ChatGPT for Excel: 6 FP&A Workflows + 1 Custom Skill
Install the ChatGPT for Excel sidebar, ship 6 documented FP&A workflows, route around the 3 it flunks, and write your first custom Skill — for variance, close, or scenario rollup.
On Tuesday, May 5, 2026, OpenAI flipped the global GA switch on the ChatGPT for Excel sidebar. Hg Capital’s private equity team had been beta-testing it for months and called it “materially accelerated our research and due diligence workflows.” OpenAI’s internal investment-banking benchmark went from 43.7% with GPT-5 to 87.3% with GPT-5.4 Thinking on the same task — building a three-statement model with proper formatting and citations. That’s the launch. This is the course that turns it into your weekly workflow.
This is a Quick Skill course — 2 hours, 8 lessons, one capstone. The first two lessons are free, so you’ll have the sidebar installed and a working first-pass model before any subscription decision. Lessons 3-8 cover the six FP&A workflows the sidebar handles reliably (three-statement model, scenario tabs, variance commentary, multi-tab audit, headcount plan, month-end close), the three workflows it still flunks (VBA/macros, Power Query/DAX, the oversized inherited workbook), the Skills system that lets you save reusable playbooks for your own recurring work, and the Plus/Pro/Business/Enterprise tier decision against your data-classification posture.
The course is built on the public primary sources — OpenAI’s launch post, the Help Center documentation, the Microsoft Office Marketplace listing, and the May 2026 practitioner walkthroughs from MindStudio, Spicy Advisory, and Pasquale Pillitteri — plus the testimonials from Hg, Bain, BCG, McKinsey, Accenture, and PwC that OpenAI named as enterprise deployment partners on launch day. Every workflow we teach comes with a verification checklist for the specific failure modes the sidebar still has: the silent hardcoded balance-sheet plug, the 15%-miss-rate sign-convention error on variance lines, and the missing-tab summary that drops four of twelve tabs without telling you.
By the end of Lesson 8, you’ll have written a custom Skill for one recurring monthly workflow from your own job — variance commentary, month-end close, or scenario rollup — and you’ll be running it against a real (or de-identified test) workbook. The Skill becomes the reusable playbook your team can run every month without restating the format. The certificate becomes the proof you can ship AI-augmented financial work with the audit discipline intact.
If you’d rather start with the comparison view — Claude for Excel vs ChatGPT for Excel side-by-side on four real FP&A tasks before you pick — start with our Claude for Excel course. If you’re looking for the broader AI-in-spreadsheets scope (Google Sheets + Excel + the marketing-ops angle), the AI for Spreadsheets course is the parent course this one branches from.
What You'll Learn
- Explain what ChatGPT for Excel is, how it differs from Microsoft Copilot in Excel and Claude for Excel, and what the May 5, 2026 global GA actually shipped
- Use the Microsoft Office Marketplace install path to add the ChatGPT sidebar to a working Excel session and confirm the permission-prompt trust signal
- Apply the three-statement modeling workflow with the Drivers-tab pattern and the cash-flow tie-out + retained-earnings + circular-reference audit checklist
- Apply the scenario-sensitivity, variance-commentary, and multi-tab audit workflows using the @-pin syntax and the DATA TABLE selection pattern
- Apply the headcount-plan and month-end close reconciliation workflows with the fully-loaded-cost-at-1.3x-base pattern
- Analyze the three failure modes — VBA/macros, Power Query/DAX, and the oversized-workbook context overrun — and route each to the appropriate alternative tool
- Evaluate the Plus, Pro, Business, and Enterprise tier choice against workload, data-classification posture, and Compliance-API audit needs
- Create a custom Skill (reusable playbook) for one recurring monthly workflow from your own job
After This Course, You Can
What You'll Build
Course Syllabus
Prerequisites
- Active paid ChatGPT plan (Plus, Pro, Business, or Enterprise — $20/month entry tier is enough for L1-L7; the capstone runs cleanly on Plus)
- Microsoft Excel desktop or Microsoft 365 (the sidebar add-in installs from Home → Add-ins)
- Working comfort with Excel: tabs, named ranges, basic formulas (XLOOKUP, INDEX/MATCH, SUMIFS), PivotTables at the consumer level
- One de-identified workbook from your own job to test against during L3-L5 (a budget, a 3-statement model template, a month-end reconciliation — sample data works)
Who Is This For?
- FP&A analysts and senior analysts at mid-sized companies doing monthly variance commentary, scenario rollups, and three-statement model maintenance
- Controllers and assistant controllers running month-end close on small-to-mid sized books
- Solo CPAs and small-firm CPAs who own the month-end close for 5-30 client books
- Finance directors and CFOs evaluating ChatGPT for Excel tier rollout for a 5-50 person finance team
- Students and job-seekers prepping for FP&A interviews who want a structured way to practice three-statement modeling against AI-graded feedback
Frequently Asked Questions
Do I need to be a financial modeler to take this course?
No. The course assumes you're comfortable with Excel and you do at least one FP&A-style workflow at work (variance commentary, a close package, a scenario rollup, a headcount plan, a three-statement model — pick one). The course is built around the documented six workflows, so any one of those entry points works. We do not teach financial modeling from scratch.
Will the sidebar work on Mac, on the web version of Excel, or in Google Sheets?
The Excel add-in ships through the Microsoft Office Marketplace and works on Excel for Windows desktop, Excel for Mac, and Excel for the web — wherever your paid Microsoft 365 account is signed in. The same OpenAI sidebar also runs in Google Sheets via a separate install path, which is the single biggest architectural difference from Claude for Excel. We focus the course on Excel because that's where FP&A lives, but the Google Sheets sidebar uses identical prompt patterns.
I'm on ChatGPT Plus. Will I hit usage limits in the capstone?
Plus is the right floor for L1-L7. The capstone — writing a custom Skill and running it against a real workbook — is well under Plus's daily agentic-usage cap. If your monthly workflow is the 12-tab month-end close (Lesson 5), you'll hit the cap faster, and that's the inflection point where Business pays for itself for solo controllers. Lesson 7 covers the exact tier-by-workload decision.
What's the difference between this and the Claude for Excel course?
Different vendor, different model family, different architecture. ChatGPT for Excel runs on GPT-5.5 and ships with a built-in Skills system (reusable playbooks). Claude for Excel runs on Sonnet 4.6 and ships with MCP connectors (S&P Global, LSEG, Moody's, FactSet wired directly into the sidebar). Most FP&A teams end up running one tier of one and a personal subscription to the other. The companion comparison blog and the dedicated Claude for Excel course go deeper on the head-to-head — link in Lesson 1.
How current does this stay? The May 5 launch is fresh.
Last reviewed 2026-05-27. We're on a 3-month review cadence — the Excel sidebar is a frontier-model product (GPT-5.5 today, future versions later), pricing tiers may shift, and the Skills syntax is still maturing. We'll refresh the prompt patterns and the tier-comparison numbers every quarter. If you find anything that doesn't match what you see in your own Excel session, ping us — the lesson content updates ahead of the review window when vendor-side material moves.
