How Scottish Water Made Its Capital Investment Data Conversational With Databricks Genie

Source

Source name: databricks_engineering
Source kind: blog
Authors / organization: Sourav Gulati, Poppy Harvey (Databricks)
Venue: Databricks Blog
Published: 2026-08-13
Collected: 2026-08-15
URL: https://www.databricks.com/blog/how-scottish-water-made-its-capital-investment-data-conversational-databricks-genie
PDF: n/a
DOI: n/a
Registry group: ai_infrastructure_and_engineering
Evidence role: engineering_practice / customer case study (vendor-published)
Language: en
Country / region: US (vendor); customer is UK (Scotland)

Plain-Language Brief

Databricks published a customer case study describing how Scottish Water built "SPARK," a natural-language interface over Databricks Genie and Microsoft Teams/Copilot, to let staff query capital investment project data in plain English instead of navigating multiple dashboards.

What the Document Says

The post states Scottish Water previously relied on fragmented reports and specialist staff to find project data, and built SPARK on governed Unity Catalog data with "metric views" standardizing measures and terminology. It reports time-savings estimates for typical lookups and for a hypothetical group of 100 users.

Key Concepts Mentioned

Databricks Genie (natural-language query layer)
Unity Catalog governed data / metric views
Microsoft Teams and Copilot integration

Claims Made by the Document

A lookup that "previously took around 8 clicks plus dashboard load time can now start with a single question in Teams."
For 100 users asking 3 questions weekly, the organization saves "roughly 10 to 25 hours per week, or roughly 520 to 1,300 hours per year."
The solution uses "governed Unity Catalog data" with standardized metric views.

Evidence Used

Customer testimonial / vendor case study; the hours-saved figures appear to be an extrapolated estimate rather than a measured before/after study, based on the retrieved text.

Methods / Approach

Case study narrative describing implementation of Genie plus a metric-view data layer integrated into Teams.

Limitations Stated by the Authors

None stated explicitly in the retrieved content.

Limitations Noted by the Agent

Agent interpretation — not from the source.

This is a vendor-published customer success story. The "520 to 1,300 hours per year" figure is described in the source as a projection based on assumed usage (100 users, 3 questions/week), not an audited measurement. Treat as a vendor case study, not independent evidence of productivity gains.

Open Questions

Whether Scottish Water independently published or verified the time-savings estimate, and what the accuracy/error rate of Genie's natural-language answers was on this dataset.

Notes for Future Q&A

May include interpretation. Labeled where it does.

Useful as an enterprise-deployment discovery signal for natural-language data interfaces; not a substitute for an independent evaluation of query accuracy.

Raw Abstract / Excerpt

"SPARK enables teams to ask practical business questions like risk assessments and project contractor lookups with immediate results."