Microsoft’s Copilot for Excel will soon be able to tap live financial data feeds and execute reusable finance-specific automations, the company announced on June 25, 2026. The update, part of the broader Microsoft 365 Copilot suite, equips the AI assistant with a trio of enterprise-grade capabilities: modifiable finance “Skills,” direct connectors to real-time market and financial data, and explicit tracking of changes it makes to workbooks. For finance professionals buried in manual data refreshes, reconciliations, and audit trails, the news signals a step change in how Excel can serve as both a modeling canvas and an intelligent, connected hub.
Finance Skills: Turning Prompts Into Reusable Automations
The most ambitious piece of the announcement is the introduction of Copilot “Skills” for finance. Microsoft describes these as editable, repeatable patterns that encode specific financial logic—effectively, macro-like shortcuts powered by natural language. A user might ask Copilot to perform a discounted cash flow analysis, and rather than rephrasing the request each time, they can save that precise sequence of calculations, formatting, and data lookups as a named Skill. The next time they need a DCF, they invoke the Skill with a single command.
This approach borrows from the automation playbook that has made tools like Power Automate and Office Scripts popular among power users, but it democratizes the creation process. Instead of writing code, finance teams can collaboratively build and refine a library of governance-friendly Skills that match their organization’s specific reporting standards and industry conventions. A published Skill can be shared across a department via SharePoint or Teams, ensuring consistency in everything from EBITDA calculations to currency conversion models. Early adopters in the Microsoft 365 Insider program—where features typically land before general availability—are expected to test the feature in the coming weeks.
Under the hood, Skills appear to be a combination of parameterized prompts and the underlying Python or Excel formula logic that Copilot generates on the fly. Microsoft has been steadily expanding Copilot’s ability to produce and run Python code within Excel since early 2024, and Skills likely leverage that runtime for complex statistical or financial functions. The key breakthrough is persistence: a Skill remembers the full context of a request, including cell ranges, data source connections, and output formatting, so that repeating it yields the same structured result without additional guesswork from the AI.
Live Financial Data Connectors: End of the Manual Refresh
The second pillar of the update addresses a chronic pain point: stale data. Excel already supports stock and geography data types via Bing, but the new live connectors go much deeper, linking directly to commercial financial data providers. While Microsoft did not publish a full list by publication time, the announcement references partners that can stream real-time quotes, fundamentals, ESG ratings, and economic indicators into worksheets. Connectors for Bloomberg, Refinitiv, and S&P Global Market Intelligence are widely anticipated, given existing partnerships across Microsoft’s cloud ecosystem.
Once a connector is set up, Copilot can not only pull the data but also interpret it conversationally. A user could ask, “What’s the three-month volatility trend for our top 20 holdings?” and Copilot will retrieve the latest values, compute the metric, and present it as a chart—all without leaving Excel. For treasury teams monitoring currency exposure or FP&A analysts tracking commodity prices, the combination of live data and natural language analysis cuts hours of manual importing and spreadsheet auditing down to seconds. Microsoft emphasized that data freshness can be controlled per connection; some feeds update every minute, while others can be scheduled to align with market close or internal reporting cycles.
Importantly, these connectors are designed with enterprise governance in mind. Administrators can manage which data sources are available via Microsoft 365 admin controls, and data flows are encrypted and compliant with the company’s existing privacy and residency commitments. This addresses concerns from regulated industries—banking, insurance, public sector—that previously hesitated to pipe live external data into Excel because of audit and compliance risks.
Explicit Change Tracking: An Audit Trail for AI
The third headline feature, “explicit change tracking,” tackles the trust gap that has lingered around AI-generated content in spreadsheets. When Copilot modifies a workbook—inserting formulas, restructuring tables, or applying conditional formatting—the system will now log those edits in a dedicated audit pane. Each entry captures what Copilot changed, the natural language prompt that triggered it, and a timestamp. Users can roll back any single modification without affecting others, and the log persists with the file, surviving saves and shares.
This moves Copilot closer to the collaboration and compliance standards already met by Microsoft’s Office apps through version history and track changes. But it goes further by attributing changes to the AI persona, distinguishing human edits from those initiated by Copilot. For finance departments subject to Sarbanes-Oxley or similar internal control frameworks, the audit trail provides the documentation required to accept spreadsheet outputs as part of a controlled process. It also helps teams learn how Copilot interprets their prompts; if an unexpected result appears, the log reveals exactly what the AI did, turning debugging into a teaching moment.
How These Pieces Fit Together
Taken separately, Skills, live connectors, and change tracking each close a gap that has slowed AI adoption in finance. Together, they sketch a vision of Excel as a fully programmable, real-time analytical environment that anyone who can type a sentence can operate. Consider a monthly board reporting cycle: An analyst could maintain a Skill that pulls the latest balance-sheet data from an SAP connector, applies the company’s FX conversion rules, generates key-ratio calculations, and formats the output into a pre-defined template. Because the Skill is reusable, the same process runs in minutes next month with a single click, while change tracking ensures the CFO’s office can inspect every cell that Copilot altered.
Microsoft is also hinting at cross-application scenarios. Skills defined in Excel could be invoked from Copilot interfaces in Outlook or Teams. Imagine a finance manager receiving an email requesting a project’s net present value; a Copilot prompt inside Outlook could execute the Excel Skill, calculate the NPV, and draft a reply that embeds the result—all without opening a workbook. This kind of seamless agentic behavior is the ultimate goal of Microsoft’s Copilot ecosystem, and Excel’s deep integration with core business data makes it a linchpin.
Context: Copilot’s Evolution in Excel
Copilot first appeared in Excel in November 2023, primarily as a text-to-formula assistant that could suggest functions, explain formulas, and create basic charts. Over the next two years, Microsoft layered on capabilities: generating Python for advanced analytics, analyzing data with custom visualizations, and integrating with Power BI datasets. The June 2026 announcement marks the first time Copilot has gained domain-specific “skills” for a vertical industry, signaling a broader strategy. Rather than one-size-fits-all AI, Microsoft is tailoring Copilot for distinct personas—finance first, with other departments likely to follow.
The pivot also aligns with the competitive landscape. Google Sheets has steadily added AI features, and specialist finance AI tools like FinChat.io or Daloopa are gaining traction. By embedding reusable finance logic and live data directly into Excel, Microsoft leverages its enormous installed base: over 800 million people use Excel worldwide, and enterprise finance teams run on it. The message is clear: you don’t need to leave your spreadsheet to get institutional-grade financial intelligence.
The Economics of Copilot and Adoption Hurdles
Copilot in Excel is not free. It requires a Microsoft 365 Copilot license, priced at $30 per user per month (business) or a $20 per user add-on for certain enterprise plans, reflecting the underlying Azure OpenAI cost. New features like live data connectors may also carry consumption-based charges for the data feeds themselves, though Microsoft has not detailed pricing as of announcement day. Finance leaders will need to model the ROI—comparing the subscription cost against the labor savings from automated data refreshes, reduced error rates, and faster decision cycles.
User trust remains the other variable. Change tracking directly addresses the “black box” problem, but finance professionals are famously conservative about tools that touch their spreadsheets. Microsoft’s own research, cited in a 2025 Gartner report on AI adoption, found that while 74% of finance teams wanted AI-driven insights, only 32% felt confident relying on them without manual checks. The addition of explicit logs should narrow that gap, but it won’t disappear overnight. Training and internal pilot programs will be essential.
What Comes Next
Microsoft indicated that the three features will hit public preview for Microsoft 365 Copilot subscribers “in the coming weeks,” with a graduated rollout starting from first-release tenants. General availability is likely by end of Q3 2026, following the pattern of previous Copilot updates. Early feedback from the Insiders program will shape final tweaks to the Skills editor and connector gallery.
The company also teased future Skills packs for accounting (close processes, reconciliations), supply chain (demand forecasting), and human resources (compensation benchmarking), suggesting that finance is just the first act. For now, the focus is squarely on the spreadsheet that runs the world’s numbers. By blending reusable automation, live market data, and AI transparency, Microsoft is betting that Copilot can evolve from a helpful sidekick into an indispensable driver of financial analysis.
For finance teams still on the fence, the message is worth watching: the tools to turn a spreadsheet into a real-time, auditable, AI-powered engine are arriving faster than most close cycles.