KLM, Royal Dutch Airlines

How much time do your employees spend in the office looking for information every day? This amount of time can be shortened with an optimal search experience!

For Koninklijke Luchtvaart Maatschappij (KLM) we help 30,000 employees find the right information. To facilitate this we have linked various information sources and developed an internal search engine.

KLM is the Dutch national airline and one of the ten largest employers in the Netherlands. Employees were frustrated that information was difficult to find. With that in mind, during a hackathon, a Google-like search engine was chosen for KLM as the best solution. Subsequently, a project called MySearch was set up to develop the proof-of concept into an enterprise search service that allows employees to search for information faster, better and without frustration.

The search service is divided into four parts:

    • a standalone frontend
    • a search API
    • an admin tool
    • an Elasticsearch cluster

LuminisĀ  mainly worked on the last three. The search API handles queries and returns results. This API is also available for other applications within KLM. Within a search query, users can indicate a context in which the search results should fall, such as language and type of content. The department in which they work, which is derived from their logged in account, can also be used to rank more relevant content higher. The admin tool is used to retrieve content from various sources within KLM and index it in Elasticsearch to make it searchable.

In addition, several factors improving relevance can be managed, such as synonyms, compound words and query parameters. By using analytics , the search and click behavior of users is examined and an attempt is made to offer relevant search results. This can be done, for example, by making a certain field in which the search terms appear more important, or by making older content of a certain type, such as news, less relevant.


Byron Voorbach

Search & Data Engineer
+31 6 46 25 18 99

Roberto van der Linden

Software Developer

Jettro Coenradie

Fellow / Search Evangelist

Sander Pagie


Related cases