Should accounting firms be worried about ChatGPT for Excel?

Worried is the wrong frame. Accounting firms should be attentive. The tool your clients use every day now has AI built into it — and that changes what clients can do independently. The firms that adapt fastest will reposition from 'the person who builds your spreadsheet' to 'the advisor who interprets what it means.' The firms that ignore it will notice clients asking fewer questions that used to generate billable work.

How can accounting firms use ChatGPT for Excel to improve efficiency?

Three high-ROI starting points: (1) variance analysis — ask ChatGPT for Excel to compare current vs. prior period financials and flag anything outside normal ranges, in plain language; (2) client report first drafts — generate a narrative summary of a client's financials directly from their spreadsheet data; (3) model building — use natural language to create projections, cash flow models, or budget templates instead of building formulas manually. In each case, the AI produces the draft and the accountant applies judgment and reviews the output.

Can clients use ChatGPT for Excel to do their own accounting?

Some routine tasks, yes. That's the honest answer. Clients who use ChatGPT for Excel can now ask their own books basic financial questions, run their own variance analyses, and generate simple financial narratives without calling their accountant. The work that survives this shift is the work that requires professional judgment: tax planning, audit, strategic advisory, exception handling, multi-entity complexity. Firms that define their value around judgment and relationships are more defensible than firms that define their value around spreadsheet access.

What data integrations does ChatGPT for Excel include?

At launch, ChatGPT for Excel integrates with FactSet, S&P Global, LSEG, Moody's, and Dow Jones Factiva for institutional-grade market data and company fundamentals. This means users can pull live market data and financial benchmarks directly into a spreadsheet using natural language — without a separate data subscription or export process.

ChatGPT for Excel: Accounting Firms 2026

Q: What is ChatGPT for Excel and how does it work?

ChatGPT for Excel is an AI assistant embedded directly inside Microsoft Excel (and Google Sheets, coming soon) that launched in beta on March 5, 2026. It can build financial models, analyze data, create and update formulas, and pull institutional-grade market data from partners like FactSet, S&P Global, and Moody's — all using natural language instructions inside the spreadsheet. It's available to ChatGPT Business, Enterprise, Pro, and Plus subscribers in the US, Canada, and Australia.

On March 5, 2026, OpenAI quietly changed the relationship between accounting firm owners and their clients.

ChatGPT for Excel launched in beta — an AI assistant embedded directly inside Microsoft Excel that can build financial models, analyze data, write and update formulas, and pull institutional-grade market data from FactSet, S&P Global, Moody's, and others using nothing but natural language instructions. No IT project. No export. No waiting for their accountant.

Your clients already use Excel. Now it has a capable AI inside it.

This isn't a distant threat. It's available now to any client with a ChatGPT Business, Pro, or Plus subscription. That's $20–$30 per month — a rounding error in any business's software budget.

Here's what this means for your firm, and what you should do about it before they figure it out on their own.

Summary

In March 2026, OpenAI embedded AI directly into Microsoft Excel — the tool your clients already use — giving them the ability to build financial models, analyze data, and pull institutional data from FactSet and S&P Global using plain English. This is not a future threat: it's available to your clients now. The question for accounting firm owners is which parts of the current service model are being commoditized and where the firm's value needs to shift.

What ChatGPT for Excel Actually Does

The AI assistant runs inside Excel (Google Sheets integration is in development). It accepts natural language instructions and responds with:

Formula building — "Create a 3-year revenue projection based on these actuals with 8% annual growth." It writes the formulas. You review them.
Variance analysis — "Compare this year's monthly revenue to last year. Flag anything that's more than 15% different and explain possible causes." It produces the analysis in readable language.
Financial model updates — "Update this cash flow model with the new inventory assumptions in column D." It applies changes across the model automatically.
Market data pull — At launch, integrations with FactSet, S&P Global, LSEG, Moody's, and Dow Jones Factiva let users pull institutional-grade company data and benchmarks directly into a spreadsheet. This is data that previously required a separate subscription and a manual export process.

The underlying model, GPT-5.4, scores 91% on Harvey's BigLaw Bench — a benchmark for complex, document-intensive professional work — and matches or outperforms humans in 83% of head-to-head comparisons across 44 occupations tested, including accounting.

That's not a toy. That's a capable tool running inside the most widely used business software in the world.

The Three Scenarios Playing Out in Your Market Right Now

When a new AI capability lands in software your clients already use, three things happen in the next 12–18 months. Understanding which scenario you're in determines what to do next.

Scenario 1: Your clients start building their own models

A client who previously called you for a quarterly variance analysis figures out they can ask ChatGPT for Excel in plain language and get a readable summary in 10 minutes. They stop calling. You don't notice at first — it's just one recurring task going quiet.

The accumulated version of this, across your client roster, is a quiet revenue leak. The work doesn't disappear from your invoice dramatically. It just stops appearing.

The response: Reposition before the leak starts. Stop defining your value as "we build your financial models." Start defining it as "we interpret what your models mean for your business decisions." That's a judgment service. ChatGPT for Excel can produce the model. It can't make the call on whether the numbers support the acquisition.

Scenario 2: Your firm adopts ChatGPT for Excel internally

You get access ($20–$30/month per ChatGPT subscription), start using it on client work, and discover you can produce a first-draft variance analysis in the time it used to take to open the client's prior-period file.

The time savings compound. A task that took 45 minutes now takes 12. You can review two more client files in the time that frees up. Or you can charge the same and take the margin. Either way, the capacity math changes.

The response: Claim this advantage before competitors do. Accounting firms that adopt ChatGPT for Excel and document the workflow templates will have a consistent efficiency edge. The ones that don't will be on the wrong side of a productivity gap when clients start comparing response times and pricing.

Scenario 3: Your firm does nothing

Your clients adopt ChatGPT for Excel on their own. Some discover it independently through their Microsoft 365 environment or their team's software exploration. Others read about it and experiment. They don't tell you — they just start doing things they used to call you about.

When they do call, they arrive with more context, more specific questions, and lower tolerance for basic explanations. The ones with the most AI confidence start internalizing more. The ones with questions about complexity still reach out — but the volume from the straightforward questions drops.

Six months from now, the delta is visible in your utilization data. Twelve months from now, it's visible in revenue.

What to Do This Week

You don't need to overhaul your practice. You need to make three moves.

Move 1: Get access and run one real client task through it.

Sign up for ChatGPT Plus or Business if your firm doesn't have a subscription yet ($20–$30/month). Connect ChatGPT for Excel. Pull up a real client file — not a test file, a real one. Ask it to run the variance analysis you'd normally build manually.

Review what it produces. Not to be impressed by it — to understand where it's right, where it needs judgment applied, and what the output quality is for your clients' actual data. That 30-minute experiment tells you more than any product demo.

Move 2: Document the workflow before a client does.

The firms that win in AI transitions are the ones that systematize their workflows before competitors or clients systematize them for you. Pick the three most common Excel-based deliverables you produce for clients: variance analysis, cash flow projections, budget-to-actual reporting. Write down the steps. Then map where ChatGPT for Excel changes those steps.

You're building the AI-enhanced workflow template that becomes your firm's standard. When a new associate joins in six months, they start with this template. When a client asks how your pricing compares to a competitor, the answer includes this efficiency.

Move 3: Update how you describe your value to clients — now, not after the conversation forces you to.

The accounting services that survive the next 18 months of AI adoption are the ones clients can't replicate with a $20/month tool: your professional judgment, your understanding of their specific situation, your accountability for the output, your relationship with their team.

That's the service you're selling. The spreadsheet mechanics are increasingly background.

Most firm owners wait until a client asks "why am I paying for this when ChatGPT does it?" before articulating the difference. The firms that get ahead articulate it first. One sentence in your next engagement update or meeting — "We use AI tools to produce first drafts faster, and we apply professional judgment to every output before it reaches you" — reframes the conversation before it becomes a negotiation.

The Bigger Pattern This Fits

ChatGPT for Excel is one of two major AI integrations landing in the platforms accounting firms' clients use every day.

The other is the Intuit/Anthropic partnership announced February 25, 2026 — Claude AI agents embedded directly into QuickBooks, rolling out to customers this spring. QuickBooks on one side, Excel on the other. The two platforms where most small business financial data lives are both getting capable AI.

Your clients are going to be able to query their own books, build their own models, and generate their own financial narratives without a phone call. That's the reality of Q2 2026.

The accounting firms that adapt fastest will be the ones that reposition around what AI can't replicate: the professional judgment call, the regulatory knowledge, the relationship, the accountability for the output. The ones that wait will have the conversation on their clients' terms instead of their own.

ChatGPT for Excel launched 9 days ago. You have a narrow window to get ahead of this before your clients become aware of it on their own.

The Crossing Report covers AI adoption for professional services firm owners every Monday. Subscribe here for the weekly briefing.

Related reading:

ChatGPT Just Moved Into Excel — What Accounting Firms Need to Do Before Their Clients Figure It Out