Copilot in Excel Review: Prompt-driven data analysis and modeling built directly into Microsoft Excel

Microsoft is a global technology provider that has embedded its generative AI assistant, Copilot in Excel, directly into its spreadsheet software as a paid add-on. For commercial real estate analysts, Excel remains the undisputed standard for financial modeling, rent roll analysis, and cash flow forecasting. By integrating a natural language interface into this environment, Microsoft targets a primary use case of data analysis, modeling, and charting with prompts. As a Tier 2 CRE-Adjacent tool, it does not bring proprietary property data, market comparables, or specialized real estate algorithms to the table. Instead, it relies entirely on the raw data the user provides within the active workbook.

The BestCRE evaluation treats this tool with specific skepticism regarding its utility for complex, multi-layered real estate financial models. While the premise of typing a prompt to analyze a 500-tenant rent roll is highly appealing, the reality of execution requires close examination. Analysts evaluate acquisitions and developments using highly specific, interconnected schedules for debt, depreciation, and waterfall distributions. Copilot in Excel must navigate these specialized structures without any native understanding of commercial real estate fundamentals. Our analysis focuses on whether this assistant accelerates the underwriting process or creates additional auditing work for the analyst who must verify every generated formula. In August 2026, the demand for efficiency in underwriting is high, but the tolerance for calculation errors remains zero. Copilot operates as a generalist assistant attempting to perform specialist tasks, requiring users to maintain strict oversight over its outputs.

What Copilot in Excel does and how it works

Copilot in Excel functions as a persistent chat interface located in the right-hand sidebar of the application. Users interact with their data by typing natural language commands rather than navigating ribbon menus or writing syntax. When an analyst opens a property operating statement or a rent roll, they can ask the assistant to perform specific actions. The primary mechanics include generating formula columns, highlighting specific data points, and summarizing large datasets into pivot tables or charts. For example, a user can prompt the tool to calculate the expiration exposure per year based on lease end dates, and Copilot will generate the necessary formulas and insert a new column or summary table.

The tool requires data to be formatted as an Excel Table to function at its highest capacity. Once a table is defined, the assistant can filter, sort, and apply conditional formatting based on text prompts. If an asset manager wants to identify all retail tenants with a base rent below a certain threshold and lease expirations within the next twenty-four months, they type this request into the sidebar. Copilot processes the request, applies the filters, and can generate an accompanying bar chart to visualize the exposure. It also includes a feature that explains existing formulas, which assists junior analysts in understanding inherited, complex underwriting models built by previous team members.

Crucially, Copilot does not autonomously alter the fundamental structure of a multi-tab financial model without user confirmation. It operates on a proposal system; when asked to analyze data or create a calculation, it presents a preview of the formula or chart in the chat pane. The user must explicitly click to apply the change to the grid. This design choice prevents the assistant from overwriting critical assumptions or breaking interconnected cell references across debt schedules and cash flow tabs. However, it also means the tool acts more as an interactive formula builder than an autonomous financial modeler.

9AI Framework: the score, dimension by dimension

CRE Relevance — 3/10

Copilot in Excel is a general-purpose application designed for every industry from retail to healthcare. It contains no native commercial real estate intelligence, property databases, or specialized underwriting templates. The tool does not know the difference between a triple net lease and a gross lease unless the user explicitly defines those terms mathematically within the spreadsheet. As a Tier 2 CRE-Adjacent tool, its relevance is derived entirely from the fact that commercial real estate professionals use Excel as their primary analytical canvas. It provides the mechanical functions to manipulate data, but the user must supply 100% of the real estate logic. In practice: Analysts must instruct the tool with highly specific, mathematically precise prompts because it lacks the industry context to infer real estate concepts.

Data Quality and Sources — 6/10

Because Microsoft does not supply external market data or property comparables through this specific interface, the data quality dimension rests entirely on the user’s internal hygiene. Copilot simply processes the numbers present in the workbook. If a rent roll contains inconsistent date formats, misspelled tenant names, or blank fields, the assistant will struggle to aggregate the information accurately. It can assist in data cleaning by identifying outliers or standardizing text strings, but it cannot verify if a tenant’s square footage matches the physical lease document. The output is strictly bound by the integrity of the uploaded file. In practice: Users must perform standard data validation and cleaning on their rent rolls before expecting the assistant to generate accurate pivot tables or charts.

Ease of Adoption — 9/10

Integration into the daily workflow is exceptionally straightforward because the interface exists within an application that analysts already keep open all day. There is no new software to install, no separate login portal to navigate, and no need to export data to a third-party platform. The chat window uses standard conversational language, removing the steep learning curve traditionally associated with mastering complex nested formulas, VBA macros, or Power Query. Users can begin testing prompts immediately upon activation. However, learning to write effective prompts that yield the correct mathematical operations requires a brief period of trial and error. In practice: A junior analyst can start using the tool within five minutes of activation, though mastering the specific phrasing required for complex financial modeling takes several weeks.

Output Accuracy — 7/10

The accuracy of the generated formulas and data summaries is highly variable depending on the complexity of the prompt. For standard tasks like calculating weighted averages, extracting years from date strings, or creating basic pivot tables, the assistant performs reliably. However, when asked to construct multi-tiered waterfall distribution formulas or complex debt amortization schedules with dynamic interest rates, the tool frequently proposes incorrect cell references or misinterprets the order of operations. The preview function is critical here, as it forces the user to audit the syntax before application. It is prone to standard generative AI hallucinations if instructions are ambiguous. In practice: Senior principals must enforce strict review protocols because the assistant will confidently generate mathematically flawed formulas if the prompt lacks structural precision.

Integration and Workflow Fit — 9/10

As a native component of the Microsoft 365 ecosystem, the integration fit is structurally flawless. It operates within the boundaries of the organization’s existing security, compliance, and data governance policies. For commercial real estate firms already reliant on OneDrive, SharePoint, and Teams, this tool inherits all established permissions. It does not require API keys or complex data mapping to function alongside existing files. However, it does not natively integrate with specialized CRE platforms like Argus, VTS, or Yardi unless that data is first exported into an Excel format. It is confined to the Microsoft environment. In practice: IT departments approve this tool quickly because it bypasses the security audits typically required for third-party AI platforms that extract data to external servers.

Pricing Transparency — 9/10

Microsoft maintains high visibility regarding the cost of its enterprise tools. The pricing details are public and structured as a paid add-on to existing Microsoft 365 subscriptions, typically requiring an annual commitment per user. While the exact tiering can vary based on enterprise agreements, the baseline cost is clearly documented on their commercial web properties. There are no hidden consumption metrics, token limits, or surprise overage charges based on how many formulas an analyst generates in a given month. This predictable flat-rate model allows real estate firms to budget accurately across their acquisitions and asset management teams without monitoring daily usage. In practice: Chief Financial Officers can model the exact annual cost per analyst without worrying about variable billing based on prompt frequency or model complexity.

Support and Reliability — 9/10

Operating under the umbrella of one of the largest software companies globally, the support infrastructure is highly dependable. Uptime is tied to the general availability of Microsoft 365 services, which rarely experience prolonged regional outages. Users have access to extensive documentation, community forums, and enterprise-grade technical support channels. When the assistant encounters an error, it typically fails gracefully by stating it cannot complete the task, rather than crashing the workbook. However, support personnel will assist with software functionality, not with debugging the specific real estate financial logic of a user’s broken underwriting model. In practice: Firms can rely on the tool being online and functional during critical deal hours, backed by standard enterprise service level agreements.

Innovation and Roadmap — 8/10

Microsoft pushes updates to its AI assistants at a rapid pace, frequently expanding the context window and improving the underlying large language model. The development trajectory points toward deeper integration with Python in Excel and enhanced capabilities in handling massive datasets without performance degradation. While the roadmap is aggressive, it is entirely focused on general productivity enhancements rather than industry-specific features. Commercial real estate users will benefit from broader improvements in reasoning and formula generation, but they should not expect updates tailored to property valuation or lease administration. The feature set will evolve to handle more complex cross-workbook referencing. In practice: Analysts will see continuous, incremental improvements in the tool’s ability to interpret complex logic, driven by global enterprise feedback rather than real estate specific requests.

Market Reputation — 10/10

The reputation of Microsoft’s AI initiatives is dominant across the corporate landscape. In the commercial real estate sector, adoption is viewed as a safe, conservative choice for firms hesitant to upload proprietary deal data to unproven startups. While specialized AI tools like Agentforce or Conduit generate excitement for their specific workflows, Copilot is seen as the baseline utility that every firm will eventually adopt. It is widely regarded as a productivity enhancer rather than a replacement for human analysts. Skepticism remains regarding its ability to handle institutional-grade financial modeling, but its reputation for basic data manipulation is strong. In practice: Investment committees and clients view the use of this tool as a standard modernization of the tech stack, carrying none of the reputational risk associated with shadow IT.

Who should use Copilot in Excel

The primary audience for this assistant consists of professionals who spend significant portions of their day manipulating raw data exports, formatting tables, and building summary charts. It serves as a mechanical accelerator for routine spreadsheet tasks rather than a strategic underwriting partner.

• Asset managers who frequently download raw rent rolls from property management systems and need to quickly format, filter, and summarize tenant exposures.
• Junior analysts who are learning complex Excel syntax and benefit from the tool’s ability to explain existing formulas inherited in legacy underwriting models.
• Research directors who aggregate demographic or market data into large tables and require rapid generation of pivot tables and charts for investment committee memos.
• Operations teams managing large portfolios who need to standardize text formatting and clean up inconsistent data entries across multiple property reports.

Who should look elsewhere

Professionals seeking an autonomous financial modeler or a tool that understands the nuances of commercial real estate valuation will find this assistant inadequate. It cannot replace the domain expertise required to structure complex capital stacks or interpret lease clauses.

• Acquisitions directors looking for an AI to automatically underwrite a property from a raw offering memorandum without manual spreadsheet setup.
• Principals who require an AI to audit complex, multi-tiered waterfall distribution models for mathematical accuracy, as the tool itself is prone to logic errors.
• Lease administrators who need an AI to read PDF lease documents and automatically populate an Argus file or Excel rent roll.

Pricing and ROI

Microsoft publishes its pricing for this tool as a paid add-on to existing enterprise licenses. Currently, Microsoft 365 Copilot requires an additional subscription, typically priced at $30 per user per month, billed annually. This requires the firm to already hold qualifying base licenses, such as Microsoft 365 Business Standard or Enterprise E3/E5. The pricing transparency is high, with no hidden consumption tokens or variable costs based on the complexity of the financial models processed. For a commercial real estate firm evaluating the return on investment, the math is straightforward. At roughly $360 per year per analyst, the tool only needs to save a user a few hours of manual data formatting or formula troubleshooting annually to justify the expense. If an asset manager saves twenty minutes a week generating pivot tables from Yardi rent roll exports, the tool yields a positive financial return. However, firms should be cautious about deploying licenses universally across the organization. The return on investment diminishes rapidly for executives or brokers who primarily use Excel for simple lists rather than deep data analysis. Procurement teams should target licenses specifically for the acquisitions, asset management, and research departments where spreadsheet manipulation is a daily requirement.

Integration and CRE tech stack fit

The integration fit within a standard commercial real estate technology stack is highly dependent on the firm’s reliance on the Microsoft ecosystem. For organizations operating entirely within SharePoint, OneDrive, and Teams, this assistant integrates natively without any complex API configurations or third-party middleware. It respects existing file permissions and enterprise data loss prevention policies, ensuring that sensitive acquisition models or partner distribution schedules remain secure. However, its integration with specialized CRE software is entirely manual. It does not connect directly to platforms like VTS, Dealpath, or Yardi. Analysts must first export data from these systems into a CSV or Excel format before the assistant can interact with it. Furthermore, the tool operates strictly within the active workbook; it currently struggles to analyze data across multiple linked workbooks, a common structure in institutional portfolio modeling. While it fits perfectly into the general productivity stack, it remains isolated from the specialized data pipelines that drive modern real estate operations.

Competitive landscape

When evaluating alternatives in the CRE AI Assistants & Copilots category, buyers must distinguish between general-purpose coding/formula assistants and specialized real estate platforms. Copilot in Excel competes most directly with other generalist tools that manipulate data. For analysts who are comfortable with Python and require advanced, programmatic data manipulation, Replit (BestCRE Score: 88) or Cursor (BestCRE Score: 90) offer far superior capabilities for building custom analytical scripts, though they require a higher technical baseline. For workflow automation that connects multiple applications, tools like Gumloop (BestCRE Score: 87) provide visual interfaces to move data between a CRM, a property management system, and a spreadsheet, whereas Microsoft’s tool is confined to the spreadsheet itself. In the realm of autonomous agents, Agentforce (BestCRE Score: 88) and Manus (BestCRE Score: 87) are designed to execute complex, multi-step tasks across different software environments, offering a more proactive approach than Microsoft’s reactive, prompt-based interface. Conduit (BestCRE Score: 87) provides more specialized financial data integrations. Copilot wins on organizational security and immediate familiarity, as it requires zero workflow changes for an analyst already working in Excel. However, it loses significantly on autonomy and specialized data handling. Firms must decide if they want a simple formula assistant or a more complex agentic workflow builder.

The bottom line

Copilot in Excel is a mandatory evaluation for any commercial real estate firm already entrenched in the Microsoft 365 ecosystem, but it is not a substitute for specialized underwriting software. It excels at accelerating the mechanical tasks of data formatting, basic formula generation, and chart creation, allowing analysts to process raw rent rolls and operating statements faster. However, its complete lack of native commercial real estate logic means it cannot independently structure a capital stack or audit a complex cash flow model. The tool requires strict oversight, as it will confidently generate mathematically incorrect syntax if prompts are not highly specific. Firms should purchase licenses for their heavy spreadsheet users to drive daily efficiency, but they must maintain rigorous manual auditing protocols for all investment-critical calculations. It is a powerful calculator, not an autonomous analyst.

Frequently asked questions