← All use cases

Prepare a business review

Combine recurring KPI, close, and forecast data into a performance story with cited sources.

Supply KPI dashboards, close workbooks, definitions of the metrics, forecast updates, earlier reviews, and notes from owners. ZeroTwo writes a narrative ready for review that explains material shifts, variances, anomalies, risks, data-quality problems, and next steps.

Intermediate
30m

Related links

Skills & Plugins

  • Documents
    Writes the narrative with an executive summary and source notes ($documents).
  • Spreadsheets
    Checks close workbooks and traces key numbers to their source tabs ($spreadsheets).
  • Google Drive
    Reads approved workbooks, prior reviews, and owner notes from the files you name.
  • Slack
    Reads approved finance-close threads when they belong in the review.

Starter prompt

Write the September monthly business review as an editable document for our leadership team. I've attached the KPI dashboard export, the close workbook, last month's review, and owner notes. Cover the headline summary, what moved in each KPI and why, variances against forecast, anomalies, risks, and follow-ups grouped by owner. Cite the file behind every number, and list any gaps in a separate section instead of filling them with estimates.

Build from the recurring review record

A review explains what changed, why, and what needs deciding. ZeroTwo drafts from dashboards, workbooks, forecasts, and notes while keeping sources visible.

Prepare the review sources

Name exact files rather than searching the workspace broadly.

  1. Set period, audience, materiality.
  2. Attach dashboard, definitions, workbooks, forecast, prior review, notes.
  3. Trace material numbers first.
  4. Run the prompt.
  5. Confirm links and owner actions.

Review the narrative

Good drafts separate supported conclusions from open questions and make follow-ups assignable. Ask for a comparison with the prior review: new variances, resolved or worsened anomalies, open risks, changed definitions, unsupported claims.

Related use cases