Fix Weak Scenario Planning in Finance with ChatGPT-Driven AI

Use ChatGPT to Design Structured Scenario Frameworks Before You Touch Excel

Most finance teams jump straight into spreadsheets when asked for scenarios. Instead, start in ChatGPT to design the scenario framework: which drivers to vary, what ranges to use, and how to label each scenario. This avoids ad-hoc, one-off views and creates a reusable structure for planning.

Provide ChatGPT with context about your business model, key value drivers, and current base case. Ask it to propose a structured set of scenarios including clear names, driver changes, and qualitative expectations. You can then feed these driver sets into your existing models for calculation.

Prompt example:
You are a senior FP&A analyst supporting financial planning.
Our business model (summary):
- Revenue drivers: number of active customers, ARPU, churn rate
- Cost drivers: COGS % of revenue, logistics cost per shipment, headcount costs
- Current base case for next 12 months: [insert summary]

Task:
1. Propose 8-10 distinct financial scenarios for the next 12 months.
2. For each scenario, specify:
   - Scenario name (short, descriptive)
   - Key driver changes vs base case (with direction and rough magnitude)
   - Narrative description of what is happening in the business
3. Make sure the set covers:
   - Demand shocks
   - Price changes
   - Supply or cost disruptions
   - Strategic choices (e.g. aggressive marketing, cost cutting)

Outcome: You get a clear scenario catalogue that can be implemented systematically in your planning models instead of improvising each time.

Automate Assumption Set Generation and Documentation

Once you know which scenarios you want, the next bottleneck is turning them into consistent, documented assumption sets. Use ChatGPT as an assumption engine that produces structured tables or JSON-like outputs you can paste into Excel, your FP&A tool, or a database.

Feed ChatGPT your base case assumptions and scenario definitions, and ask it to generate modified assumptions, including explicit justifications you can later audit.

Prompt example:
You are helping with driver-based financial planning.
Base case assumptions (next 12 months):
- Volume growth: 5% YoY
- Average selling price (ASP): +1% YoY
- COGS: 62% of revenue
- Logistics cost: 4 EUR per shipment
- Marketing spend: 10% of revenue

Scenario definition: "Sudden input cost spike"
- Raw material prices increase sharply in Q2 and stay elevated
- Management delays price increases to protect demand

Task:
1. Propose a consistent set of modified assumptions for this scenario.
2. Output them in a structured table with columns:
   - Driver
   - New value
   - Timing (from which month/quarter)
   - Rationale (1 sentence)
3. Ensure the changes are realistic and internally consistent.

Expected outcome: Faster creation of well-documented assumption sets, with less manual retyping, and a clear audit trail of why each assumption changed.

Let ChatGPT Turn Model Outputs into Executive-Ready Scenario Briefs

After running scenarios in your models, finance spends a lot of time turning results into slides and narratives. Use ChatGPT to draft scenario summaries that highlight key impacts, risks, and recommended actions for management, based on exported tables or key figures.

Export your scenario outputs (e.g. revenue, EBITDA, cash, key KPIs per scenario) into a clean text or CSV format and feed them to ChatGPT with clear instructions on target audience and tone.

Prompt example:
You are preparing an executive briefing for the CFO.
Below is a summary table with key financial outputs for 4 scenarios
for FY2025 (Base, Demand Shock, Cost Spike, Aggressive Growth):
[Paste simplified table or CSV]

Task:
1. Write a concise 1-2 page narrative comparing the scenarios.
2. For each scenario, summarize:
   - Headline story in 2-3 sentences
   - Impact on revenue, EBITDA, and cash vs base case
   - Key risks and operational implications
3. End with 3-5 clear decision points the leadership team should discuss.

Expected outcome: Scenario packs that are ready for leadership review in hours, not days, while finance can focus on interpreting and challenging the story instead of drafting it from scratch.

Use ChatGPT to Stress-Test and Challenge Key Assumptions

ChatGPT can also play the role of a "critical reviewer" of your assumptions. Instead of relying only on internal debate, ask it to identify where your forecast assumptions might be optimistic, internally inconsistent, or blind to external risks.

Share your core planning assumptions and ask ChatGPT to challenge them from different angles (macroeconomic, industry, operational). This is especially useful for stress tests and downside scenarios.

Prompt example:
You are an independent risk and scenario planning expert.
Here are our core assumptions for the next 24 months:
[Paste key volume, price, cost, and capex assumptions]

Task:
1. Identify 10 specific risks or vulnerabilities in these assumptions.
2. For each, explain why it might be unrealistic or fragile.
3. Suggest 1-2 alternative assumption values or ranges we should
   test in downside scenarios.
4. Propose 3 additional stress scenarios we are currently missing.

Expected outcome: A richer set of stress-test cases and a more robust understanding of where your plan is most exposed, without weeks of manual what-if analysis.

Create Repeatable Scenario-Planning Playbooks and Prompt Libraries

To move beyond ad-hoc use, formalise your ChatGPT scenario planning workflows into internal playbooks. Capture the best prompts for scenario design, assumption generation, summarisation, and stress-testing. Store them in a shared knowledge base and update them after each planning cycle.

Define simple usage patterns: which prompts are used in monthly reviews, quarterly forecasts, and annual planning; who is responsible; and how outputs are archived. Over time, you build a reusable "AI assistant" that is tailored to your business model and planning cadence, rather than starting from scratch every time.

Integrate Lightly with Your Data and Tools to Avoid Copy-Paste Overload

While a full system integration may come later, you can already reduce friction by defining a standard way to export and feed data into ChatGPT (or a ChatGPT-based internal assistant). For example, agree that all scenario result tables follow the same column structure and naming conventions, so prompts can reference them reliably.

Work with IT and your FP&A platform owner to explore simple automations: generating CSV exports for all scenarios, or calling ChatGPT via API from a lightweight internal tool. This is where Reruption’s AI engineering and PoC experience can help you move from manual pasting to a pragmatic, secure integration without a multi-year IT project.

Across these practices, finance teams typically see outcomes such as: a 30–50% reduction in time to prepare scenario packs, 2–3x more scenarios considered per decision, and faster alignment between finance and business stakeholders. These are realistic, measurable improvements that compound over each planning cycle.