Microsoft has unveiled a major update to Copilot in Excel that directly targets finance professionals, introducing customizable AI 'skills,' expanded data connectors, and powerful new planning and traceability features. Announced on June 25, 2026, the capabilities are rolling out to Microsoft 365 Copilot subscribers and mark the most significant finance-focused update since Copilot's debut. The move underscores Microsoft's strategy to embed AI deeply into mission-critical workflows—and Excel remains the undisputed hub for financial modeling, reporting, and analysis.

For years, Excel has been the backbone of finance teams. Now, with Copilot's new toolset, Microsoft is addressing the gap between traditional spreadsheet work and the growing demand for automated, auditable, and connected financial processes. From generating variance analyses with a single prompt to pulling live data from ERP systems, the new capabilities are designed to reduce manual grunt work and let finance professionals focus on strategic decisions.

What Are Copilot Skills in Excel?

At the heart of the update are 'skills'—prebuilt, repeatable workflows that Copilot can execute on demand. Think of them as AI-powered macros that understand context. Instead of writing complex formulas or VBA scripts, users can invoke a skill like "Analyze budget vs. actuals for Q2" and Copilot will automatically retrieve the relevant data, perform calculations, generate a formatted summary, and even highlight anomalies.

Microsoft says these skills are specifically tuned for common finance tasks. Initial skill templates include:

  • Variance Analysis: Compares actuals against budgets or forecasts, identifies drivers of deviation, and explains the impact in natural language.
  • Financial Statement Generation: Creates income statements, balance sheets, and cash flow summaries from raw transaction data.
  • Forecasting: Extrapolates trends using multiple models (linear regression, exponential smoothing) and provides confidence intervals.
  • What-If Scenario Modeling: Adjusts key assumptions—like revenue growth or COGS percentages—and instantly recalculates downstream impacts.
  • Reconciliation: Matches entries across two datasets, flags discrepancies, and suggests possible resolutions.

Users can customize these skills by adding their own business rules, naming conventions, or formatting preferences. The skills also support chaining: a variance analysis can automatically trigger a forecast update, all with full auditability. Copilot logs every step, making it possible to trace how a final number was derived—a non-negotiable requirement for finance teams under regulatory scrutiny.

New Data Connectors for Live Financial Feeds

Static spreadsheets are becoming a liability when data changes by the minute. With this update, Copilot in Excel introduces new connectors to bring live data directly from common financial systems:

  • Dynamics 365 Finance: Pulls general ledger entries, accounts payable/receivable, and cash flow data.
  • SAP ERP: Connects via Microsoft's existing SAP gateway to extract financial reports in real time.
  • Workday Financial Management: Retrieves payroll, procurement, and expense data.
  • QuickBooks Online: Suitable for small and medium businesses, now with Copilot-assisted analysis.
  • Custom Connectors via Power Query: Finance teams can build connectors for proprietary systems using the existing Power Query framework—and now Copilot can suggest and even generate connector logic.

These connectors aren't just simple imports. Copilot understands the schema of each source, so a request like "Pull month-end revenue from SAP and compare with our Excel forecast" works without the user having to specify tables or fields manually. The AI maps dimensions (like cost centers, regions, periods) intelligently, reducing the setup time that typically plagues cross-system reporting.

Planning and Traceability: From Black Box to Glass Box

Finance departments have long demanded transparent AI. The new 'traceability' feature addresses this by turning Copilot's outputs into something that can be audited. When a skill generates a number—say, a gross margin forecast of 42.7%—users can click 'Show Work' to see the exact logic chain: which data sources were used, how outliers were treated, which forecasting algorithm was selected, and any assumptions made.

This is a direct response to the 'black box' criticism that many professionals have leveled against AI tools. "Finance leaders won't trust AI-generated numbers unless they can verify the methodology," said one early adopter from a Fortune 500 company during a preview briefing. "This traceability function is the difference between a toy and a tool."

On the planning front, Copilot can now assist with full financial planning cycles. Users can define a planning model—say, an annual operating plan broken down by month and department—and Copilot will:

  • Ingest historical data and current run rates.
  • Suggest driver-based assumptions (e.g., headcount growth, inflation factors).
  • Populate the plan template with initial values.
  • Allow iterative adjustments via natural language: "Increase marketing spend by 10% in Europe and re-allocate the difference to R&D."
  • Track version history so that plan iterations are never lost.

Planning and traceability together mean that a CFO can finally have an AI-generated budget that they can defend to the audit committee. The system maintains a full lineage from source system to final plan cell.

Impact on Windows Users

Excel is, of course, a premier Windows application—and the new Copilot features light up natively on Windows 11 and Windows 10 devices running the latest Microsoft 365 apps. The features are deeply integrated into the Windows environment: use the Copilot side pane, voice input via Windows Speech Recognition, and touch-friendly interfaces on Surface devices.

For Windows-based finance professionals who spend their days toggling between Excel, Outlook, Teams, and specialized ERP clients, the update promises a dramatic reduction in context switching. Copilot can pull data from an SAP system, generate an Excel report, and then draft a summary email in Outlook—all from a single prompt. That scenario, which previously required three or four separate applications and manual data transfer, now happens almost instantaneously.

IT admins will also appreciate that the new features are manageable via Microsoft 365 admin center and Intune. They can enable specific skills, restrict certain connectors, and enforce data loss prevention policies—critical when dealing with sensitive financial data.

How It Compares to the Competition

Google Sheets has been gradually integrating generative AI through its Duet AI features, and specialized tools like Anaplan and Adaptive Insights have long offered cloud-based financial planning. However, Microsoft's advantage lies in Excel's ubiquitous install base and the depth of Copilot's integration with the entire Microsoft 365 suite.

Now that Copilot can natively understand financial schemas and connect to nearly any data source, it begins to level the playing field against dedicated corporate performance management (CPM) platforms. While those tools will still hold an edge for extremely complex, multi-dimensional planning, they often carry six-figure licensing costs. Copilot's new capabilities are included in the Microsoft 365 Copilot subscription, which many enterprises already license for productivity AI.

One area where competitors maintain a lead is real-time collaborative planning. Google Sheets allows multiple users to edit simultaneously with minimal latency. Excel on the web and desktop has improved, but it's still not as seamless. However, Microsoft is clearly betting that AI-assisted solo work will reduce the need for massive simultaneous editing sessions—Copilot becomes a force multiplier for the individual analyst.

Availability and Licensing

The new finance skills, connectors, and planning features are available starting June 25, 2026, to all Microsoft 365 Copilot subscribers. Specifically, they require a Microsoft 365 E3 or E5 license with the Copilot add-on, or a standalone Copilot for Microsoft 365 plan. A phased rollout means some users may see the features in their update channel within the first week, while others might wait up to a month.

The features are available both on the desktop Windows app (Version 2406 or later) and on Excel for the web. Mac users with Microsoft 365 subscriptions will also get access, though some advanced connector setup may be limited to the Windows version initially. Mobile apps will receive a preview of the skills by late summer 2026.

Microsoft has not announced any price increase for Copilot, though the CFO community will likely scrutinize whether the enhanced capabilities justify the existing $30 per user per month surcharge. Given the potential time savings in finance departments, the ROI story is compelling: even a 5% productivity boost for a financial analyst making $80,000 per year would break even on the Copilot license cost in just over one month.

Early Community Reaction

On Windows enthusiast forums and financial technology communities, the announcement has been met with a mix of excitement and cautious optimism. Many users recall the early days of Copilot in Excel, when hallucinations and formula errors made it unreliable for serious work. The introduction of traceability directly addresses that reliability concern.

"Finally, AI that shows its work," commented one user on the Windows News forum. "I can't take an AI forecast to my board and say 'trust me, the model said so.' Now I can show them the steps—game changer."

Others raised concerns about data sovereignty when using connectors that bridge on-premises ERP systems with cloud AI. Microsoft has emphasized that data processed via Copilot is handled according to existing enterprise data protection agreements, and that skills can run within the user's compliance boundary. But for highly regulated industries like banking and insurance, any AI that touches financial data will face intense scrutiny.

Some independent analysts note that the skills templates, while powerful, will likely require customization to fit each organization's unique chart of accounts and reporting standards. "Out-of-the-box, it'll do 80% of what a standard FP&A team needs," said one industry expert. "The last 20% will still need human tweaking, but that's a massive acceleration."

Looking Ahead

The June 2026 update is clearly not the endgame. Microsoft's roadmap suggests further verticalization: healthcare, retail, and manufacturing Copilot skills are likely in development. For Excel, the next frontier could be autonomous agents—Copilot not just suggesting or generating, but actually executing month-end close processes with appropriate human checkpoints.

Finance leaders should start preparing now. That means cleaning up data hierarchies, standardizing naming conventions across ERP systems, and training team members to shift from manual Excel jockeying to AI supervision. The value of these new tools scales directly with the quality of underlying data and the clarity of business logic.

One thing is certain: the spreadsheet isn't going anywhere. If anything, AI is making Excel more indispensable than ever. The real winners will be those who use Copilot not to replace analysis but to deepen it—turning hours of data wrangling into moments of insight.