Airline data your teams can question in plain English, from Slack
FareIntelligence works with large volumes of airline fare and pricing data. We provided data and AI engineers who built a governed semantic layer over that data and a Databricks Genie space on top, so people can ask questions in natural language from Slack.
- Service
- Data for AI delivery · AI-ready data · Data & AI engineers
- Engagement
- Embedded data and AI engineers
How it fits together
- 01Airline fare & schedule data
- 02Curated tables in Unity Catalog
- 03Semantic layer
- 04Genie space
- 05Slack
The challenge
Airline data is wide and changes constantly: fares by route and travel date, booking classes, origin and destination pairs, carriers and competitor fare movements. The answers were in the data, but getting them meant waiting for someone to write SQL, and the same measure could come out differently depending on who wrote the query. FareIntelligence wanted business users to get reliable answers on their own, without losing control of definitions or access.
What we did
- Placed data and AI engineers inside the FareIntelligence team, working in their Databricks workspace and delivery process.
- Modelled the core airline entities (routes, O&D pairs, fares, fare classes, carriers, travel and booking dates) into curated, documented tables in Unity Catalog.
- Built a semantic layer on top: agreed metric definitions, business names and descriptions, and join paths, so a term like "lowest fare" means the same thing every time.
- Set up a Databricks Genie space over the semantic layer, with instructions and example SQL so it understands airline terminology and common question patterns.
- Connected Genie to Slack through the Genie Conversation API, so people ask questions where they already work and can see the query behind each answer.
- Kept it governed: the space only covers curated, approved tables, and Unity Catalog permissions decide what data is exposed.
Asked in Slack
- What was the lowest fare on DEL to BOM for travel next week, by carrier?
- How have competitor fares on our top ten routes moved over the last 14 days?
- Which routes had the biggest fare changes yesterday?
Illustrative examples of the questions teams ask.
Roles we provided
Data Engineers
Curated airline datasets and pipelines in Databricks
Analytics / Semantic Layer Engineer
Metric definitions, documentation and Genie space tuning
AI Engineer
Genie Conversation API and Slack integration
The outcome
Business users get answers on airline fare data without waiting for an analyst, and every answer comes from the same governed definitions. The semantic layer is also a reusable foundation for the next step: AI agents that work on the same trusted data.
Working on something similar?
Tell us about your data and what you want people or AI to do with it.