Use Copilot in Excel to build your brackets
The matchups are finally set, and the annual question is back: how do you pick a bracket that’s fun and gives you a real shot at predicting the winner—whether you’re following the men’s tournament, the women’s tournament, or both? This year, you can use Copilot in Excel as your bracket sidekick—turning past tournament patterns into quick “what-if” scenarios, stress-testing upset paths, and sanity-checking your picks against historic data. Instead of manually building an analysis, copy/pasting data, and building multiple versions, you can ask in natural language and let Copilot build the analysis for you right inside an Excel workbook. Below are a few fast, practical ways to use Copilot in Excel to build a bracket workbook, explore upside picks (hello, Cinderella runs), and model “if this happens, then what?” outcomes so you can fill your bracket with more confidence than the rest of the group. 1) Set up a bracket workbook Open a new workbook, then open Copilot in Excel. Make sure “Edit with Copilot” is turned on. Start by asking Copilot to create a bracket template: “Create a 2026 [men’s or women’s] college basketball bracket including all the latest teams and seeds. Build dropdowns for each round so I can choose the winner of each matchup all the way to a champion, formatted like a standard bracket. For each dropdown, show only the 2 teams in the matchup based on the winners chose in previous rounds using helper columns." With that foundation in place, you’ve got a clean structure for picks and scenario assumptions. From here, you can make your picks and Copilot can help you add calculations, create “what-if” views, and summarize the implications of different upset paths. Bonus: Want to theme your brackets around your favorite team? First have Copilot generate a simple skills sheet and ask it to follow the instructions when creating brackets. “Create a skills sheet for my favorite team, [Team]. Include the official team colors (with hex codes), mascot/nickname, text colors, and conditional formatting rules for winners/losers.” 2) Stress-test your bracket with real-world scenarios Now for the part that can actually give you an edge: use Copilot to spin up scenario tabs and see how your bracket performs under outcomes that happen all the time in March—Cinderella runs, unexpected seed collapses, and “hot team” momentum that goes against conventional logic. Try some follow-up Copilot prompts like: Cinderella path: “Pick a 10–13 seed to reach the Sweet 16 based on past tournament frequency. Create a version to reflect that upset path, and show which higher-seeded teams I’m fading.” All the 1-seeds don’t make it: “Create a version of my bracket where at least two 1-seeds lose before the Elite Eight. Identify the earliest-round upsets needed and how my champion pick changes.” Favorite team: “Assume my favorite team is [Team]. Build two paths: (1) optimistic (reach the Final Four) and (2) realistic (based on projected path). For each path, show who they’d likely face by seed line and which matchup round matters most for them.” Momentum model: “Calculate a “momentum multiplier” using the conference tournament games and recent performance for each team and use that to fill out a version of my bracket weighted by momentum.” 3) Compare bracket variants and choose your entry Once you’ve built a few scenario versions, Copilot can help you compare them—so you’re not guessing which bracket is better, you’re choosing the one whose risk/reward matches your needs. Use a prompt like: “Create a comparison analysis for all my bracket scenarios in this workbook including charts. Include number of upset picks by round, and my top 5 most ‘contrarian’ picks across all my brackets. Give a recommendation for which to submit if I’m trying to win my bracket challenge or simply play it safe.” Copilot can generate the comparison table, highlight the key differences, and summarize the tradeoffs in plain language—so you can decide whether you want a safer entry, a balanced upset strategy, or a bold bracket designed to win big. Your turn: build your bracket with Copilot Ready to try it? Open Excel, start a new workbook, and use Edit with Copilot to create your brackets. Once you’ve got a bracket you like, share it with your league, your family, or your coworkers.Building Agent Mode in Excel
Excel is the world’s most trusted canvas for working with data, powering everything from household budgets to Fortune 500 companies, scientific research, operational planning, and classroom learning. It’s where millions turn to think, plan, and build. Agent Mode takes that impact even further, unlocking expert-level capabilities and making advanced analysis, modeling, and automation approachable for everyone, across every domain. Agent Mode lets you describe a task in natural language and then works with you to plan, reason, iterate, and validate the outcome. After introducing Copilot in Excel, it quickly became clear that our users wanted more — richer insights and more direct action on the sheet. Agent Mode aims to deliver on these expectations with a resilient experience that works across domains and data shapes, taking meaningful action directly in your workbook. We’ve developed Agent Mode to take advantage of the full richness of Excel artifacts, including table structures, formula syntax, dynamic arrays, PivotTables, charts, and more. It can create workbooks that are refreshable, auditable, and verifiable. This leap is powered by advances in our reasoning engine and the deeper expression of Excel as a rich modeling language. These breakthroughs allow Agent Mode to not only generate and execute solutions but also evaluate results, fix issues, and repeat the process until the outcome is verified. SpreadsheetBench instructions and obtained an accuracy rate of 57.2%. In our testing environment, Agent Mode makes direct workbook modifications via Excel APIs in a JavaScript runtime. We measure accuracy using the script provided by the SpreadsheetBench authors that grades output using the open-source openpyxl library. For evaluation on Claude and Shortcut.aI, we manually ran the SpreadsheetBench tasks (including answer location information needed for reliable evaluation) and downloaded the Excel files that were produced. These downloaded files were then graded using the same evaluation script provided by the SpreadsheetBench authors. Note that our evaluation with Claude completed on 895 of 912 instructions. Accuracy numbers were calculated using only completed tasks. All OpenAI benchmark results were originally published by OpenAI here. We measure Agent Mode on both our internal evaluation sets and the public SpreadsheetBench benchmark. Our results on SpreadsheetBench place Agent Mode at the leading edge of current systems, accurately completing 57.2% of the benchmark’s tasks. But we want to be clear: we don’t optimize for benchmarks, we optimize for real user jobs in Excel. That means solving messy, ambiguous, and complex tasks that reflect how people actually work. And while SpreadsheetBench is a strong signal, it doesn’t capture everything that makes Excel powerful — like dynamic arrays, PivotTables, charts, and formatting — or the customer need for refreshable, auditable, and verifiable solutions. That’s why we have also developed internal evaluation sets, AI grading, and user feedback loops to guide improvements. We also acknowledge that we have plenty of room for improvement, particularly around things like formatting and presentation-worthy layouts. But we believe our foundation is strong, and the direction is clear: Agent Mode is here to make Excel more powerful, more intuitive, and more helpful than ever before. Designing an Intelligent Spreadsheet Agent At the center of Agent Mode is a reasoning and reflection loop — powered by the latest generation of advanced reasoning models — that can interact directly with Excel workbooks. Rather than jumping straight into action, our system generates model-ready context from a given workbook and leverages an advanced reasoning model to begin planning for a given task. The system then interacts with the workbook by writing and executing code to carry out that plan, reflecting on the results, and evaluating whether the outcome matches the intent. If gaps remain, the loop continues: revising the strategy, pulling in additional context, and exploring alternative approaches. This cycle of planning, execution, and reflection continues until the system determines the task is complete. By combining planning with reactivity, the agent can chart a path, adjust when needed, and ultimately deliver solutions that feel intentional and well thought out. The reasoning engine of our system architecture is model-agnostic by design, allowing for rapid integration of new models as they become available. Loose coupling between our reasoning and workbook interaction layers allow us to quickly swap in and evaluate new models. Managing spreadsheet context Excel workbooks are living systems. They're often large, constantly changing, and filled with rich objects like PivotTables, slicers, and charts. For an agent, trying to absorb every detail all at once is simply impractical. Passing the entire dataset into context, along with the metadata for every object, would overwhelm any current model. Even exposing the thousands of read APIs Excel provides is far too heavy-handed. Instead, the agent approaches the workbook strategically: it pulls in just the pieces of context it needs, when it needs them, navigating the complexity step by step. This makes the agent not just a passive processor of data, but an active explorer of your workbook’s inner workings. To enable this selective exploration, we’ve developed a document context producer that operates within a coordinated push-and-pull system. On the push side, the document context producer proactively sends a compact “blueprint” of the workbook along with the user’s prompt — a summary of spatial layout, values, objects, and the formula dependency graph — encoded as JSON for complex objects and Markdown for tabular data. When deeper inspection is required, the reasoning engine can then request and pull additional information on demand, ensuring it can always operate with the context it needs. This hybrid design balances completeness with efficiency and lays the foundation for future improvements around caching, indexing, and search that will make context retrieval faster and more robust. Engineering domain knowledge of Excel Managing context gives the agent a clear view of the workbook. The next challenge is action: knowing which of Excel’s thousands of functions and APIs to call to get the job done. Excel spans thousands of API controls, including formulas, objects, and advanced features — a surface far too large for any current model to memorize or control directly. Instead of brute-forcing that complexity, we built distilled documentation into our reasoning engine — a compact, structured reference of Excel functions, objects, and specialized tool calls. Agent Mode can draw on this distilled knowledge to execute sophisticated tasks like building PivotTables, charts, slicers, and financial models. By embedding only the essential information, the model gains expert-level fluency in Excel’s internal workings without overwhelming its context window, enabling accurate reasoning across the full feature set of the application. Validation-driven generation In developing and evaluating our core coding and reflection loop, we observed that many spreadsheet errors are silent — formulas return values, but subtle mistakes remain hidden until they cascade into bad analysis. Relying on a single execution step is risky when the goal is trustworthy automation. To counter this, Agent mode in Excel reframes each tool call as an auditable, verifiable workflow. Before executing an action, our reasoning engine first generates lightweight tests to establish expected outcomes. These checks act as verifiable guardrails, ensuring that each step can be inspected and reproduced. Crucially, rather than hardcoding values, Agent Mode carries out all computations directly on the grid. This preserves the full dependency structure of the spreadsheet, allowing users to audit intermediate results, trace formulas, and verify correctness at every stage. Across our quantitative evaluations, we have been able to drive double-digit accuracy improvements with this validation-infused approach. Scaling quality with AI graders As we evolve Agent Mode into a deeply integrated, context-aware companion for data workflows, AI graders have emerged as one of the most critical technical enablers driving quality, trust, and usability. They serve not only as evaluators of accuracy but also as definers of excellence—ensuring that results are not just correct, but also useful, complete, relevant, and delightful. Graders are the mechanism through which we translate abstract quality goals into measurable, actionable standards. In Agent mode, they underpin both offline evaluation pipelines and live user experience metrics, helping us answer key questions like: Did Agent Mode fulfill the user’s intent? Was the output accurate and verifiable? Did the result feel native to Excel? Was the experience satisfying and accessible? Without graders, we would risk optimizing for superficial metrics — like response time or token count—while missing the deeper signals of user success. Looking ahead An early preview of Agent Mode in Excel is available starting today via the Frontier program for Microsoft 365 Copilot licensed customers and Microsoft 365 Personal, Family, or Premium subscribers (under the Microsoft Services Agreement). Agent Mode works in Excel on the web and is coming soon to desktop. To try it, look for Agent Mode in the Tools menu of Copilot in Excel. Learn more about it in our announcement blog. This preview is just the beginning of our journey. We’re continuing to build a complete, M365 integrated experience that is trustworthy, reliable, and transparent — one that you can depend on for critical work. And from a developer perspective, we’re exploring extensibility solutions that would allow customers and partners to build custom solutions on top of our Agent Mode capabilities. Over the coming weeks and months, we plan to fully integrate and iterate on this experience across all Excel clients. We’ll continue to improve core output quality, refine the Agent Mode interfaces in chat and on the grid, and incorporate user feedback to ensure the experience feels at home in Excel, while unlocking entirely new ways to model, analyze, and automate.What's New in Excel (February 2026)
Welcome to the February 2026 update. This month we are excited to announce expanded availability for Agent Mode in Excel, as well that you can now query modern Excel workbooks (like .xlsx, .xlsb, .xlsm, .ods) stored locally on your device using Microsoft 365 Copilot Chat, on Windows and Mac. In addition, we've heard your feedback that working across multiple Copilot entry points can feel fragmented; and to address this, the editing capabilities that App Skills provided will be integrated into Copilot Chat and Agent Mode in Excel, which became generally available earlier this year. Click here to read more > Excel for Windows and Mac: - Agent Mode in Excel expanded availability - Query your local Excel files with Copilot Chat Excel for Windows and Mac Agent Mode in Excel expanded availability Agent Mode in Excel is now also available for Copilot in Excel users in the EU, including Current Channel and Monthly Enterprise Channel. Read more here > Query your local Excel files with Copilot Chat Copilot in Excel now works with locally stored modern workbooks. This gives users faster, more consistent assistance across all their files, improving productivity without requiring changes to how workbooks are stored. Previously, insights and analysis from Copilot Chat were limited to Excel workbooks stored in the cloud. With this new feature, analyzing your locally saved Excel workbooks with Copilot Chat makes it possible to stay productive even when you’re offline. This feature is currently rolling out on Windows and Mac. Read more here > Check if a specific feature is in your version of Excel Click here to open in a new browser tab Many of these features are the result of your feedback. THANK YOU! Your continued Feedback in Action (#FIA) helps improve Excel for everyone. Please let us know how you like a particular feature and what we can improve upon—"Give a compliment" or "Make a suggestion".. You can also submit new ideas or vote for other ideas via Microsoft Feedback. Subscribe to our Excel Blog and the Insiders Blog to get the latest updates. Stay connected with us and other Excel fans around the world – join our Excel Community and follow us on X, formerly Twitter. Special thanks to our Excel MVPs David Benaim, Bill Jelen, Alan Murray, and John Michaloudis for their contribution to this month's What's New in Excel article. David publishes weekly YouTube videos and regular LinkedIn posts about the latest innovations in Excel and more. Bill is the founder and host of MrExcel.com and the author of several books about Excel. Alan is an Excel trainer, author and speaker, best known for his blog computergaga.com and YouTube channel with the same name. John is the Founder & Chief Inspirational Officer at MyExcelOnline.com where he passionately teaches thousands of professionals how to use Excel to stand out from the crowd.App Skills is evolving with Copilot in Excel
We are continuing to streamline the ways people engage with Copilot in Excel, and we've heard your feedback that working across multiple Copilot entry points can feel fragmented. To address this, Copilot in Excel is transitioning toward a more unified experience so users can easily choose between conversational assistance and direct editing capabilities. As part of this effort, the editing capabilities that App Skills provided will be integrated into Copilot Chat and Agent Mode in Excel, which became generally available earlier this year. This update is part of our broader work to simplify the Copilot entry points and make it clearer how to interact with Copilot depending on the task. We know that when you find a workflow that works, change can feel disruptive. That's why we want to give you a clear picture of what's evolving in how you use Copilot in Excel and how these changes provide an even better experience. Why this change matters We've heard your feedback about wanting a more capable experience when working with Copilot inside the spreadsheet. The enhanced editing experience that Agent Mode introduced was built specifically to address this. It's designed to handle more complex requests, work across multiple steps, and give you greater control over how Copilot edits your workbooks. Rather than maintaining separate experiences that can feel fragmented, we're bringing everything together so you have: More power: Copilot with Agent Mode can now handle complex, multi-step reasoning tasks that go beyond what App Skills could do. Better clarity: You'll know exactly where to go depending on what you need—direct editing happens with Agent Mode; quick answers and simple actions work best in Copilot Chat. Continued innovation: By focusing on improvements toward a unified experience, we can deliver new capabilities faster. As part of this, we also plan to better integrate Agent Mode’s editing capabilities into the Copilot Chat experience in the coming months. What’s changing? App Skills entry points in Excel are going away. You will no longer find an App Skills button in the ribbon or be able to use App Skills from the context menu. Existing skills are consolidating into Copilot Chat and Agent Mode. When you want Copilot to make changes directly in your workbook, Agent Mode is designed to support many core editing tasks, such as creating or updating tables, applying formatting, or generating charts and PivotTables. Copilot Chat remains available for tasks that don’t require modifying content, such as interpretation or exploration of your data and using agents like Analyst. If you can open the App Skills chat pane and submit a prompt, you may receive an error message instead of a response. During this transition, some users may still see App Skills entry points for a short time. In some cases, opening the App Skills pane may result in an error message indicating that App Skills is no longer available. If this happens, you can continue your workflow using Agent Mode or Copilot Chat based on the type of assistance you need. When is it happening? This update is rolling out now. Depending on your Excel version, you may see the App Skills entry point up until the end of February. App Skill scenarios not yet available Certain scenarios that previously used App Skills—specifically the Advanced Analysis mode that used Python in Excel and advanced text analysis capabilities—are not yet available within Copilot Chat or Agent Mode. We are continuing to expand support for these capabilities in Copilot Chat and Agent Mode—watch for updates as these become available over time. Note: this was originally communicated to commercial customers via the M365 Message Center (MC1184407) on November 10, 2025.Agent Mode in Excel is now generally available on desktop
Agent Mode in Excel, part of Microsoft 365 Copilot, is now generally available on Windows, with Mac rolling out over the coming days — extending access beyond Excel for the web, which launched in December. Since our initial public preview, we’ve expanded availability, added web-grounded search, and introduced a new multi-model reasoning system that allows customers to choose between OpenAI and Anthropic models. Under the hood, we’ve significantly improved task success, performance, and reliability across core Excel scenarios, including workbook creation, formula repair, and chart and PivotTable generation. Evolving Copilot to become an active collaborator Excel is where people think with data. It’s where budgets, forecasts, and operating plans take shape, and where decisions get made. Agent Mode turns Copilot into a true partner in that work, able to take your goals, plan next steps, act directly in your workbook, iterate, and validate outcomes. With today’s release, that capability is now available in the desktop versions of Excel customers rely on every day. What’s improved during public preview We launched Agent Mode in public preview to validate the experience with real Excel customers working on real workloads. Our goal has always been to support the way Excel is actually used inside organizations — messy data, ambiguous goals, and multi-step workflows that need to be refreshable, auditable, and verifiable. Throughout the preview, we continued to invest in this by introducing: Expanded availability: Agent Mode now works across Excel for the web, Excel for Windows, and Excel for Mac, so you can leverage its power no matter where your access your work. It’s integrated directly into Copilot in Excel, and we’ll continue to improve the experience in the coming weeks and months. Integrated web search: Instantly get up-to-date information with source citations, perfect for “pull in the latest data” scenarios. Model choice: A new model switcher lets you choose between our OpenAI-powered experience and the latest Claude models from Anthropic, so you can try different AI approaches. Model choice We’ve learned — alongside our colleagues at GitHub — that different reasoning models excel at different kinds of work. Some are better suited for fast, structured problem-solving while others shine when tasks require explanation, iteration, or more open-ended reasoning. Our goal is to build an intelligent system that can make the right model choices on your behalf. At the same time, we believe customers should have visibility and control — so Agent Mode also gives you the ability to explicitly choose the model you’d like to apply to your task. new model switcher choices supported. When in the default Auto mode, Copilot will attempt to choose the best model for you. You can also choose one of the specific models before running a prompt. The latest models from OpenAI (GPT 5.2) and Anthropic (Claude Opus 4.5) are available today for Microsoft 365 Copilot and Microsoft 365 Premium licenses. Learn more here about using Claude with Agent Mode in Excel and Anthropic as a subprocessor. How to try it Open Excel on Web, Windows or Mac. Open Copilot and select Agent Mode from the Tools menu. Start with an outcome-based prompt, like “Build a loan calculator that computes monthly payments based on user inputs for loan amount, annual interest rate, and term in years. Generate a schedule showing month, payment, principal, interest, and remaining balance. Present the results in a clear, formatted table.” Availability Agent Mode in Excel is generally available today across Excel for web, Windows, and Mac for commercial Microsoft 365 Copilot licenses and Microsoft 365 Personal, Family, and Premium subscribers. Note that Personal and Family subscriptions use an AI credit model and Agent Mode in Excel is not yet available to customers in the EU or UK. For more on availability and access, check out Agent Mode in Excel.Excel as a Literate Computation Surface for AI
[This article was originally published by Sumit Chauhan on LinkedIn.] AI systems perform robust computation, but their outputs are typically dissociated from the structure of the computation itself. Answers are delivered as fluent summaries, scripts, or static artifacts. Explanations may accompany results, but the execution path remains opaque and nonexecutable. This separation constrains inspection, audit, and collaborative verification. Excel bridges this separation through a longstanding but underappreciated design property: computation and explanation coexist in the grid. Values persist as first class objects, accessed and connected through a network of formulas and calculation objects. The dependency structure is explicit and intermediate results remain live within the model. Assumptions remain live inputs rather than fixed premises embedded in prose. A spreadsheet is therefore a runnable representation of reasoning. The Excel Agent extends this property to AI-driven analysis. Instead of returning an answer, it writes computation directly into spreadsheet primitives: cells, formulas, tables, and references. Analytical intent is encoded structurally—not narratively—resulting in computational instantiation that is inspectable, addressable, and mechanically verifiable. Respecting established analytical practice This distinction becomes explicit when updating real analytical models. In one internal evaluation, the Excel Agent was prompted to update an existing public company financial model following newly released quarterly results. The spreadsheet contained a structured income statement, linked calculations, margin rows, and derived percentage outputs sourced exclusively from GAAP financial statements. We observed that the Excel Agent pulled precise reported figures from the GAAP filings and updated only the newly available quarterly actuals, leaving guidance untouched. The agent preserved the existing model structure, row ordering, and dependency relationships. Calculated fields—totals, margins, growth rates—were not overwritten. Instead, they recalculated mechanically from updated inputs. Number formatting remained intact, preserving distinctions between dollar values and percentages. Changes were immediately auditable by inspection: updated cells represented inputs, while formulas remained unchanged. The contrast to general purpose AI tools is instructive. In parallel tests, a comparable update performed through a chat based model relied on headline summaries rather than reported figures, conflated guidance with actuals, overwrote calculated fields with inferred values, and introduced silent formatting errors. The resulting spreadsheet appeared plausible but was structurally compromised. The result was incomplete and not auditable. Inspectability and error localization Because Excel externalizes reasoning as explicit dependency graphs, errors localize narrowly. A disagreement targets a specific cell, formula, or source assumption rather than an opaque explanation. Review is incremental. Validators can inspect references, confirm lineage, and trace downstream effects mechanically. Excel Agent inherits these affordances. Its output is not sealed; it invites modification. Alternative scenarios become addressable inputs rather than regenerated answers. This changes the cost structure of verification. Inspection becomes cheaper than regeneration. Corrections are edits, not prompts. Trust derives from structure rather than narrative coherence. Durability through artifact-centered collaboration Spreadsheets are durable analytical objects. They are shared, versioned, reviewed, audited, and revisited independent of their origin. The Excel Agent produces artifacts that persist beyond the original interaction. The analysis remains runnable months later without replaying a model, and knowledge accumulates as structured computation rather than transient output. This reframes AI’s role in analytical work. Excel Agent does not replace analytical judgment. It relocates reasoning into a medium designed for inspection, modification, and reuse. AI output becomes a starting point rather than an endpoint, expressed as runnable structure rather than fluency. The value is not in producing convincing answers, but in creating durable, collaborative, inspectable computation that can be reviewed, extended, and trusted over time. The shift is from answers to artifacts, and opaque intelligence to shared reasoning. Sumit Chauhan Corporate Vice President, Office Product Group
