CONTINUING EDUCATION FOR TAX & FINANCIAL PROFESSIONALS

AI in Accounting: How CPAs Can Leverage Microsoft Copilot

Western CPE Staff

Most CPAs already spend their workday inside Microsoft 365. Excel for analysis, Word for memoranda, Outlook for client correspondence, PowerPoint for presentations, and Teams for everything in between. Microsoft Copilot brings artificial intelligence directly into those applications as an embedded assistant that works within the tools you already use. That integration is what makes Copilot especially relevant for Certified Public Accountants.

More Than an Excel Feature

Most of the conversation around Copilot in accounting focuses on Excel, and for good reason—it can generate formulas, analyze datasets, identify variances, and clean up messy spreadsheets using plain-language instructions. But limiting Copilot to Excel undersells its value to a CPA practice.

Copilot’s integration into all of Microsoft 365’s products give you an assistant that can help you cut down on busy work. Draft technical accounting memoranda in Word, transform financial data into executive-ready presentations in PowerPoint, or capture meeting outcomes and develop agendas from prior discussions in Teams. Copilot can handle all of it. For CPAs managing multiple clients and deadlines, that breadth across the full suite of Microsoft products is where you’ll see real gains in your productivity.

Prompting Is a Skill, One that You Can Foster

Copilot’s output quality depends entirely on the quality of the instructions it receives. A vague request produces a vague result. This is where CPAs have a built-in advantage: your profession already demands precision in documentation, clear communication of scope, and specificity about frameworks and standards. Those same habits translate directly into effective prompts.

The difference between a generic AI output and a genuinely useful one often comes down to providing the right context. There are a few things you’ll need to keep in mind for your prompts depending on your goals:

Specify the accounting framework that applies to your engagement, such as U.S. GAAP or IFRS.
Define the relevant time period, the intended audience, and the purpose of the analysis.
Build your instructions in structured layers and set clear boundaries on Copilot’s scope.

Giving your intentions and frameworks for prompts a clear structure and goal can narrow Copilot’s response, increasing its ability to align with your professional expectations. Think of Copilot as a new staff member: the more context and guardrails you provide upfront, the less work you’ll be doing to correct any errors or inconsistencies that might crop up.

Professional Judgment Still Belongs to the CPA

Adopting Copilot does not change a CPA’s fundamental obligations. Due care, competence, and accountability remain with you—not a tool. Copilot can generate a first draft of an audit narrative or a variance explanation, but it cannot assess whether that output is appropriate, complete, or consistent with the facts of an engagement. It does not exercise professional skepticism, and it can produce results that appear confident but are factually incorrect.

The ethical dimension matters, too. CPAs need to consider transparency with clients about how AI is being used, what documentation obligations arise when AI assists in generating workpapers or deliverables, and how to guard against the risks that come with over-reliance on any single tool. The profession’s existing standards around objectivity and due care all apply—they just need to be understood in a new context.

Governance and Security Are Not Optional

Before rolling out Copilot across a practice, firms need to think about how it fits within their existing controls. Copilot operates within Microsoft’s security architecture—including tenant isolation and role-based access—but that doesn’t eliminate the need for firm-level governance. Questions about data retention, what information Copilot can access, how AI-generated outputs are stored and reviewed, and who is responsible for oversight all need clear answers before adoption scales beyond a pilot.

Firms should also consider establishing formal usage policies that define acceptable and prohibited use cases, documentation standards for AI-assisted work, and review expectations. These practice management decisions will affect your engagement quality, client confidentiality, and regulatory compliance.

Getting Started

The CPAs who get the most out of Copilot won’t be the ones who try to automate everything overnight. Start with a single workflow—drafting a recurring client communication, summarizing a dataset during close, or preparing a first draft of a standard memo—and evaluate the results critically. As you develop confidence in what Copilot does well and where its limitations show up, you can expand into more complex applications across accounting, audit, tax, and advisory work.

New technologies are always being implemented in the accounting and tax industry. Copilot is the next step in that progression—not a replacement for professional judgment, but a tool that can amplify it when used with the right framework, the right controls, and the right understanding of both its capabilities and its limits.

Want to read more about how AI is changing the tax and accounting industry? Be sure to check out our article AI in Accounting: 5 Ways to Incorporate AI into Your Firm.

You can learn more about how to leverage AI in your practice with some handy coursework: