What did ChatGPT add to Excel in 2026?

In March 2026, OpenAI launched ChatGPT work for Excel — a native AI integration that allows users to build financial models, analyze data, create and update formulas, and pull institutional-grade market data from FactSet, S&P Global, and Moody's directly within Excel. The key difference from Copilot: ChatGPT work can access external market data, not just the data already in your spreadsheet. For accounting firms, this means a junior staff member can now run market benchmarking and financial analysis that previously required specialized software or manual data sourcing.

How should accounting firms use ChatGPT for Excel with client data?

Client financial data can be used with ChatGPT work only through ChatGPT Enterprise or a business account with a data processing agreement that prohibits OpenAI from training on your inputs. Free-tier and consumer ChatGPT accounts should never handle client financial data. For in-house benchmarking and scenario modeling with non-client data, any account tier is appropriate. The practical rule: if the data would be confidential in any other context, use an enterprise-tier account.

What can ChatGPT for Excel do that standard Excel cannot?

ChatGPT work in Excel adds four capabilities Excel lacks natively: (1) natural language formula building — describe what you want and the AI writes the formula, (2) external market data access — pull pricing, financial ratios, and market benchmarks from FactSet, S&P Global, and Moody's without a separate subscription, (3) AI-assisted data analysis — ask questions about a dataset in plain language and receive structured analysis, and (4) financial model generation — describe the model structure and the AI builds it, including sensitivity tables and scenario tabs.

How is ChatGPT for Excel different from Microsoft Copilot in Excel?

The core difference is data access. Microsoft Copilot in Excel works with the data already in your spreadsheet. ChatGPT work in Excel can also access external market data from OpenAI's financial data partners. For accounting firms doing client benchmarking, valuation work, or financial advisory, this external data access is the more significant capability. For firms primarily doing formula management and internal analysis, the two tools are more comparable.

ChatGPT for Excel Accounting Firms 2026

Published April 18, 2026 · Updated April 2026 · By The Crossing Report · 6 min read

Summary

ChatGPT work for Excel launched March 5, 2026, with AI financial modeling and institutional-grade market data from FactSet, S&P Global, and Moody's accessible natively in Excel
The implications for accounting firms: a capability gap is forming between firms that have tested this in client work and firms that haven't — and the gap will be visible to clients
Three immediate moves for accounting firms: get access, run one real client task, document the workflow before a client asks you to
Intuit's QuickBooks-Anthropic integration is rolling out spring 2026 in parallel — AI is entering the accounting workflow from multiple directions simultaneously

What Happened on March 5, 2026

OpenAI launched ChatGPT work for Excel on March 5, 2026.

This is not a minor update to spreadsheet software. This is institutional financial data — FactSet, S&P Global, Moody's — accessible through natural language queries inside the tool where most accounting firms already do their financial analysis work.

Before March 5: a junior staff member who needed market benchmarking data for a client valuation pulled it manually from a subscription data service (if the firm had one) or approximated it from public sources. The process took 1–3 hours depending on the data needed and the analyst's experience.

After March 5: the same staff member types a natural language query in Excel — "pull current P/E and EV/EBITDA multiples for the top 10 companies in the SIC code for commercial printing" — and the data appears. The process takes under 10 minutes.

The time differential is obvious. What's less obvious but more important for accounting firm owners: your corporate clients now have access to the same capability. The sophisticated financial analysis that used to differentiate your firm's advisory work is now a tool your clients' finance teams can run themselves.

This is the pattern that defined the transition in legal (where GC AI is now deployed at 1,400+ companies) arriving in accounting. The time to prepare is not after your clients realize it. It is now.

What ChatGPT Work for Excel Actually Does

AI Formula Building

Describe what you need in plain language, and ChatGPT builds the formula. "Create a formula that calculates the year-over-year percentage change for each row in column B compared to column C, and flag anything over 15% in red" becomes a working formula with conditional formatting in seconds.

This is not revolutionary for experienced Excel users, but it meaningfully reduces errors from junior staff who would otherwise search Stack Overflow or ask a senior for help. More importantly, it speeds up formula construction for complex models where even experienced analysts spend 20–30 minutes on formula architecture.

External Market Data Access

The differentiating capability: FactSet, S&P Global, and Moody's data accessible through natural language queries without a separate data subscription.

What this enables:

Comparable company analysis for valuation work
Industry financial ratio benchmarking
Market pricing and yield data for financial advisory
Credit rating and debt structure analysis for corporate clients

For accounting firms that do business valuation, M&A advisory support, or financial benchmarking, this changes the economics of data gathering significantly.

AI Financial Model Generation

Describe a model structure and ChatGPT builds it. "Create a three-statement financial model with an income statement, balance sheet, and cash flow statement linked through net income and working capital changes, with a 5-year projection tab and three scenario columns (base, upside, downside)" generates a working model framework.

The output requires review and customization — it is a starting point, not a finished deliverable. But a starting point that takes 5 minutes instead of 90 minutes changes how quickly your team can respond to client requests.

Three Moves for Accounting Firms

Move 1: Get Access and Run One Real Client Task

ChatGPT work for Excel is available through ChatGPT Plus ($20/month) and higher tiers. The first step is getting one person — your most Excel-proficient professional — access and assigning them one real client task to complete using the tool.

Do not run a training session before real use. Real use is the training. The learning that happens in 90 minutes of actual work on a client deliverable exceeds the learning from any abstract demonstration.

Document what you learned: which tasks it accelerated, where it needed correction, what you would do differently. That documentation becomes your firm's starting prompt library.

Move 2: Document the Workflow Before a Client Asks

Your clients' finance teams are learning this tool at the same time you are. The firms that will be positioned as AI-capable advisors are the ones who have a tested workflow and can demonstrate it — not the ones who are figuring it out after a client asks whether you've tested the new Excel AI.

One documented workflow — "here is how we run comparable company analysis using ChatGPT for Excel, here is the data we pull, here is our review process" — is enough to signal AI competence to a sophisticated client audience.

Move 3: Update How You Describe Your Value

When a client can run a comparable company analysis in Excel using publicly available AI tools, the value of "we ran comps" as a service description erodes. The value that doesn't erode: your judgment about what the comps mean for this client's situation, the context of the firm's history and goals, and the accountability of a licensed professional behind the analysis.

Update your service descriptions to emphasize judgment, interpretation, and accountability — not the analytical mechanics. "We benchmark your performance against industry peers and interpret what the gaps mean for your next three strategic moves" is the value proposition that AI cannot replicate.

The Parallel Development: QuickBooks and Anthropic

On the same timeline as the ChatGPT-Excel launch, Intuit announced a partnership with Anthropic to integrate Claude AI directly into QuickBooks. The integration — entering beta spring 2026 — allows accountants to run natural language queries against client QuickBooks data inside the platform: "Show me this client's cash flow trend for the last 18 months with three key drivers" surfaces in seconds.

The significance: AI is entering the accounting workflow from multiple directions simultaneously. Excel is the analysis layer. QuickBooks is the data layer. General-purpose AI (Claude, ChatGPT) is the drafting and research layer. Each integration reduces a friction point that previously kept AI occasional rather than daily.

Accounting firms that treat these as isolated tool updates will find themselves perpetually in catch-up mode. Firms that see them as a single shift — AI becoming the operating system of professional accounting workflow — and build their practice around that shift will have a durable competitive position.

Sources

OpenAI: ChatGPT work for Excel launch announcement, March 5, 2026
Intuit: QuickBooks-Anthropic AI partnership announcement, Spring 2026
FactSet, S&P Global, Moody's: Financial data partnership announcements with OpenAI, March 2026

The Crossing Report covers the AI transition for professional services firm owners. The premium tier includes a ChatGPT-for-Excel workflow guide for CPA firms, including tested prompts for comparable company analysis and client financial benchmarking. Subscribe here →

ChatGPT for Excel in Accounting Firms: What You Need to Know (2026)