If you’ve used ChatGPT, you know it types an answer back. You ask it a question, and it gives you a reply. That’s its main job: generating text. It’s like having a very knowledgeable, very fast typist.
But imagine an AI that doesn’t just talk about things. Imagine one that actually does things.
This is the big shift we’re seeing. Instead of just generating words, a newer kind of AI can take action. It can open applications, pull data from different places, and complete multi-step tasks. Think of it less like a chatbot and more like a highly capable assistant who can roll up their sleeves and get work done.
Why does finance need AI that acts, not just chats?
For decades, Excel has been the backbone of finance. It’s where the quarterly numbers get locked down, where forecasts are debated line by line, and where every single figure needs to trace back to a verifiable source. Finance professionals are always looking for a better tool, something that helps them model reality with more precision than yesterday.
Here’s the thing: that precision comes at a cost. Finance teams spend a surprising amount of time on repetitive, data-gathering tasks. They hunt for information, manually rebuild analyses, and format reports. It’s crucial work, no doubt, but it often pulls skilled professionals away from the high-level judgment and strategic thinking only humans can provide.
Many people think AI in finance means just asking a bot to summarize a spreadsheet or maybe do a quick average. That’s a good start, but it’s not enough for the real-world complexity of financial planning, accounting, or treasury work. A simple chatbot can’t build a discounted cash flow (DCF) model from scratch, using external market data, and then format it exactly how your CFO prefers. It just can’t.
The real power: AI agents with “skills”
This is where Copilot in Excel comes in. It’s not just a fancy chatbot living inside your spreadsheet. It’s an AI agent (a piece of software that can understand your instructions, decide which steps to take, and then actually perform those steps using other tools). It’s designed to automate those multi-step, complex workflows finance teams tackle every day.
Think of it like this: You have a brilliant new junior analyst. You give them a task: “Build a three-statement financial model for Company X, using our standard DCF template, and pull the latest financials from FactSet.”
Now, imagine that analyst already has a detailed playbook for “building a DCF.” They know exactly what data points they need, where to get them, and how to put them together. That playbook is essentially a skill (a pre-defined, repeatable set of instructions for a complex task).
Copilot in Excel works much the same way. You can give it an instruction, like “Refresh the monthly reporting model.” It then accesses a library of sample finance skills (pre-built playbooks for common tasks) or even your own custom skills (playbooks you’ve created yourself). These skills guide Copilot through the exact steps: collecting data, applying specific structures and formatting, and generating the final output. It’s like a highly trained junior analyst who never gets tired and always follows the rules you set.
Grounded in data you trust
A major hurdle for AI in finance is trust. If an AI gives you an answer, you need to know where it came from, how it was calculated, and that the underlying data is solid. Copilot in Excel addresses this by connecting directly to the data financial professionals already rely on.
It uses financial connectors (digital bridges that allow Copilot to pull data from specific, verified external sources) to bring market data, company fundamentals, and research straight into your workbook. This means your analysis starts from the latest, most reliable sources, instead of relying on manual data pulls that can introduce errors or use outdated information. Microsoft has partnered with leading providers like LSEG and Moody’s, and in June 2026, expanded options to include even more:
- CB Insights: Provides predictive intelligence on private companies and markets. Teams use this for sourcing companies and evaluating emerging markets.
- Daloopa: Offers audit-ready fundamentals sourced from SEC filings and investor presentations. Analysts use it to update operating models and reduce manual data entry.
- FactSet: Connects Excel to financial and alternative data for investment professionals. It’s used for modeling, screening, and market analysis.
- Morningstar: Brings investment research, ratings, and portfolio analytics into Excel. Investment teams use it to evaluate holdings and compare funds.
- PitchBook: Delivers institutional-grade private capital market intelligence, including company profiles and deal histories. Teams build target lists and screen investments with it.
- S&P Global – Deterministic Retrieval (powered by Kensho): Provides structured, API-driven access to S&P Global data for AI systems. This helps with company research and multi-entity comparisons, giving predictable, cited results.
You can see how this changes things. Instead of spending hours logging into different terminals, downloading CSVs, and copying-pasting, Copilot fetches it for you, directly into the cells.
Imagine an analyst needing to update an operating model. Historically, this might involve manually extracting financials from multiple SEC filings or investor presentations. With Daloopa’s connector, Copilot can pull that audit-ready data straight into the workbook, saving a significant amount of time. Microsoft’s own finance teams use Copilot in Excel in their real workflows, and they report spending less time hunting for information and rebuilding analyses.
Here’s a simplified look at how Copilot in Excel orchestrates these tasks:
*Copilot in Excel uses the Skills Library and integrates via Financial Connectors to perform complex tasks.*What this means for your business
This isn’t about replacing your finance team. Not at all. It’s about giving them superpowers. By automating the grunt work, Copilot in Excel frees up your most valuable asset: human judgment.
Your finance professionals can spend more time analyzing, strategizing, and making truly informed decisions, rather than spending 37%, give or take, of their day just collecting and formatting data. This shift allows for deeper insights, quicker responses to market changes, and ultimately, a more agile and competitive finance function.
Microsoft has even partnered with the Financial Modeling Institute (FMI), a global body for credentialing top modelers, to pressure-test Copilot’s capabilities. Its library of real-world financial modeling cases is now a foundational part of how Microsoft evaluates the AI, ensuring it meets the high bar of professional finance work. That’s a strong signal of commitment to accuracy and reliability.
The catch
No tool is perfect. While Copilot in Excel offers incredible capabilities, it still needs human oversight. You’ll need to define good “skills” that reflect your company’s specific processes and standards. It’s not a magic button that solves everything without any input.
Also, those powerful financial connectors? Many of them require separate licenses or subscriptions directly from the data providers themselves. So, while Copilot brings the functionality together, you’ll still need to budget for access to those premium data sources like FactSet or PitchBook if you don’t already have them. It’s a tool that enhances existing capabilities, not a free pass to all market data.
Ultimately, the goal is to make Excel even more powerful, turning it into a collaborative hub where AI agents handle the repetitive tasks, and your human experts focus on the strategic decisions that truly move the needle. It’s an interesting time for finance, isn’t it?
Sources: microsoft.com