Session

Sponsored by: Thoughtspot | Creating Data Conversations: How Baker Hughes shifted to Real-Time AI-Driven Insights in Two Years

Overview

ExperienceIn Person
TrackAnalytics & BI
IndustryEnergy & Utilities
TechnologiesAI/BI
Skill LevelBeginner
DOWNLOAD SESSION SLIDES

As a leader in energy technology, Baker Hughes’ Data & AI Office transformed from a traditional report factory into a strategic platform powerhouse in just two years. By uniting a Databricks lakehouse with ThoughtSpot’s agentic analytics, they moved beyond legacy dashboard sprawl to a governed, AI-native foundation. This shift transformed enterprise data into trusted business context, ensuring every AI agent returns actionable insights that leaders can trust instantly. Awarded the 2026 Databricks Data Intelligence Industry Leader, Baker Hughes replaced fragmented workflows with a culture of curiosity. Using ThoughtSpot Spotter and Databricks Genie, global teams in Finance and Manufacturing now maintain a 1:1 relationship with their data. By enabling expertise hubs to talk directly with the Gold data layer, they have automated time-to-insight to drive operational resilience. Join ThoughtSpot and Baker Hughes to hear their transformation story, key learnings, and wins along the way.

Session Speakers

Jennifer Lowell

/Senior Data Architect
Baker Hughes

Paul Thompson

/Delivery Leader - Data & AI
Baker Hughes

Full Summary

How Baker Hughes rebuilt its data foundation to unlock trusted AI insights

Baker Hughes spent two years moving from fragmented, spreadsheet-driven reporting to a governed, automated data foundation that business teams can trust. The journey paired disciplined engineering with a modern lakehouse on the Databricks Platform, replacing offline snapshots with live, directly sourced views and enabling conversational analytics that surface real process issues.

FAQ


Executives had years of exposure to Excel-scrubbed, hand-filtered reports that masked operational reality. When zero-touch data from source systems replaced those artifacts, numbers often differed from previous, manually adjusted figures, which triggered skepticism until validation closed the gap.

Agile backlogs, clear acceptance criteria, and sprint cadences aligned to each business unit kept priorities focused. A single path to production with Git, CI/CD, automated tests, change data capture, and schema evolution reduced breakage and rework so engineers could deliver faster.

Genie aligns with engineering-oriented users who are comfortable with SQL and explicit instructions. Spotter is tuned for business users, using conversation to capture context, request clarifications, and refine definitions without requiring code.

Freeform exploration surfaced issues like prepaid customers with open shipments and shipments not being properly closed in source systems. Those gaps skewed KPIs and were rooted in process lapses rather than data pipeline errors.

Business teams receive training on standards and tooling, then develop on the shared stack. The central team reviews and approves pull requests, ensuring consistency and governance while enabling faster, domain-led delivery.