Quality · Govern
Analyze Workbook Quality
Run a quality report on one workbook, compare PROD against UAT, or QA a whole project.
Platform guarantee
Nothing changes until a person says so.
The AI reads, explains and proposes. Anything it suggests arrives as a card that routes into the same approval gate a person would use, with the same signature and dual-control rules. Publishing always creates a new copy and never overwrites your work.
In the product
Quality, as it actually looks.
The register
What Quality does for you.
Every line below was read out of the product’s own source on 14 August 2026.
Twenty-four semantic difference kinds, compared model to model rather than as text, because Tableau rewrites the file on every save.
Three differences always fail, including a required calculation quietly swapped for a similarly named direct field.
Live-data verification compares the numbers actually served, plus a perceptual image hash, catching extract staleness that XML cannot show.
Promotion is signature-gated. A failure is never dismissible, and every waiver is recorded.
Specification
- Version3.10.0
Shipping at the 24 August 2026 research pass.
- Packageboreon-tabquality
The repository this behaviour lives in. Every claim above has somewhere to be checked.
- Model useOptional, on your own key
Switched on by you, running on the provider key you supply.
Assurance
The AI is yours to switch on.
Where Quality uses AI, it runs on the provider key you supply. Your prompts go to the provider you picked, and nowhere else.
The model drafts. Anything that would change your Tableau content arrives as a card that routes into the same approval gate a person would use, with the same signature and dual-control rules you already run. Your team keeps the pen.
Govern
Prove what changed, who can reach it, and what it exposes.
Quality sits with these, and one login opens all of them with your Tableau permissions.
See Quality in your own Tableau Environment.
A month on your own content is a better test than any description of it.