GermanyTravel and aviation2026 – ongoingEurowings
Semantic search on the booking website
Eurowings raised search recall from 50% to over 75% by replacing keyword search with semantic search
Problem
Eurowings, part of the Lufthansa Group, used keyword-based search on its website. Results were less relevant and slower than the airline wanted.
Approach
Eurowings replaced keyword search with semantic search built on Databricks vector search and Model Serving. It used MLflow to manage the model lifecycle.
Results
Takeaway
Better matching and faster responses came from the same change. The company gave no booking or revenue figures.
Categories
- Size
- Large enterprise
- Industry
- Travel and aviation
- Business function
- Marketing and sales
- Innovation type
- Product and service innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Databricks reports that Eurowings moved its website from keyword search to semantic search on the Databricks platform. The page gives a 109% lift in search relevance, recall rising from 50% to over 75% and latency falling from six seconds to under two.
Full details are in the original sources.
