Microsoft is giving Excel users a potent new set of AI capabilities as it rolls out finance-focused skills, partner-driven workflow extensions, and deep data integrations to Copilot in Excel, the company disclosed in a late June 2026 update for Microsoft 365 Copilot subscribers. The enhancements transform the already capable spreadsheet assistant into a specialized tool that can handle complex financial analysis, connect directly to business data sources, and enforce enterprise-grade governance—all through natural language commands.
For the millions of accountants, financial analysts, and business planners who live inside Excel, the upgrades address a long-standing gap: until now, Copilot’s general-purpose AI was handy for formula writing and charting, but lacked the contextual understanding and connectivity needed for rigorous finance work. The new release changes that by embedding domain-specific intelligence and secure data access directly into the familiar grid.
Finance skills come natively to Copilot
At the core of the announcement is a suite of finance-oriented AI skills that allow users to ask questions and issue commands in plain English—or dozens of other supported languages—and receive accurate, audit-ready results. Instead of painstakingly building complex models, users can now instruct Copilot to “forecast quarterly revenue based on historical trends and account for seasonality” or “perform a variance analysis on departmental budgets versus actuals and highlight anomalies.”
Early demonstrations suggest that the AI draws on a blend of classical statistical methods and machine learning models, all executed within Excel’s calculation engine. It can generate cash flow projections, simulate what-if scenarios, and even propose journal entry adjustments that comply with GAAP or IFRS standards. For finance teams, this means routine tasks like consolidating P&L statements from multiple sheets or calculating weighted average cost of capital can be reduced to a few conversational prompts.
Crucially, the finance skills are not isolated functions. They sit alongside Excel’s existing Copilot capabilities—formula suggestions, data visualization, Python-powered analytics—creating a seamless experience. A user can ask Copilot to pull in live sales data, generate a rolling forecast, identify cost drivers, and then produce an executive-summary slide deck, all without leaving the application. The AI respects Excel’s security model, meaning sensitive financial data remains within the tenant’s compliance boundary.
Industry observers note that this moves Excel closer to becoming a complete financial analysis hub, potentially displacing specialized tools for certain mid-market use cases. “What Microsoft is doing is embedding the kind of FP&A (financial planning and analysis) logic that used to require separate software,” said one enterprise architect in a community discussion thread about the update. “If Copilot can handle driver-based budgeting and sensitivity analysis out of the box, that’s a huge win for small to medium finance departments.”
Partner-built workflows extend Excel’s reach
Perhaps the most strategic element of the update is the introduction of partner-built workflows—custom extensions created by third-party software vendors that integrate their services directly within the Copilot experience. Microsoft is leveraging the Copilot extensibility platform, launched earlier in 2025, to allow ISVs to develop “skills” that hook into their own back-end systems.
In practice, this means a controller working in Excel could type “fetch our AP aging report from SAP and flag overdue invoices” and Copilot would call a partner-built connector that reaches into the SAP system, retrieves the data, and displays it in a structured table. Similarly, a tax analyst could ask Copilot to “calculate deferred tax liability using Thomson Reuters ONESOURCE rates,” and the tool would invoke the partner’s tax engine to return the computation.
The partner ecosystem is expected to cover popular finance platforms—ERPs like Microsoft Dynamics 365, SAP, Oracle NetSuite, and Workday Financials, as well as specialist tools for treasury, tax, and risk management. Microsoft says it has worked closely with a group of launch partners but has not yet disclosed the full list. However, the early preview suggests that any partner with a supported API can build a Copilot connector, subject to Microsoft’s security and certification process.
This move not only enriches Excel’s native capabilities but also positions Copilot as a central orchestration layer for financial workflows. Instead of toggling between half a dozen applications, a finance professional can conduct an entire closing process—from data aggregation to variance explanation to sign-off—within a single Excel window. Embedding these workflows directly in Copilot also reduces the risk of errors that arise from manual copy-paste and increases the speed of financial close cycles.
Direct data integrations eliminate middlemen
The third pillar of the June 2026 update is a revamped direct data integration model that gives Copilot the ability to query external databases and cloud services without requiring users to first build Power Query connections or ODBC imports. Starting with a curated set of Microsoft sources—including Azure SQL Database, Dynamics 365 Finance, and the Microsoft Fabric lakehouse—and expanding to third-party data providers via the partner framework, Copilot can now access live data on command.
Analysts can ask natural-language questions that span multiple data systems. For instance, “compare this month’s CRM pipeline value against the actual revenue recorded in our ERP and calculate the conversion rate” would trigger Copilot to fetch real-time data from both Dynamics 365 Sales and the finance module, then run the computation. The results appear in a dynamic array that can be refreshed with a single click.
Under the hood, Microsoft has upgraded Excel’s data engine to handle authenticated, on-the-fly queries while respecting row-level security and data masking policies. IT administrators retain control through the Microsoft 365 admin center, where they can define which data sources Copilot is permitted to query and set rate limits to avoid excessive consumption of database resources.
This direct connectivity is especially significant for enterprise customers who have been hesitant to use AI tools that require duplicating data into Excel files. Now, the AI can operate against governed, server-side data, maintaining a single source of truth. Data sensitivity labels are preserved, and all queries are logged for audit purposes, satisfying compliance teams that every access is traceable.
Governance features put IT in the driver’s seat
With any AI-powered tool that handles financial data, governance is non-negotiable. Recognizing this, Microsoft is delivering a robust set of administrative controls alongside the new Copilot features. Key among them is the ability for Microsoft 365 Copilot administrators to define prompt policies—rules that constrain what Copilot can do with certain types of data.
For example, a policy could prevent Copilot from generating forecasts that incorporate data columns tagged as “personally identifiable information,” or block any prompt that tries to export financial figures to an external service. Organizations can also require that certain Copilot-generated analyses be reviewed and signed off by a designated approver before they are shared outside the finance team.
Audit logs capture each Copilot interaction, recording the exact prompt, the resulting action, and any data sources accessed. This creates a forensic trail that can be ingested by SIEM tools for security monitoring and is available for internal or external audits. Sensitive spreadsheet cells can be excluded from Copilot’s context window, ensuring the AI never sees, say, salary figures or board-level confidential notes.
The governance framework integrates with Microsoft Purview, allowing data classification labels and retention policies to extend to Copilot-generated artifacts. If a user asks Copilot to create a forecast that incorporates data from a sensitivity-labeled document, the output automatically inherits the highest classification. These controls give chief financial officers and chief information security officers the confidence to roll out AI assistance broadly.
Availability and platform support
Microsoft confirmed that the new finance skills, partner workflows, and data integrations will be available to Microsoft 365 Copilot subscribers across Excel on Windows, Mac, and the web. The features will begin reaching first-release tenants in early July 2026, with a broader rollout expected to complete by the end of the quarter. Some partner-built workflows may require an additional license from the third-party vendor, but the core Copilot functionality is included in the Microsoft 365 Copilot subscription.
The company also noted that the updates apply to both the consumer and commercial SKUs of Excel, though the governance capabilities are enterprise-only and require an E3 or E5 license along with the Copilot add-on. Education and frontline worker plans will see a subset of the features at a later date.
Community reaction and early expectations
On windowsforum, where the announcement first appeared in a thread titled “Copilot in Excel Adds Finance Skills, Data Integrations, and Governance,” users expressed a mixture of enthusiasm and pragmatic concern. One commenter wrote, “Finally, Excel is getting the AI it deserves. I’ve been waiting for native forecasting without having to learn Python.” Another pointed out the potential cost: “If every partner workflow requires its own subscription, this could get pricey fast.”
The biggest point of contention in the discussion surrounded governance. A finance manager weighed in: “The audit log feature is great in theory, but will it capture downstream actions like when someone copies a Copilot-generated output into an email? Our compliance people will ask these questions.” Microsoft’s documentation on precisely how audit trails follow data beyond the Excel file is still sparse, leaving some enterprise adopters cautious.
There was also notable excitement about the direct database integrations. A data engineer shared, “We’ve been building custom connectors for years. If Copilot can talk to our Azure SQL warehouse with just a prompt, that’s going to save our analysts hours of manual work every week. But I want to see how it handles complex schemas with hundreds of tables.”
The broader Copilot strategy
These Excel enhancements fit into a larger pattern unfolding across the Microsoft 365 suite. Throughout 2025 and into 2026, Redmond has steadily been weaving domain-specific AI into each core productivity app—Copilot in Word gained legal document drafting, Copilot in PowerPoint got brand-aware design tools, and Copilot in Teams now ingests meeting transcripts to auto-generate project status updates. Excel, with its unmatched base of business users, was always the ultimate prize.
By adding finance as a vertical, Microsoft is leveraging its installed base of over 1.2 billion Office users and the troves of financial data their organizations produce daily. The partner-workflow model also positions Excel as the interface of choice for financial data, a direct challenge to dedicated FP&A platforms like Anaplan, Adaptive Insights, and Vena Solutions. Whether those platforms respond by building their own Copilot connectors or doubling down on standalone AI is an open question, but Microsoft’s pivot clearly raises the stakes.
Industry analysts predict that the next phase will see Copilot in Excel tackling regulatory compliance and ESG reporting, pulling data from sustainability tracking systems and automatically generating mandated disclosures. Given the governance groundwork now in place, such an expansion would be a logical next step.
What this means for Excel users
For the everyday financial professional, the immediate impact is a noticeable reduction in the friction that eats up hours of each workday: no more exporting data, re-importing, and manually updating spreadsheets only to redo the work when source data changes. Copilot’s ability to connect to live systems and refresh analyses on demand turns Excel from a static canvas into a dynamic command center.
Training requirements will be minimal. Because the interface is conversational, even junior analysts can perform sophisticated tasks by describing what they need in natural language. This could democratize financial analysis across organizations, letting marketing managers run profitability scenarios or HR leads model workforce costs without waiting for the finance team’s availability.
Yet, with that power comes responsibility. The success of these features will hinge on how well organizations can implement the governance controls and educate users on the boundaries of what Copilot can—and should—do. The risk of “shadow AI” where employees bypass approved workflows is real, which is why Microsoft has built the administrative guardrails in parallel with the flashy new skills.
Looking ahead
Microsoft has stated that it will continue to iterate on Copilot in Excel based on feedback from the initial wave of finance users. Roadmap hints suggest upcoming support for real-time streaming data—think stock tickers or IoT sensor feeds—and the ability to chain multiple Copilot commands into automated “finance scripts” that run on a schedule.
The June 2026 release cements Excel’s evolution from a simple spreadsheet program to an AI-augmented analytical engine. For the millions of people who rely on it daily to make million-dollar decisions, that’s a transformation worth watching closely.