AI chatbots like ChatGPT, Microsoft Copilot and Google Gemini are only as effective as the prompts they are given. With the right prompts, they can summarize reports, benchmark peers, draft clear documents and even test executive-level scenarios — all in seconds or minutes rather than hours or days.

Getting bogged down in manual tasks is no one’s preference, and with the help of AI, finance professionals can now focus on what they do best: interpreting results, telling the story behind the numbers and driving strategic decisions.

Keep reading for examples of prompts that finance professionals on the AFP FP&A Advisory Councils use in their daily work. Use the links below to jump ahead to a specific section.

• Conduct research
• Act as a thought partner
• Act as a virtual assistant
• Provide technical assistance

A word of caution before proceeding: Responses generated by AI chatbots vary in quality, so it’s important to always verify AI responses. AFP cannot guarantee the accuracy of AI responses, nor do we recommend any particular tool or technology.

Conduct Research

Research is the foundation from which corporate finance operates. Doing it the “traditional” way can take hours or days, whereas AI can do the same thing in a much shorter period of time by gathering, summarizing and contextualizing information from multiple sources in mere seconds.

Prompt: Compare and contrast reports from different sources and companies for benchmarking purposes.

Expected Result: A side-by-side view of how different companies or industries perform across key metrics, such as revenue growth, margins or cost ratios, so you can quickly spot the performance gaps, best practices and outliers that inform your benchmarks and forecasts.

From Raymond Cheung, Vice President, Corporate Finance and Strategy, SML Group Limited

Prompt: Give me a summary of the most recent 10-Q filing or investor research report for the company, call out revenue, cost of goods sold and EBITDA, and provide the year-over-year comparison for that period YTD. Provide any insights into the business segment or anything mentioned in the Management Discussion & Analysis section. Write this in bullet-point format.

Expected Result: In bullet-point format, you receive a snapshot of the company’s financial performance — key figures, trends and management commentary. You can quickly surmise what’s driving results, how performance compares year over year, and where risks or opportunities may be emerging.

From Lisa Bartko, Senior Finance Manager, Uline

Prompt: Act as an FP&A mentor. Drill down into the topic of [insert FP&A topic, e.g., driver-based forecasting]. Explain the concept step by step, include practical examples and sample calculations, and show how it connects to decision-making and planning. End by summarizing how these insights could enhance analytical strategy or forecasting accuracy.

Expected Result: A mini-lesson that explains the concept, walks you through real numbers and sample calculations, and ties each step to planning and decision-making, so you can easily apply it, avoid common pitfalls, and bolster your analytical strategy and forecast accuracy.

From Sultan Mujallid, FP&A Director

Act as a Thought Partner

One of the primary ways people across industries are using AI is as a sounding board. For finance professionals, this can take the form of brainstorming, anticipating questions and preparing for discussions with colleagues across the business.

Prompt: As a VP of [insert department], what questions would you ask of this analysis [attached or pasted], and what would be most important to you?

Expected Result: A good sense of how senior leaders might interpret your analysis — what questions they’d raise, which metrics they’d prioritize and what concerns they’d flag — so you can anticipate feedback, refine the narrative and tailor your presentation to their priorities.

From Tyler Vonderheide, Senior Manager, FP&A Systems, Southwest Airlines

Prompt: Read through my summary of results [attach or paste]. What am I missing? What else should I look at more closely?

Expected Result:An objective review that is able to point out blind spots, missing context or overlooked drivers, thereby helping to refine your analysis, reinforce your conclusions and ensure the findings stand up to scrutiny.

From Jesse Todd, Director, Finance Transformation, Microsoft

Prompt: What is the best way to explain why this matters to this [e.g., marketing, senior leadership] audience?

Expected Result: An audience-specific message that clearly communicates financial insights in terms relevant to a specific audience. This goes a long way toward helping you explain the “so what” in a way that resonates with non-finance stakeholders and drives engagement or action.

From Cassie McCombs, finance leader

Act as a Virtual Assistant

AI can take on the tasks of a virtual assistant, helping you rewrite, refine and reframe your work for maximum clarity and impact. From tightening a report to translating technical findings into executive-ready language, these prompts turn rough drafts and dense notes into polished, professional outputs that save time and elevate your communication.

Prompt: Rephrase the following email to make it clear, effective and appropriate to the situation.

When giving AI this prompt, be sure to include the following information:

• Purpose – what the sender is trying to achieve (e.g., request, persuade, thank, follow up).
• Audience – who the recipient is and their relationship to the sender.
• Tone – desired voice and style (e.g., confident, warm, formal, concise).
• Context – any background shaping the message (e.g., delayed response, ongoing project).
• Constraints – word limits or key facts to keep.

Rewrite the email to fit these goals while keeping the sender’s intent and natural voice. Provide a short note explaining the tone and key improvements.

Email: [paste here]

Details: [purpose, audience, tone, context, constraints]

Expected Result: A polished, well-structured email that strikes the right tone for your audience. The AI’s note explains what changed and why, helping you understand how to fine-tune future messages.

From Raymond Cheung, Vice President, Corporate Finance and Strategy, SML Group Limited

Prompt: Provide three versions of the attached [something I've written] and improve [e.g., clarity, tone and flow].

Expected Result: Multiple refined drafts to choose from, each offering a different take on clarity, tone and flow. You decide which one is the best fit for your audience and get to see how certain edits can make your writing more effective.

From Cassie McCombs, finance leader

Prompt: Translate the attached report into a one-page summary for executives using the language from the attached sample report [attach sample report, too].

Expected Result: A one-page executive summary that captures the report’s main takeaways using the right tone, structure and language. It delivers the “so what” up front, so leadership can quickly grasp key insights without reading the full report.

From Kavin Soni, Senior Financial Analyst, Google

Prompt: Review the attached meeting summary and surface the key points that connect to my ongoing workstreams [assuming your tool remembers history].

Expected Result: A recap that focuses on the points most relevant to your active projects. This allows you to easily see the decisions, risks or next steps that affect your workstreams and act on them without having to read the full meeting summary.

From Kavin Soni, Senior Financial Analyst, Google

Prompt: Act as my junior analyst and build a quick summary/model/explanation for …

Expected Result: A fast, first draft of the analysis, complete with summarized data, key assumptions, and a simple model or explanation you can refine.

From Majid Darvishan, Co-Director of FP&A Workshop, Indiana University

Provide Technical Assistance

With the right prompt, AI can handle technical tasks in minutes that would take humans hours to produce. AI can automate the setup, complete formatting and even run the initial analysis, providing finance teams with a fast, accurate starting point for deeper review and refinement.

Prompt: Generate a monthly CFO-ready financial one-pager. Include visualizations and summaries for revenue trends, operating margin, cash runway and budget vs. actuals. Make it suitable for board reporting.

Expected Result: A board- and executive-ready financial snapshot that highlights key metrics and visuals so leadership can quickly grasp the organization’s financial health and trajectory at a glance.

From Cassie Wang, CFO, Head of Finance, Lightship Security Inc.

Prompt: Using Copilot and Python in Excel, build me a 12-month forecast for this product line, taking seasonality into account.

Expected Result: A ready-to-use Excel forecast with seasonality built in, complete with assumptions, documented logic, charts and a 12-month projection you can update or change by adjusting the drivers or plugging in new actuals.

From Jesse Todd, Director, Finance Transformation, Microsoft

Prompt: Act as an IFRS expert. Review the following analysis or financial case and verify which IFRS standard(s) apply. Explain whether the case meets or does not meet the conditions of the relevant standard(s), citing key clauses or definitions where appropriate. Identify any mismatches or compliance issues and suggest corrections if needed. [attach or paste FP&A scenario or summary]

Expected Result: A structured first pass at identifying which IFRS standards might apply, with clear references to clauses and definitions. AI can provide a strong starting point for further review, but because it can misinterpret context or hallucinate citations, you should always verify the references and conclusions against the official standards before relying on them.

From Sultan Mujallid, FP&A Director

Prompt: Create a detailed prompt that I can use to ask Copilot to build a financial model. The model should look like this: a company with four products, and output on tabs. 1) Assumptions tab: Revenue, COGS, opex, capex, working capital, inflation, depreciation, interest, tax. 2) Income statement tab. 3) Cash flow statement tab. 4) An output tab with a dashboard and relevant charts. I want good formatting and dynamic formulas (not hard-coded numbers). Dates should run horizontally across columns, and the model should be populated with formulas that calculate each cell. Test two scenarios: 1) volume increases by 8% and 2) price drops by 5%.

Expected result: Copilot will create a prompt for itself that meets the requirements above and can be copied and pasted back into Copilot to produce a downloadable spreadsheet. As with all prompts, the quality of the output depends on how accurate and detailed the prompt is. When it comes to complex models like this, do not expect the LLM to produce the correct output the first time. And of course, you are responsible for defending the integrity and output of any model you present.

From Hesham Mokhiemer, International Expert Trainer, The Financeer

Want to learn more about AI prompting?

Watch this 10-minute video to learn from Matthieu Hafemeister, Co-founder of Concourse, on how to improve your AI prompting skills.

Ready to explore other applications of AI in finance?

Download the AFP FP&A Guide to AI-Powered Finance for guidance on beginning your AI journey.