Organising travel is a huge challenge, and it can be made quicker, simpler and cheaper with the support of technology. Imagine searching for flights without the support of platforms comparing prices, travel times, baggage requirements. We open the first browser tab, the second, the third, the fourth... the thirteenth - each with an offer from a different flight provider. Eventually, platforms are created that compare a few selected airlines, the number of tabs opened decreases and search time is reduced by 50%. Now imagine a solution that integrates all these tools. With just one tab, one filter selection and one click, you can benefit from the best offer of all those available on the market.
Getaway - imagined vs reality
One tab, one filter selection and one click are all it takes to benefit from the best offer of all those available on the market. This is what the client comes to us with. The creation of a travel portal as a travel assistant meets the needs of the target group. There is no such solution on the market to date. The travel assistant is the gap that Getaway wants to fill. It might seem - an ideal situation.
The biggest challenge, however, turns out to be ignorance of the specifics of software development. This unawareness compounds the problem of finding the right Software House.
#1 Conceptual phase of integration with flight providers
Getaway starts with the first partner who cannot understand its needs. This project needs to start well before software development. The work starts at the grassroots in the form of a proof of concept. The aim is to integrate 4 platforms that enable flight searches:
- Tripstack offers cheap Ryanair and EasyJet flights.
- Duffel is the platform that is the most developed of the planned integrations with flight providers.
- Kiwi excels in terms of the best programming API and offers the most flexible search parameters.
- Amadeus is a global platform partner of Ryanair.
Each platform brings specific, unique value to the project. The differences between them create further challenges and problems that would be impossible to solve without specialist knowledge.
#2 Aligning data, parameters and learning to search
What distinguishes each of the platforms in integrating with flight providers is also the biggest challenge. They all take a different approach to search, introducing distinct parameters according to which they select the best solution for the user.
Prerequisites
✅ I plan to create a platform integrating 4 different flight supplier platforms
✅ Each platform has an API at a different level of sophistication and development
Description of the problem
The platforms differ in the area of collecting and converting data into search results for the user. The biggest challenge is to unify them and match filters without generating 4xx errors.
What to do?
I have the same problem. Did you manage to solve it?
I have searched on Google, ChatGPT, but without success. I have searched all available databases, including Discord.
Three solutions need to be put in place to help eliminate possible 4xx errors.
Firstly, the format of the data should be standardised - so that suppliers' data is available in the same form as in other platforms
Secondly, it is necessary to introduce support for the search parameters of interest to the user
Thirdly, in order to make the integration work better, parallel searches should be designed and the system should be taught to search multiple directions and terms using all suppliers
What if we do not serve a particular supplier at a particular time?
Yes, you need to take into account the possibility that the partner(s) can be switched off at any given time, and ensure support from dedicated partner teams through certification in case of problems.
#3 Reduce loading times and automate performance evaluation with AI/ML
The main objective is to reduce the time taken to find the best offer. Therefore, the ability to update found flights in the background should be planned as part of the integration of flight providers for Getaway.
The platform is designed to help the user choose the cheapest or fastest connection. In this case, we used AI/ML technology. With the help of GPT Chat, we have created a unique algorithm that assigns a score to the selected connections, allowing the user to select the optimal solution for the chosen destination and date.
#4 Relevance Getaway for baggage handling
Integration with flight providers requires building a system that can combine the unique mechanisms of the four platforms and create a single mechanism to cover all needs.
Thank you for all the answers, but I have another problem. Each provider has a unique approach to baggage handling. Which approach do you suggest?
All approaches need to be analysed and handled correctly through the Relevance mechanism. This will avoid providing incorrect information or errors for missing data.
And how did you manage to implement this solution?
Yes, we analysed each supplier, which allowed us to define the key problems to be solved:
- Kiwis offer baggage handling with their margin.
- Duffel and Tripstack do not return complete information for all lines, but is gradually introducing the service.
- Amadeus has complete data, but integration is difficult and complex.
Based on this information, we concluded that the best solution would be to create a 'Getaway Relevance', i.e. the ability to pre-define baggage information per line.
#5 Effect of work on Getaway project
- The system finds the best (in Getaway's opinion) flight among the offers available from the 4 partners mentioned.
- The tool is prepared to integrate more providers in the future.
- We minimised the risk of 'not finding' a flight for a customer.
- We have created a compliance system that updates flight prices in the background.
- We have added the ability to define new airports/airlines in the system as they come into the world.
- We have created a unique, dedicated flight search tool with the help of Amadeus, which allows you to search for flights in parallel within a specific date range and for more airports.
The lead time is five months.