Weekly practitioner insights — tools, soft skills, and craft.
Claude can construct three-statement models, DCF valuations, and M&A analyses by generating Python code that builds interactive Excel files with embedded formulas and real-time sensitivity tables. The approach includes data integration through Model Context Protocol to verify information across sources and create audit trails linking claims to original sources.
Python in Excel embeds code directly into cells for FP&A workflows including data cleaning with pandas (handling missing values, unpivoting), forecasting with scenario modeling, and month-end closing automation paired with generative AI for variance analysis. Workflows pull from ERP/Excel, apply transformation rules, and export formatted reports to Power BI.
The most effective pushback on unrealistic budgets from business partners uses a shared 'driver dictionary' of 8–15 core performance drivers (volume, pricing, headcount, utilization) to shift conversations from subjective targets to objective assumptions. Pair this with integrated FP&A models and rolling forecasts to show cascading financial consequences of partner assumptions on revenue, cash flow, and working capital.
Effective board forecasts require three scenarios (base, optimistic, pessimistic) grounded in credible data sources (historical trends, run-rate analysis, customer conversations, 2025 results review), transparent assumptions on macroeconomic factors and operational risks, and stress-tested cash flow plans with specific course-correction actions for downside cases. Present as a dynamic 12-month rolling model updated quarterly.