Helping finance professionals master FP&A, Excel, data, and CFO advisory services through learning experiences, masterminds, training + community | Adjunct Professor in Data Analytics | Microsoft MVP
Hello to the new COPILOT function in Excel. This is going to change the way we use Excel and FP&A modeling in a major way. Here's some of what to expect: (1) Generating Summaries: COPILOT will examine large data sets and summarize them into clear, concise narratives. It will help you identify trends that you may not have spotted otherwise. (2) Creating lists and dynamic tables: The COPILOT function can create lists and dynamic data tables that can drive financial models. While Power Query is still a go-to option for connecting Excel to back-end data, this offers a new way to tee up financials for use in forecasting and analysis. (3) Classifying data: You can use COPILOT as a function to categorize non-financial data such as written feedback or survey responses into useable formats. This allows seamless work within Excel, instead of needing to leave the software or use half-baked add-ins that don’t work very well. Suggestions for getting the most from the Excel COPILOT function: (A) Be concise: Like with any generative AI tool, how you write your prompts makes a huge difference. Be objective, be clear, and be contextual. The better your instructions, the better your outputs. Treat Copilot like it's a child who may not know what you're talking about. We're not yet at a stage where Copilot can read our minds and our work habits. (B) Be direct: Use instructions like "rank" or "summarize" or "categorize" to instruct COPILOT to perform those commands. Being vague is not a good strategy for getting the most from the tool. (C) Be easy to work with: COPILOT (the function) only uses data available within the LLM itself. If you need the COPILOT function to analyze workbook data, first import that data into your file before referencing it. Like with any AI-generated results, especially in finance, outputs should be reviewed and validated for accuracy. These are great tools, but they aren't fool-proof. -------------------- When I developed the first Copilot in Excel for FP&A course for LinkedIn Learning, I knew that it was only the beginning of AI-powered FP&A within the most widely-used finance tool on the planet. It's been interesting witnessing the development and rollout of this tool. Now we get to see one small step with a function but a giant leap forward for capability.
Copilot won’t replace finance brains—it’ll expose which ones were hiding behind manual grind. Summaries and classifications are helpful, but the edge stays with leaders who can turn AI’s shortcuts into sharper, faster decisions.
Love this, Carl
Carl Seidman, CSP, CPA The data classification piece could be huge for variance analysis instead of manually coding dozens of budget vs actual explanations, COPILOT could categorize them by theme (volume, pricing, timing, etc.). This would make monthly commentary much more systematic and help identify recurring drivers across departments.
Thanks for sharing, Carl
Fully agree Carl Seidman, CSP, CPA
Really interesting update Carl! I’m still a new learner with Copilot in Excel and have been experimenting with how it can fit into FP&A workflows. I am also exploring how AI copilots can support finance processes more broadly specifically in the FP&A space. If anyone here has been pushing the limits with Copilot in Excel, I would love to connect and learn from your experience.
Great post, Carl Seidman, CSP, CPA! One of the best things I've learned to do recently is to describe the use case and then to ask the LLM to generate the best prompt, I'll then tweak the methodology from there; for instance the suggested prompt may contain something I didn't think of, and I can take things one step further, or the prompt is structured in a way that won't make the result quite right and it'll require a bit of a revision, either way I say I tend to get quicker and better results this way.
Thanks for sharing, Carl
Thanks for sharing, Carl. Interesting. I am Yet to Start this.
Great summary, Carl! Thank you for sharing