Practice — Mild · Assistants Rehearsed
Skills focus. Each AI feature, run once with its check performed out loud.
Ungraded. Start from your A11 model. Tenant-lane note: where a feature is unavailable, read its section and complete the written check instead.
Tasks
- Q&A: add the visual; type "total revenue by region in 2025" and read the field-readout. Then type "average revenue" and write down which of the two averages the parser chose. Add the synonym that would make "sales" resolve to [Total Revenue]; retest.
- Key influencers: target [Return Rate], explain by Category and Region. Read the result against what you know (the 2.44–3.82% band). Write the one-sentence finding — no "drives" allowed.
- Decomposition tree: [Total Revenue], splits for Region/Category/ Subcategory. Decompose Midwest → Furniture and read the leaf against your drillthrough page's figures.
- Forecast: 6 months on the monthly line; band on; write the assumption caption. Then look at where the forecast band sits for next February and ask: does the model "know" about the February trough? How?
Reflection
- The parser's choice for "average revenue" — was it the one a retail VP would have meant? What does that say about shipping Q&A unconfigured?
- The forecast reproduced the seasonal dip. Which part of YOUR work made that possible — the tool's intelligence, or something in the model?