CPAs are revolutionizing their workflows—from automating emails and drafting technical memos to analyzing data and executing marketing strategies.

With advancements in language processing, decision-making, and automation now more accessible than ever, accountants have a wide array of innovative AI applications to choose from.

“There are so many new tools,” said Tricia Katebini, CPA, a partner at GRF CPAs & Advisors in Bethesda, Md. She has helped the firm implement numerous new AI-based tools. “It’s nice to see, but it’s overwhelming.”

To navigate the buzz and overwhelming range of new tools, the JofA invited six CPAs to share the most compelling AI use cases they’ve encountered — from tailored solutions to ready-made software (see the sidebar 'How to Determine AI Technology Needs').

Here are their stories.

CASE 1: SNAIL MAIL AUTOMATION

Karl Spanbauer, CPA, oversees finances at Capital Area Food Bank, a leading nonprofit that distributes around 60 million meals each year to residents in and around Washington, D.C.

After stepping into the role, one major frustration stood out for him: the mail. Not emails, but stacks of physical mail — vendor invoices, tax notices, statements, and more — that flooded his small accounting team.

“I lost a couple documents. It was literally hundreds of pieces of mail,” Spanbauer said. “I knew we had to automate it somehow.”

Leveraging his expertise with Microsoft Power Automate and other tools, Spanbauer developed a custom solution to scan, summarize, and respond to incoming mail.

Here’s how it works:

1. Staffers run all business-related mail through a scanner.

2. The scanned file is automatically saved to a shared drive.

3. An image analyzer extracts the text of the image.

4. The text is analyzed by a secure large language model (LLM), which generates a summary.

5. A ticket is opened in Jira, the team’s workflow system, which includes the summary and the original mail.

6. The team tackles the tickets, responding to the mail as needed.

He has expanded the system with more automations, allowing for quicker access to invoice statuses and smoother execution of follow-up actions.

Case 1 at a glance

Degree of difficulty: Drawing on extensive experience with Power Automate and other tools, Spanbauer swiftly implemented the solution—built upon months of groundwork to integrate systems and consolidate data.

Cost: The project incurred no extra implementation costs, as it utilized the team’s existing software licenses.

Payoff: The team now handles mail in focused 20-minute sessions every 'Mail Monday,' saving approximately four hours of collective staff time each week.

The takeaway: The project succeeded because Spanbauer already had a strong digital infrastructure—what he needed was a way to integrate physical mail into it. 'AI can’t interact with a piece of paper sitting on my desk. I had to get that piece of mail into that ecosystem,' he explained.

CASE 2: TECHNICAL ACCOUNTING MEMOS

Glenn Hopper leads AI research and development at Eventus Advisory Group in Memphis, Tennessee. His firm provides a diverse range of services, including fractional CFO and controllership roles, catering to both private enterprises and small to mid-sized public companies.

One of his recent initiatives includes developing a specialized bot capable of drafting detailed technical accounting memos.

We get asked to do these accounting memos all the time, Hopper explained. For instance, a memo might address the issue of goodwill impairment during an acquisition. He noted the challenge lies in striking a balance: ““You have to get very technical, and you want to make it clear and understandable.”

Leveraging his expertise in generative AI, Hopper processed the organization’s extensive collection of accounting memos for analysis by a large language model, employing a method called retrieval-augmented generation (RAG).

“We’ve been writing these forever,” Hopper said, “and we have a style we use for them.”

This process entailed breaking the memos into smaller, thematically focused segments that a large language model can more easily interpret. Hopper then transformed these segments through vectorization and stored them in a vector database powered by OpenAI technology.

"The final outcome is a rich repository of information that the bot can access to deliver responses that are both coherent and contextually accurate," Hopper explained. This memo generator is built on OpenAI’s Assistants API platform, enabling organizations to develop sophisticated, tailor-made bots.

Hopper noted that he worked diligently to “pound out any possibility for hallucination,” a common issue in generative AI where the model invents information when data is lacking. “I think we’ve minimized it a lot by giving it very specific examples to refer to, very specific instructions,” he said. This fine-tuning process involved training the custom bot on hundreds of examples. However, Hopper declined to share further details, citing Eventus’ proprietary methods.

Once the tool produces a draft of the accounting memo, Eventus ensures a “human-in-the-loop” approach, with experts reviewing and refining the content as necessary.

“The people who are using it are kind of blown away — and for many of them, it was their first foray into even using ChatGPT,” he said.

Case 2 at a glance

Degree of difficulty: This implementation of OpenAI’s technology was highly tailored, demanding approximately 60 hours of Hopper’s time and drawing heavily on his deep expertise.

Cost: In addition to the cost of ChatGPT licenses, running the memo-generating bot—and similar advanced bots—can cost several dollars per use. Paid ChatGPT plans begin at $20 per month for individual users and $25 per user in team environments.

Payoff: Hopper estimates that generating accounting memos has gone from a four-hour task to just 30 minutes—including thorough human review. He has also developed additional bots capable of handling complex tasks, such as analyzing 10-K filings.

The takeaway: While the memo bot proved to be a success, it was only made possible through Hopper’s specialized expertise. For many other organizations, a more practical and cost-effective approach may be to wait for an off-the-shelf solution from a vendor. “It takes time to do, and you need to have a certain skill set to do it,” Hopper said. “It’s not cheap to get these things done.”

CASE 3: MARKETING THE BUSINESS

Barrett E. Young, CPA, serves as the marketing partner at GWCPA in La Plata, Maryland, a firm dedicated to guiding family-owned businesses through leadership transitions. Since joining GWCPA in 2017, Young has led its 15-member team in embracing technology more effectively—including the integration of ChatGPT and related AI tools over the past year.

The firm has also launched a client-facing tool called the GWCPA Generations Advisor—a tailored ChatGPT implementation that provides guidance on ownership transitions. While it requires an active ChatGPT license to use, it doesn’t carry per-use fees, as it’s less complex than the Eventus bot discussed earlier.

“I’m using ChatGPT every night when I have ideas in my brain at two in the morning,” Young said. “I’m asking it questions and making notes, ‘What if I gave that to my customers?'”

Creating a custom GPT like this is relatively simple: the developer provides written instructions to ChatGPT, outlining the bot’s purpose, preferred tone, response structure, and other relevant guidelines.

The goal was to create “something I could give away to our target audience that would start conversations and add value in their lives,” Young said.

To enhance the custom GPT, Young uploaded GWCPA’s marketing materials, enabling the bot to reference key information and mirror the firm’s communication style—an application of the retrieval-augmented generation (RAG) approach. Access is granted through a simple sign-up to GWCPA’s free marketing email list, which Young also uses to share weekly prompt suggestions with subscribers. The goal is for the GPT to handle basic succession planning questions, increase awareness of GWCPA’s services, and help pave the way for more meaningful conversations when clients are ready to engage with the team."

Case 3 at a glance

Degree of difficulty: For Young, implementing the Generations Advisor was a swift and straightforward process, building on his prior experience creating custom GPTs for personal use. These more basic custom GPTs can be easily developed within the ChatGPT interface using simple, plain-language instructions.

Cost: Building a custom GPT necessitates a paid ChatGPT subscription, with pricing starting at $20 per month and rising to $200 per month based on the required level of access.

Payoff: Young’s mailing list has seen steady growth, with the custom GPT contributing around 50 new subscribers over the past three months. It consistently attracts a modest but targeted group each month. Although the audience remains relatively small, Young emphasizes that it’s reaching the right individuals. ‘One or two clients from that makes the entire thing worth it because we’re high-value, high-touch services,’ he explained."

The takeaway: Generations Advisor represents a modest customization of the core ChatGPT platform, leveraging its capabilities to deliver a valuable tool for clients that Young hopes will spark greater interest in the firm’s services.

CASE 4: DATA ANALYSIS WITH AI-ASSISTED CODING

Don Tomoff, CPA, a director at Invenio Advisors in Cleveland, leverages ChatGPT for coding and data analysis as a consultant for executives and companies, helping to save time and reduce costs while enabling more efficient project delivery.

“I used to outsource coding for projects, and I haven’t had to do that since January of 2023,” Tomoff said. “That’s been just a huge win.”

He now relies on platforms like ChatGPT to generate code for tasks in Excel and other software. For instance, when he needed a script to locate a specific piece of text and delete all rows above it in an Excel spreadsheet—regardless of its position—Tomoff simply described the problem in natural language to ChatGPT, which then produced functional Visual Basic for Applications (VBA) code. Other AI models, such as Anthropic’s Claude and Google’s Gemini, offer similar capabilities. For an example prompt used with Claude, see the sidebar "Claude AI Prompt for VBA."

Tomoff approaches larger projects by breaking his requests into multiple stages. He first asks ChatGPT to outline the necessary steps for a given task, then directs it to generate code for each step. If errors occur, he describes the issues and requests corrections, enabling iterative refinement.

According to Tomoff, outsourcing coding used to cost as much as $150 per hour, totaling up to $2,000 for a typical project. Additionally, the time required to develop, test, and refine solutions has been reduced from days to just hours.

“You want to explain as precisely as you can what you want,” he said. “You don’t want to eat the elephant in one effort, because it will stop on you if you give it too much to digest. It’s iterative.”

Brianne Smith, CPA/PFS, Ph.D.—a financial adviser, managing member of her own accounting firm, and assistant professor of accounting at Auburn University at Montgomery, Alabama—guides her accounting students through four key phases of data analysis when using ChatGPT: "Ask the question, master the data, perform the analytics, and share the story."

Smith explained: “You wouldn’t load ChatGPT and ask it to do all four things at the same time — but we would focus in on various areas and chunk them out into intermediary goals at the end.”

ChatGPT and similar tools can directly analyze and transform files, going beyond merely generating code for users to run.

Case 4 at a glance

Degree of difficulty: The learning curve is low, provided the user has some familiarity with generative AI bots. Success hinges on clearly defining the problem and having the patience to experiment and troubleshoot.

Cost: While code generation is available in the free versions of major generative AI models, paid versions often offer greater convenience and enhanced capabilities.

Payoff: TTomoff is now able to handle more complex projects independently. “Projects that I either would not undertake or would take too long that I would undertake, I’m doing in minutes,” he said.

The takeaway: Tomoff advises embracing the bot as a powerful extension of your skills. “If you’re an accountant, you no longer can say, ‘I don’t know how to do this,'” he said.

CASE 5: BRAINSTORMING AND RESEARCH

LLMs can be powerful research assistants — when they’re used correctly.

When researching a topic, Tomoff often inputs an outline into ChatGPT or Perplexity, an LLM-powered search engine, and asks: “What other areas would you add, and why would you add them?”

The responses can reveal research gaps he might have missed, while follow-up questions offer foundational insights into those topics. According to Tomoff, the true strength lies in the sheer volume of ideas these bots can generate.

“People will say it’s not any more creative than you and me,” he said. “It can sit and come up with ideas 100 times faster.” He also highlighted ChatGPT’s recent introduction of a ‘deep research’ feature, a more advanced tool available to paid users.

However, Tomoff warned that chatbots typically bring users about 90% of the way to a solution. Due to the possibility of hallucinations, users must always verify any facts derived from the responses.

Meanwhile, a new generation of research tools is customizing LLMs to meet the specific demands of accounting. Young’s firm recently began using Ask Blue J, a bot designed exclusively for federal and state tax laws. It provides precise answers and citations for detailed tax inquiries and can also draft memos and client emails. However, it isn’t without flaws. According to Young, the answers are about “95%” accurate, but the model frequently has difficulty producing correct calculations when dealing with frequently changing, inflation-adjusted figures.

“I’m always telling my staff, ‘Double-check that,'” he said.

Case 5 at a glance

Degree of difficulty: Chatbots offer intuitive research support, yet preventing hallucinations demands a combination of subject expertise and familiarity with the bots.

Cost: Free chatbot versions are useful for brainstorming and general research, while costs for more advanced or industry-specific solutions can vary widely.

Payoff: At their best, chatbots enable accountants to swiftly address inquiries and uncover new avenues for research. “This is the tax manager I always wish I had, where I could ask any question and not get judged,” Young said.

The takeaway: Chatbots aren’t magical genies that can instantly solve every question, but they serve as valuable tools to expand one’s knowledge. “It just augments what you know,” Tomoff said.

CASE 6: AUDIT AUTOMATION AND ASSISTANCE

Katebini serves as an adviser to not-for-profit organizations at GRF CPAs & Advisors, where she has guided clients through the recent surge of automation and AI-driven tools, particularly in the field of auditing

The firm has recently adopted Trullion’s Audit Suite, a platform offering a variety of valuable features. It is capable of:

1. Search through documentation to find evidence related to an audit sample. “You’re not scouring through 80 pages of a bulk PDF. It’s very easy in that respect,” Katebini said. This technology enables staff to quickly determine if any evidence is missing for a specific audit point.

2. Generate concise summaries from documents like board minutes and lease agreements.

3. Automate various aspects of financial statements, such as footings and linking disclosures directly to the corresponding statement balances. “Our production team doesn’t have to go through the process of manually footing everything,” Katebini said.

Trullion is just one among many audit-focused solutions available today. For example, the AICPA’s Dynamic Audit Solution examines large datasets to detect patterns and potential risks. Other tools can automatically identify discrepancies and missing information or quickly reformat data when it’s submitted in an incorrect format.

“It’s the press of a button for us now,” Katebini said.

Collectively, the firm’s new tools are enhancing staff capabilities to perform advanced analyses and strengthen client relationships.

“Our staff is so inundated, there’s so much work to be done, it’s hard to take a step back and talk with your client sometimes, and technology advancements in assurance engagements is creating efficiencies to allow us more time to advise,” Katebini said.

Case 6 at a glance

Degree of difficulty: While off-the-shelf products are becoming more widely available, successfully selecting and implementing these solutions demands a strong understanding of the technology, its practical applications, and solid project management skills.

Cost: Varies by product.

Payoff: Significantly enhanced efficiency in handling and producing documentation.

The takeaway: A new generation of technological solutions is emerging, along with a growing role for finance technologists to deploy them.

How to determine AI technology needs

With the rapid expansion of artificial intelligence (AI) technologies, navigating the myriad of options can be overwhelming. Tricia Katebini, CPA and partner at GRF CPAs & Advisors in Bethesda, Md., shared that she dedicates a significant amount of time to technology, explaining, “I’m a person in the firm that has been designated to head up and lead the technology implementations, the vetting.”

Katebini offered the following guidance for discovering and deploying new solutions:

Identifying pain points: Instead of looking for what “looks cool,” Katebini said, the firm began by asking staff about the tasks and work that were the most time-consuming and unexciting — such as creating note disclosures or getting data in the right format. “How can we solve that problem?”

Seeking solutions: The team looks for solutions, including by asking vendors directly. “We’re bringing our pain points and problems to our vendors, too, and they’re very willing to listen.”

Making a choice: The right tool is a matter of cost, utility, and compatibility with the rest of an organization’s tech stack.

Brianne Smith, CPA/PFS, Ph.D., who manages her own accounting practice and financial advisory firm while serving as an assistant professor of accounting at Auburn University’s Montgomery campus in Alabama, noted that consulting on technology implementation can even evolve into a standalone line of business for a firm.

“My firm and other firms are starting to consult on this type of process,” Smith said. “We make sure that they have proper security and privacy. We also recommend various tools for pain points.”

She and others recommended beginning modestly with AI—perhaps simply by experimenting for personal use—and seeking input and assistance from others throughout the process.

“There’s a whole community. We are all very willing to collaborate and talk about what tools we use,” Katebini said. “It’s kind of exciting to be at the crossroads.”

Claude AI prompt for VBA

Below is an Invenio Advisors LLC prompt instructing the Claude AI model to create a visual basic (VBA) bot to move items from one list to another list in Excel.

Prompt Structure

ClaudeAI PROMPT to move items from a list to another list (VBA)

You are an Excel VBA expert. I would like your help in creating a VBA macro that will MOVE (not copy) rows from one table to another when they’ve been checked. Specifically:

1. Source worksheet: {INSERT WORKSHEET NAME} with table named {INSERT EXCEL TABLE NAME}
2. Destination worksheet: {INSERT WORKSHEET NAME} with table named {INSERT EXCEL TABLE NAME} (create these if they doesn’t exist).
3. The destination table should start at cell A10 if it needs to be created
4. Move only rows where a checkbox in the last column is checked
5. Transfer all columns EXCEPT the last 3 columns from the source table
6. Prevent duplicates based on the video title (column 2) in the destination table
7. After moving a row, DELETE it from the source table
8. Include error handling for cases where tables don’t exist
9. Alert the user with a message upon completion

The macro should first check if items are already in the saved list before moving them.

Once you have created the code, please advise if you have any recommendations for improvement. I will then let you know if I want to make any changes to the code you provide.

After you complete this task, please provide instructions on how to implement the process in my workbook. Take a deep breath and let’s do this!

Source: Invenio Advisors LLC