name: Build a variance driver bridge tagline: Explain movements across actuals, budget, and forecast with source-backed drivers. summary: Give ChatGPT actuals, budget, forecasts, KPI data, thresholds, and owner notes, then ask it to rank the drivers behind material movements and produce an editable bridge with reconciliations, questions, and source citations. skills:
-
token: $spreadsheets description: Reconcile finance inputs, calculate material movements, and create an editable driver bridge with formulas and checks.
-
token: $documents description: Package source-backed driver commentary and owner questions into a reviewable memo when needed. bestFor:
-
Forecast-to-actual, budget-to-actual, and forecast-to-forecast reviews.
-
Analyses spanning revenue, margin, operating expense, cash, or balance-sheet drivers.
-
Teams that need ranked drivers, reconciliations, and owner questions in one deliverable. starterPrompt: title: Build the variance bridge body: >- Use $spreadsheets to explain the [period] movement between [actuals, budget, forecast, or prior forecast].
Use the attached close workbook, budget file, prior forecast, KPI dashboard, operating-expense tracker, cash view, and finance-owner notes. Build an editable variance bridge across the relevant revenue, margin, operating-expense, EBITDA, free-cash-flow, and balance-sheet lines.
Rank material drivers by impact, reconcile source breaks, draft owner questions, and cite the workbook tab, dashboard, tracker, or note behind each driver. Flag movements that are below [threshold] separately. Do not write a definitive explanation when the sources only support a hypothesis. suggestedEffort: medium relatedLinks:
-
label: Agent skills url: /codex/build-skills
Introduction
A variance bridge connects a reported movement to the operational drivers behind it. ChatGPT can reconcile actuals, budget, forecasts, KPI data, and owner notes, then rank the drivers and separate supported explanations from questions that still need an owner.
Choose the comparison
Define the periods and versions you want to compare, such as forecast to actual, budget to actual, or current forecast to prior forecast. Provide the materiality threshold, sign conventions, and the finance lines that matter for this review.
- Attach the actuals, budget, forecasts, KPI data, and owner notes.
- Define the comparison, materiality threshold, and required finance lines.
- Run the starter prompt and ask for an editable bridge workbook.
- Review source breaks, sign conventions, and driver rankings.
- Continue in the same task to resolve owner questions and prepare commentary.
Challenge the explanations
Driver commentary should follow the evidence. Ask ChatGPT to label hypotheses clearly when the source files show a movement but do not establish its cause.