Beyond Compliance – The Ways CPAs and CMAs Can Embrace the AI-Powered Future

By Kip Holderness, Ph.D., CPA, CMA, CFE, CVA

When I began my career, I entered the accounting field because I was told accounting was the language of business and those who could speak the language could solve problems. That message still holds true today, but the nature of those problems is changing. Over the years, something shifted.

Many of my students at West Virginia University, despite landing solid jobs, found themselves performing work that felt rigid and overly focused on compliance. Creativity and problem-solving – the things that once made accounting feel dynamic – took a back seat to documentation and regulatory interpretation. Students weren’t designing solutions; they were checking boxes. In many ways, they became translators of business and bureaucracy.

Now, however, the accounting profession is transforming, underpinned by generative artificial intelligence (GenAI). ChatGPT, Claude and Gemini are changing how accountants write, summarize and communicate information. These large language models (LLMs) can draft reports, translate regulatory language and assist in coding, making once-technical tasks accessible to non-technical professionals.

This transformation led me to develop and teach a university course on artificial intelligence applications in accounting. The insights I gained have shaped my perspective on how CPAs and CMAs can begin using GenAI in their work.

How AI is Shifting the Accounting Landscape

Automation of Routine Tasks. Data entry, bank reconciliation and expense classification can now be handled by bots or machine learning models with little oversight. GenAI adds even greater capabilities to these tools by increasing CPAs’ and CMAs’ abilities to work with unstructured data. Traditional automation requires highly structured and repeatable inputs, but GenAI can interpret varied formats, phrasing and context. Processes don’t have to look exactly the same for the system to be useful. This flexibility enables accountants to handle ambiguity more effectively and expand automation into areas that were previously too inconsistent for bots to manage.

Better Insights, Faster. Natural language processing and machine learning allow for rapid summarization of financial documents, anomaly detection and predictive forecasting. Leading audit firms now use AI to evaluate the risk of every transaction within a general ledger. Auditors can rely on AI to prioritize high-risk, high-impact items, dramatically improving audit efficiency and strengthening assurance quality.

In tax, professionals increasingly use firm-provided AI chatbots trained on current tax codes and regulatory guidance. AI tools are also being used to scan company documents to extract pertinent information and pre-fill tax forms, significantly reducing manual data entry and ensuring greater compliance with evolving regulations.

CPAs and CMAs in industry or managerial roles are experiencing a shift from manual, backward-looking processes to technology-augmented, forward-thinking decision-making. Cost analysis, budgeting, forecasting, performance evaluation, variance analysis, inventory control, and cost allocation are now supported by real-time data and predictive analytics.

Evolving Skillsets. Modern CPAs and CMAs must be tech-savvy advisors. GenAI is pushing the profession from compliance to consultation. Those who understand how to leverage data, communicate insights and apply technology thoughtfully will be best positioned to lead in this new landscape.

How to Get Started with GenAI: Lessons from the Classroom

My AI in Accounting course focused on demystifying GenAI and giving students hands-on experience with practical tools. You don’t need a tech background to follow the same path.

Here are five ways CPAs and CMAs can start experimenting.

1. Understand the Basics. LLMs excel at generating and organizing text, drafting summaries, translating jargon and suggesting code. They can also hallucinate facts, misinterpret context and produce results that sound authoritative but are partially or completely incorrect. By experimenting with different prompts and evaluating outputs critically, accounting and finance professionals can develop intuition for using LLMs effectively and ethically.

My students quickly learned what LLMs cannot do well. We have seen several instances where AI confidently supplied quotations or references that turned out to be misattributed or entirely fabricated. Recognizing these weaknesses is valuable because it helps students avoid potentially careerdamaging mistakes.

In recent years, several professionals have faced embarrassment or even disciplinary action after relying on AIgenerated text that included fictitious legal or technical citations. For example, one court filing referenced cases that did not exist, while a government report included citations to research that had never been published. Both situations were ultimately corrected, but they illustrate the real reputational and legal risks of unverified AI content. Although GenAI is improving rapidly, it still cannot be relied upon to provide authoritative citations without human oversight.

By testing prompts and verifying all references, quotes and citations, CPAs and CMAs can build the discernment needed to use GenAI effectively and ethically. Understanding what the technology does well and where it fails most often is the first step in integrating it responsibly into professional practice.

2. Try AI-Assisted Coding. ChatGPT and GitHub Copilot can help generate code for Excel macros, SQL queries or Python scripts, even if you’ve never coded. I gave my students a dataset of sales transactions that included salesperson, products, prices, order and shipping dates, and customer satisfaction ratings. Their assignment required them to calculate sales commissions, identify top-selling products, evaluate average customer satisfaction, determine which orders were most frequently cancelled and compare in-person versus online purchasing methods. They also analyzed shipping times and trends across salespeople and regions.

Using GenAI, the students generated Python scripts that automated these analyses and produced data visualizations such as bar charts of top performers, pie charts showing sales by product category and scatterplots linking customer satisfaction with delivery speed. Each group then built a concise report that summarized findings and presented them in a format suitable for management review. Although it took a couple of hours to refine the scripts, students quickly recognized the long-term efficiency gains for recurring analytics and reporting tasks.

3. Build a Custom GPT. Beyond AI tools' basic chat capabilities lies the power to build a custom GPT, a tailored AI assistant designed for a specific workflow or organization. In my own teaching, I discovered firsthand how valuable this can be. I built a simple chatbot trained on my course syllabus that could reduce repetitive student questions about due dates and grading policies.

I also encouraged students to explore the creative side of custom GPTs. One particularly inventive student built a custom GPT that gave personalized golf club recommendations based on a player’s handicap and budget. The avid golfers in the class were impressed with the quality of the recommendations. While the exercise didn't relate specifically to accounting, it left a lasting impression about how these tools can specialize remarkably well when trained on targeted knowledge and rules.

For accounting professionals, the same principle applies. A custom GPT could be trained on a firm’s policies, templates and terminology to answer employee or client questions, draft engagement letters, summarize tax guidance or guide new clients through onboarding. It could also perform higher-value analytical tasks such as analyzing data sets for trends, preparing variance analyses or generating first drafts of budgets and management reports. The possibilities are broad. What matters most is tailoring the custom GPT to the practitioner’s specific needs and ensuring that human oversight remains part of the process.

4. Low-Code/No-Code Automation Tools. Make.com and Zapier allow users to build powerful automations without writing code. These platforms connect seamlessly with tools such as email, QuickBooks, Google Sheets, Slack, Microsoft Teams, Dropbox, and Salesforce.

The capstone project for my class required students to interview a business professional to understand their daily responsibilities and pain points. They then had to design, build and deliver an AI-driven tool to address those needs. This required them to conceptualize how AI could improve efficiency and implement a working solution that the professional could actually use. Nearly every project incorporated some level of automation, a clear sign that these tools align closely with the practical realities of accounting and finance. Much of our profession involves gathering, verifying and communicating information across teams and systems, so automation naturally serves as a powerful extension of what accountants already do.

With just a bit of guidance, students created impressive solutions using low-code tools. Some of these projects were relatively simple. One student designed a system that automatically generated and emailed invoices for a local sewing shop. Some projects were more ambitious. Another student automated an HR workflow so that when a handwritten employee complaint was received, the program read and cataloged the report, emailed those involved for more details and drafted a plan of action aligned with company policy.

In each case, students could quickly see the value of these automations. One of my students working for a local tax firm built an automation that created secure client folders for new engagements and sent onboarding emails requesting required documents. This process had previously been handled manually. His boss was so impressed with the system that for the remainder of the semester, my student was allowed to bill the firm for the time he spent working on his homework.

These projects underscored a key lesson for students: Automation isn’t about replacing people; it’s about amplifying what accountants already do best, improving efficiency, ensuring accuracy and strengthening communication.

5. Start Small, Think Big. Don’t wait for your firm to announce an enterprise-wide AI strategy. Start by identifying one specific process that could benefit from greater efficiency, perhaps automating engagement letters, creating a custom dashboard or using AI to draft initial versions of management reports. Every small success builds your confidence and demonstrates the tangible value of AI integration. Over time, these incremental improvements can compound, positioning you as an innovator and a more strategic advisor within your organization.

About the Author: Kip Holderness is an accounting professor at West Virginia University and a CMA instructor at UWorld, helping thousands of candidates pass these high-stakes accounting exams. He also coordinates WVU’s Master of Forensic and Fraud Examination (FFE) and serves as the Director of Research for the Association of Certified Fraud Examiners Research Institute. He received a B.S. in Accountancy and an M.S. from Brigham Young University and a Ph.D. from Bentley University.

Thanks to the Sponsors of Today's CPA Magazine

This content was made possible by the sponsors of this issue of Today's CPA Magazine:

Accounting Biz Brokers

Accounting Practice Sales

Capstan Tax

Goodman Financial

Poe Group Advisors

Professional Accounting Sales