.
Two of these suppliers, however, needed deeper work. One had almost no documentation and required its own microservices migration to current API versions with no prior blueprint to follow. The other supplier, which alone covers more than 18 individual airlines, brought a long list of quirks: separate credentials for some of their carriers, which interfered with compatibility on the booking platform and low-cost carriers with access only to the live production system and no sandbox environment to test in.
The team leveraged AI with Cursor, helping them parse supplier API contracts running past 100,000 lines, cutting typical two-to-three-day troubleshooting down to about half a day. A custom Python proxy, another product of that AI-assisted software development, allowed engineers to route local requests through the client's Kubernetes environment. The runtime also moved from Java 8 to Java 21, with shared logic pulled into reusable JAR components so the next integration would not start from zero.
Thoughtworks helped us redefine our platform to seamlessly handle nine different airline supplier integrations who all worked from different rulebooks, and understood the systems at play well enough to help us build something better.
Outcome: Faster searches, complete data coverage, new airline suppliers
Leveraging AI-assisted software development to work fast, we brought measurable gains in speed and reliability as part of modernizing the travel technology platform.
Round trip search latency p95 improved from 10 seconds to 8 seconds, with overall average latency dropping from 8.5 to 7.5 seconds.
Baggage coverage reached approximately 99.5%, with a failure rate of roughly 0.4% at booking.
Fare rules coverage rose from 80% to 90% overall across 18 airlines on the Verteil connection, with one airline improving from 2% to 93%.
Cancellation and refund automation improved by roughly 70%, reducing reliance on the manual operations queue.
Two low-cost carriers were brought onto the platform, strengthening Cleartrip's footprint in Middle Eastern markets.
Paid seat and meal selection launched across major international routes.
Star Air, added as a new direct supplier, recovered to approximately 350 daily segments with a full end-to-end booking and auto-cancellation flow live.
Not only did Thoughtworks fix individual integrations, but they helped us rethink how we can approach every new supplier going forward as we continue to build our platform to scale
Looking ahead: More airlines, international currency support, greater reach
Cleartrip's roadmap continues with the same efficiency as the legacy system modernization and microservices migration, with plans to scale toward over 200 airlines for international travel as well as adding full support for Indian Rupee currency transactions to support a large portion of their currency base without relying on conversions. The reusable components of the platform built through AI-assisted software development during this engagement give Cleartrip the runway needed for each new airline supplier integration, allowing them to scale well into the future with nothing but clear skies ahead.
