More in Spreadsheets & money data.
Build an Excel forecast, scenario and CFO memo with Claude, Copilot and ChatGPT
PartialEvidence: Article or interviewPosted Last checked
Claude beat Copilot and ChatGPT in Mike Dion's Excel forecast tests, but ChatGPT's model missed key costs: a partial result.
How to use this
- Tap Copy prompt.
- Open the tool the case used (or a similar one) and paste the prompt; results may differ from the original.
- Paste it into the message box.
- Send it, then read the answer before acting on it.
New to this? What's a prompt?
A prompt is the message you send to the AI tool. Copy the one on this page, put your own details where it asks, and read the answer before you act on it.
Prompt
Recreated prompt
The source didn't share its exact prompt, so we wrote one from what it describes. The result shown is the source's, not ours.
The person didn't share their exact words, so we wrote this from what they described. We haven't tested it, and they didn't use this exact version.
Build me a forecast model based on this data, twelve months forward, with trend assumptions, structured so I can run scenarios on it. Then show what happens if foot traffic falls 10% while food costs increase 5%, and write a short CFO memo on the results.
Try this in:
Sends the Recreated prompt shown above. ChatGPT and Grok open with it filled in, and ChatGPT may send it straight away. Gemini and Claude open empty, so we copy the prompt for you to paste. Every button also copies it.
Opens in a new tab. We don't see what you send.
These links are only a shortcut. We're not affiliated with OpenAI, Google, Anthropic or xAI, and your results may not match the source.
Optional stronger version
Our rewrite with extra checks. The poster didn't test it.
Before you start, ask me 1–3 short questions about the details below and wait for my answers. Once I've answered, confirm the plan in one line and run the whole task start to finish. Goal: Build a 12-month forecast from my actuals, run one scenario on it and write a short memo for leadership. Questions to ask me first: - What does the business sell, and how many locations or segments are in the data? - Which assumptions should drive the forecast (traffic, average ticket, costs)? - Who reads the memo, and what decision do they need to make? Inputs (fill in or confirm): - Monthly actuals (revenue, cost of goods sold, operating expenses): [SHEET] - Budget figures, if any: [SHEET] - Scenario to test: [FOR EXAMPLE, TRAFFIC -10% AND FOOD COSTS +5%] - Memo reader and decision: [READER / DECISION] Rules: - Before building, tell me your plan and wait for my OK. - Put assumptions on their own tab and link every forecast cell to them; no hard-coded numbers. - Include every line in the source data, including cost of goods sold. - In the scenario, adjust costs for the lower volume first, then apply the price increase. - Lead the memo with the finding and end with a recommendation. Checkpoint: After the base forecast, show me the assumptions and totals before running the scenario. If you get stuck: If the data has gaps or odd months, flag them instead of smoothing them over. Verify before you finish: Check that the base forecast ties to the actuals and that the scenario changes flow through to profit. Output: An assumptions tab, a monthly forecast, a scenario comparison and a one-page memo. Done when: The model is linked, the scenario math is right and the memo is ready to send.
AI tools
Outcome
Mike Dion of F9 Finance gave Claude, Copilot and ChatGPT inside Excel the same data: six months of actuals for a fictional coffee shop business across three locations. Each tool got the same three tasks: a 12-month forecast model with trend assumptions, a scenario where foot traffic falls 10% while food costs rise 5%, and a CFO memo. Claude asked to confirm its plan first and finished first on all three, at roughly 90% usable on the first pass; its memo was the only one close to send-ready. Copilot came second at 60 to 70%, with the scenario's direction right but the math wrong. ChatGPT came last: its forecast left out cost of goods sold entirely, and that broken base carried into the scenario and memo.
Evidence: Article or interviewA newsletter, journalist or interview account of what happened.
Known limitations
PartialThe source says it only partly worked, or had a significant problem a person had to deal with.
What went wrong: Only Claude's work was close to ready. Copilot's scenario math was wrong, and ChatGPT's forecast was missing cost of goods sold.
Source
- Posted
- by Mike Dion (F9 Finance)
- Last checked
- Method:
- How we verify
https://www.f9finance.com/chatgpt-vs-claude-vs-copilot-for-excel/
Context: The tests ran in April 2026 on fictional data, and the tools may have changed since. F9 Finance also promotes its own prompt library and templates.
More in Spreadsheets & money data
Clean up a messy spreadsheet and list issues with Copilot in Excel
Copilot in Excel made a strong manual-review sheet for Tony Phillips but didn't fix many safe issues, so the cleanup was only partial.
PartialCopilotEvidence: Article or interviewView CaseCategorize bank transactions into a rolling 12-month budget with Claude
Kirill Markin had Claude categorize 43 transactions and update a 12-month budget in about 10 minutes of review.
WorkedClaudeEvidence: Described by posterView CaseBuild a monthly expense tracker spreadsheet from one line with Claude
One line to Claude got Amanda Caswell a categorized expense tracker with a summary and working total formulas in about 90 seconds.
WorkedClaudeEvidence: Screenshots or logsBeginner-friendlyView CaseMake a father's handwritten crop ledger, kept since 1971, searchable with ChatGPT
A farmer digitized a father's crop ledger, kept since 1971, so past planting dates and yields are searchable with ChatGPT.
WorkedChatGPTEvidence: Company case studyView CaseTurn a year of bank statements into a business expense spreadsheet with Claude Cowork
Claude Cowork turned a year of statements into a categorized expense spreadsheet; the poster fixed a handful of items by hand.
WorkedClaude CoworkEvidence: Described by posterView CaseBuild a credit risk scorecard in Excel with Claude in Excel
Claude in Excel built a polished 28-sheet credit scorecard but broke the brief's method and overstated accuracy: partial at best.
PartialClaude (Excel)Evidence: Shown on videoView Case

