AI for Sensitivity Analysis: Claude Tools for Scenario Modeling and What-If Analysis

Sensitivity Analysis and AI

Sensitivity analysis quantifies how changes in key assumptions affect model outputs — transforming a single-point estimate into a range of outcomes. Every financial model has critical drivers whose uncertainty determines whether a project is viable, a deal is attractive, or a forecast is credible. Claude with ClaudeFinLab builds one-way and two-way sensitivity tables, runs conceptual Monte Carlo frameworks, constructs tornado charts to rank driver importance, and helps finance teams communicate uncertainty to decision-makers.

One-Way Sensitivity Tables

• "One-way sensitivity table for DCF: base case enterprise value = $485M at WACC of 10.5% and terminal growth rate of 2.5%. Build a one-way sensitivity showing EV as WACC varies from 8.5% to 12.5% in 50bps steps: at 8.5% WACC → $612M; 9.0% → $581M; 9.5% → $553M; 10.0% → $527M; 10.5% → $485M (base); 11.0% → $465M; 11.5% → $444M; 12.0% → $426M; 12.5% → $409M. Key insight: a 200bps increase in WACC from base reduces EV by 15.5% ($485M → $409M). Conclusion: WACC is a highly sensitive driver; the buyer should hedge interest rate risk in the deal financing."
• "Revenue growth sensitivity for SaaS DCF: model assumes 28% NTM revenue growth. One-way table across growth scenarios from 10% to 40%: at 10% → EV $320M; 15% → $375M; 20% → $435M; 25% → $468M; 28% base → $485M; 30% → $505M; 35% → $548M; 40% → $598M. The EV is nearly linear with growth rate — a 10% reduction in growth from base (28%→18%) reduces value 12.4%. This tells management that retaining NTM growth above 20% is critical for maintaining valuation."

Two-Way Sensitivity Tables

• "Two-way sensitivity table (WACC × Terminal Growth): 5×5 matrix showing enterprise value. Rows: WACC from 9.0% to 11.0% in 50bps steps. Columns: terminal growth from 1.5% to 3.5% in 50bps steps. Build the full matrix: [9.0%, 1.5%]=$545M; [9.0%, 2.5%]=$581M; [9.0%, 3.5%]=$623M; [10.0%, 1.5%]=$499M; [10.0%, 2.5%]=$527M; [10.0%, 3.5%]=$560M; [11.0%, 1.5%]=$458M; [11.0%, 2.5%]=$480M; [11.0%, 3.5%]=$505M. Color code: green (EV > $550M), yellow ($450-550M), red (<$450M). The buyer's return threshold at $500M EV implies acceptable at WACC ≤10.5% and terminal growth ≥2.0%."
• "LBO two-way sensitivity (Entry Multiple × Revenue Growth): PE fund evaluating acquisition at $480M EV (10x EBITDA). Two-way table of 5-year IRR across: Entry EV/EBITDA (9x, 10x, 11x, 12x) vs. Revenue CAGR (5%, 7%, 9%, 11%, 13%). Key cells: [10x entry, 9% growth] = 22.4% IRR (above 20% hurdle); [11x entry, 7% growth] = 18.6% IRR (below hurdle); [9x entry, 7% growth] = 24.1% IRR (attractive). Table reveals: at current offer price (10x), need at least 9% growth to meet hurdle. Negotiate down to 9x if growth visibility is uncertain."

Tornado Charts and Driver Ranking

• "Tornado chart construction for project finance NPV: base case NPV = $42M. Test each key driver at high and low scenarios (±10% or ±1 standard deviation). Ranked results (largest to smallest swing): (1) Electricity price: NPV range $28M-$61M (swing $33M); (2) Construction cost: $31M-$56M (swing $25M); (3) Capacity factor: $35M-$54M (swing $19M); (4) Discount rate: $36M-$51M (swing $15M); (5) O&M costs: $38M-$48M (swing $10M); (6) Debt rate: $39M-$46M (swing $7M). Tornado chart conclusion: electricity price and construction cost are the two most critical risks — these deserve the most due diligence and hedging attention."
• "Revenue driver sensitivity decomposition: SaaS company revenue = Users × Conversion Rate × ARPU. Base case: 100K users × 3.2% conversion × $1,800 ARPU = $5.76M ARR. Sensitivity: users ±20%, conversion ±20%, ARPU ±20%. Impact on ARR: users swing $2.30M; conversion swing $2.30M; ARPU swing $2.30M. All three are equally sensitive (multiplicative model). Next layer: what drives user count? Paid acquisition ($240K CAC), organic (60% of new users). Paid acquisition efficiency: ±20% change in CAC conversion → ±12% change in total users → ±12% change in ARR = ±$0.69M."

Monte Carlo Simulation Framework

• "Monte Carlo setup for real estate underwriting: property has base case NOI of $4.2M and assumed cap rate of 5.5%. Five uncertain inputs: (1) Market rent growth: Normal(3%, 1.5% std); (2) Occupancy: Normal(94%, 3% std); (3) Expense ratio: Normal(38%, 4% std); (4) Exit cap rate: Normal(5.5%, 0.5% std); (5) Exit year: fixed at 5. Run 10,000 simulations. Output distribution of exit equity IRR: mean 16.8%, median 16.2%, std 4.8%, 10th percentile 10.2%, 90th percentile 23.4%. Probability of IRR > 15% (sponsor hurdle): 62%. Probability of loss (IRR < 0%): 3.2%. Key tail risk: exit cap rate expansion to 6.5%+ combined with occupancy below 88% = worst 5% of outcomes."
• "Monte Carlo for credit analysis: company has base EBITDA $48M, DSCR covenant 1.25x, total debt service $28M. Run 5,000 scenarios where EBITDA is drawn from Normal($48M, $8M std) reflecting business volatility. P(DSCR breach) = P(EBITDA < 1.25 × $28M) = P(EBITDA < $35M). From N($48M, $8M): z = (35-48)/8 = -1.625, P(breach) = 5.2%. Over 3 years with partial correlation between years (0.4 correlation): P(at least one breach in 3 years) = 14.7%. Bank should consider this in covenant headroom design — recommend 1.35x covenant (not 1.25x) to provide additional buffer."

Break-Even Analysis

• "Operating leverage and break-even: company has fixed costs $18M/year, variable costs 42% of revenue, current revenue $38M. Contribution margin = 58%. Break-even revenue = Fixed Costs / CM% = $18M / 58% = $31.0M. Current operating leverage (revenue / break-even revenue) = $38M / $31M = 1.23. This means: if revenue falls 18.4% from current level, company reaches break-even. At $28M revenue (26% decline), company loses money. In a recession scenario with 20% revenue decline to $30.4M, EBIT = $30.4M × 58% − $18M = ($0.37M) → breakeven barely missed. Compare to lower fixed cost model: if $6M of fixed costs converted to variable at 15% of revenue, new break-even = $12M/(43%) = $27.9M — much more resilient."
• "Financial covenant break-even: term loan has leverage covenant of 4.5x. Current EBITDA $48M, current debt $168M, leverage = 3.5x. What EBITDA decline triggers covenant breach? 4.5x × EBITDA = $168M → break-even EBITDA = $168M / 4.5x = $37.3M → requires EBITDA decline of 22.3% ($48M → $37.3M) before covenant breach. How much buffer does that represent? At current run rate ($48M EBITDA) with $6M/year growth: EBITDA declining 22% requires a fairly severe recession scenario. However, if EBITDA is on a downward trajectory (declining business), calculate time-to-covenant-breach at various EBITDA decay rates: at -10%/year, breach in Year 3.2."

Related Articles

• Financial Modeling Guide
• WACC Calculation AI
• Real Options AI