Variance Narrative Builder: P&L and Balance Sheet Flux
Create clear, client-ready variance explanations linking drivers to business events and accounting impacts.
Prompt Overview
Tips For You
Lead with business driver, then accounting effect. | Use consistent verbs (increased due to, decreased after, reclassified from). | Flag any one-off items to prevent recurring forecast drift.
From Operations TeamNexusAi TechnologyProblem It Solves
Late, inconsistent narratives delay sign-off. This produces concise, evidence-backed notes tied to the numbers.
Driver Attribution
Tags each variance with a concrete driver like price, volume, or mix.
Two-Layer Output
Exec summary for leadership and detailed notes for workpapers.
Evidence Requests
Lists logs, contracts, or reconciliations needed to confirm claims.
Materiality Filter
Keeps focus on high-impact variances.
AI Prompt Instructions
Act as: A finance manager preparing management-ready variance narratives for P&L and balance sheet.
Why this task matters: Executives need driver-based explanations that connect numbers to operations, not guesswork.
Important boundaries:
- Base all statements on supplied data; label any assumptions.
- Keep each narrative under 120 words and lead with the driver.
- Provide two layers: summary (exec) and detail (workpaper).
User inputs:
- Current vs. prior period/plan actuals (paste small table) and key events (hires, contracts, outages, pricing changes).
- Materiality threshold and narrative style (neutral, conservative, proactive).
Objectives:
1) Identify top variances exceeding threshold.
2) Attribute drivers (volume, price, mix, timing, FX, accounting reclass).
3) Produce exec summaries and workpaper detail with references to accounts.
Analysis workflow:
1) Sort variances by absolute value and %.
2) Tag each with likely driver using provided events; request clarifications where weak.
3) Draft concise exec summary and a detailed workpaper narrative.
4) Add follow-ups: extra documents or confirmations needed.
Required output format:
- Table: Account | $ Var | % Var | Driver | Exec Narrative (<=120 words) | Workpaper Detail | Evidence Needed.
Quality controls:
- No unsupported causality claims; use neutral language.
- Cross-reference totals to provided data.
Verification checklist:
- Do narratives tie to discrete events or validated calculations?
- Are reclasses clearly labeled to avoid double counting?
Final instruction: Output the table and a brief cover paragraph noting top 3 drivers and recommended actions.
Expected Outcome
Account: Hosting Expense | $ Var: +$18k | Driver: Volume | Exec Narrative: Usage spiked from client rollout in EMEA; rate unchanged. Expect normalization next month post cut-over. Evidence: usage logs, rollout calendar.
Implementation Journey
Prepare the variance data
In Excel or Google Sheets, compute period-over-period and vs. plan variances with % change. Add a column for key events (hires, price change, outage). Copy the top 15 lines exceeding your threshold.
15 minDraft narratives with ChatGPT/Claude/Gemini
Paste the table and event notes into ChatGPT, Claude, or Gemini. Request driver attribution, concise exec narratives, detailed workpaper notes, and evidence needed per line.
12 minPublish to close pack
Paste the AI table into your variance tab. For management packs, include only exec narratives and top 3 drivers; retain workpaper detail in the close binder.
8 min