ChatGPT Comes to Excel and Google Sheets

OpenAI has put ChatGPT directly inside Excel, a move that lands squarely in Microsoft’s home territory.

On May 5, ChatGPT for Excel and Google Sheets reached general availability across platforms and subscription tiers. The Excel beta began in March and the Sheets beta followed in April, so both pilots moved to GA in about two months. For a product that spans two rival productivity ecosystems, that pace is unusual.

The more important detail is not simply that ChatGPT now sits inside a spreadsheet. OpenAI announced a group of data partners at the same time: FactSet, S&P Global, Moody’s, MSCI, LSEG, Dow Jones Factiva, Daloopa and Third Bridge.

Taken together, those names change the product’s profile. This is not just an assistant for writing formulas. It is aimed at the financial-data stack that Wall Street analysts use every day.

What ChatGPT Can Do Inside a Spreadsheet

  • Build full tables: a user can ask for a three-year cash-flow statement covering operating, investing and financing sections, and ChatGPT can create the structure, formulas and references.
  • Clean data: pasted-in messy data can be normalized through natural-language instructions, such as standardizing dates to YYYY-MM-DD or marking negative numbers in red.
  • Explain formulas: users can open a complex model and ask what a specific formula is calculating.
  • Work across tabs: ChatGPT can help build models that connect multiple worksheets.
  • Pull financial data: integrations with partners such as S&P Global and FactSet can bring market data into the workflow without constantly switching to a separate terminal.

“You no longer need to leave your spreadsheet to get help.”

In plain terms, OpenAI wants to compress a familiar analyst workflow. What used to involve Excel, a Bloomberg terminal, web research, internal knowledge bases and occasional ChatGPT prompts could increasingly happen in one place.

Why This Matters for Microsoft

Microsoft’s own Copilot has been running inside Excel for more than a year. In theory, Excel is Microsoft’s strongest ground, and OpenAI should have a difficult time entering it. But this GA release changes the balance in a few ways.

  1. It does not require a Microsoft 365 Copilot subscription. ChatGPT for Excel comes through ChatGPT subscriptions, including Plus, Pro, Business, Enterprise, Edu and K-12 plans. Copilot is sold separately on a per-user monthly basis.
  2. It also supports Google Sheets. Microsoft Copilot is not going to live inside Sheets. OpenAI can now cover both spreadsheet ecosystems.
  3. It runs on GPT-5.5. Copilot uses GPT models deployed through Azure, but Microsoft controls the versioning and feature cadence. ChatGPT for Excel is tied to OpenAI’s current GPT-5.5 experience.

For a financial analyst, the choice may come down to procurement and workflow. A company already paying for Microsoft 365 E5 with Copilot may keep analysts on Copilot. An individual user, a startup team, or a Mac and Google Workspace shop may be more likely to use ChatGPT for Excel and Sheets.

The data-partner list also points directly at finance. FactSet and S&P Global Market Intelligence are standard tools on Wall Street. MSCI is a major name in indexes and ESG data, Moody’s covers credit ratings, LSEG supplies equity-market data, Daloopa structures SEC filings, and Third Bridge offers expert-interview research.

Data sourceMain use
FactSetCompany financials, valuation and research
S&P GlobalCredit, market indexes and industry data
Moody’sCredit ratings and risk data
MSCIIndexes and ESG data
LSEGGlobal equities and derivatives data
Dow Jones FactivaFinancial news archives
DaloopaStructured SEC filing data
Third BridgeIndustry expert interviews

That lineup presses directly against Bloomberg’s core territory. A Bloomberg Terminal seat costs roughly $27,000 a year, while Refinitiv, now part of LSEG, has been priced around $20,000 a year. OpenAI is positioning some of this data access inside a ChatGPT product that starts far below that level for consumer users.

The Practical Limits

The promise is strong, but several caveats matter.

First, financial-data rights are not controlled by OpenAI alone. OpenAI has described the partner access as app-based integrations, but it has not spelled out how much usage is included each month or how overages are billed. Professional analysts can hit FactSet hundreds of times a day, and that kind of usage is unlikely to be cheap.

Second, hallucination risk is much higher in finance. If ChatGPT writes the wrong formula or misreads a company ticker in a valuation model, the resulting forecast can be off by hundreds of millions. OpenAI says GPT-5.5 has a hallucination rate 52.5% lower than GPT-5.3 Instant, but financial work has a much lower tolerance for error than ordinary Q&A.

Third, enterprise compliance remains hard. Uploading core financial models to a cloud AI service is the kind of issue that large banks’ IT and legal teams scrutinize closely. OpenAI says its enterprise products provide data isolation, but each investment bank’s general counsel will still need to review the actual contract language.

Bottom Line

OpenAI’s recent strategy has not only been to make ChatGPT smarter. It has been to put ChatGPT inside the tools where people already work: Excel, Word, Sheets, Slack and Microsoft Teams.

Every new integration reduces Microsoft Copilot’s room for differentiation. If every Office workflow eventually supports two AI assistants, the contest comes down to model quality and price. Those are the two dimensions where OpenAI is most comfortable competing.

How Microsoft responds may be clearest not in analyst notes, but in the next set of Copilot penetration numbers it chooses to disclose in earnings.

Sources: Introducing ChatGPT for Excel and new financial data integrations (OpenAI); ChatGPT for Excel and Google Sheets is now out of beta, and it's kind of a big deal (Digital Trends); OpenAI launches ChatGPT add-in with Excel, Google Sheets for all users (The Peninsula Qatar); CocoLoop; OpenAI Releases ChatGPT Excel Add-In with Advanced Financial Data Integrations (mlq.ai)