Copilot for Finance

Course subjects

Prompting Effectively with Copilot

• Why vague prompts create misleading financial outputs

• The Goal-Context-Sources-Expectations prompt model

• Using constraints to control scope and assumptions

• Preventing fabricated numbers or implied certainty

• Refining a prompt once, then validating manually

Data Preparation and Initial Analysis

• Reviewing raw financial data

• Cleaning and structuring data (tables, headers, consistency)

• Identifying gaps, anomalies, and assumptions

• Preparing the workbook so Copilot can operate correctly

• Output: prepared Excel dataset with documented assumptions

Analysis and Financial Commentary

• Using Copilot in Excel to generate calculations, summaries, or trends

• Verifying formulas and results

• Drafting financial commentary that reflects validated findings

• Avoiding over-statement or unsupported conclusions

• Output: validated analysis with draft financial commentary

Financial Recommendations and Options

• Developing alternative financial scenarios or approaches

• Assessing trade-offs (cost, cash flow, risk, timing)

• Using Copilot to structure options, not decide them

• Output: financial recommendation outline

Executive Decision Support

• Presenting multiple financial options clearly

• Making assumptions, risks, and dependencies explicit

• Drafting executive-ready justification language

• Output: executive finance options summary

Researcher Agent: Scaling Finance Research After the Fundamentals

• Why Researcher is not a replacement for financial analysis

• When Researcher is appropriate (market conditions, benchmarking, regulatory scanning)

• How to translate manual finance research steps into a Researcher task

• Responding to Researcher clarification questions

• Reviewing and validating Researcher outputs before use