Refactoring dashboards with Claude Code (without breaking prod)
Can AI safely update production dashboards?
In this Lightdash demo, Oliver shows how AI can refactor and improve BI dashboards. All without breaking production, using preview environments, Git diffs, and pull requests.
Instead of manually editing charts one by one, he uses AI to:
Add a 12-hour cycle time target to an engineering dashboard
Automatically introduce better visualizations (including a goal-based gauge)
Reorganize layout using dashboard best practices
Review every change as code
Open a pull request for team approval
All changes run inside a Lightdash preview environment, so production stays untouched until your team reviews and merges.
This is Agentic BI built for real engineering workflows.
No chatbot UI or risky production edits. Just dashboards as code, version control, and governed AI automation.
If you’re exploring Agentic BI, or modern semantic layer workflows, this demo shows what safe, production-ready AI looks like.
Ready to use try Agentic BI for your team?
🔗 Start your free trial → https://www.lightdash.com/start?utm_s...
📣 See how teams use Lightdash → https://www.lightdash.com/customer-st...
📖 Read the docs → https://docs.lightdash.com/introducti...
💬 Join the community → go.lightdash.com/community