Claude for Excel
Claude for Excel: an analyst that reads the whole model
Claude for Excel is Claude working inside the spreadsheet itself. It reads the model, follows how the tabs connect, explains what each formula does, and can build or extend it from an instruction written in plain language.
For a finance team that means fewer hours spent decoding a model inherited from someone who left, and less risk from a formula nobody checked. Our workshop teaches how to work this way safely, on your real files.
What changes when Claude sits inside the sheet
Until now, getting AI help with Excel meant copying data out and pasting an answer back. With Claude inside the file it sees the whole structure: the assumptions, the calculations, the links between sheets, and the output the board reads.
That lets you ask where a number comes from, what happens to profit if one assumption moves, or why the summary total does not tie out, and get an answer that points at the cells so you can verify it.
Who it is for
The audience is CFOs, controllers, FP&A teams and analysts: people who build budgets, forecasts and pricing models, and people who must understand somebody else's model before a meeting. You do not need to know AI. You need to know the work.
Most people bring a file that hurts: an enormous model with broken links, a budget grown over years, a monthly report assembled by hand. We work on those files, and we agree up front how sensitive data is handled.
What a session covers
We start with reading: Claude explains a model you know well and you judge whether it is right. Then we move to auditing, which means checking formulas, finding errors, hardcoded values and gaps between sheets, and then to building: adding a scenario, a new tab or a new calculation.
The emphasis is trust. You learn to ask for the reasoning behind each change, to confirm no figure moved that you did not ask about, and to document what Claude did so a controller or an auditor can follow it. Hebrew or English, on your own files.
Where finance teams get stuck
The common mistake is asking AI to build a whole model in one request and receiving something that looks professional and calculates wrongly. A financial model is not taken on faith. Without a way of checking, the extra verification work eats the saving.
You leave with a method: which tasks to hand to Claude with confidence, which to double-check, and how to document the work. You practice the rest in the full course, on financial models.
Keep working on real models
The full course practices reading, auditing and building financial models. Sign-in is required to enter.