How AI is transforming the audit — and what it means for CPAs

The latest wave of artificial intelligence (AI) technology promises to offer the technological leap that large and small firms have long awaited to transform the audit.

Many rote tasks make aspects of auditing highly suitable candidates for automation, but firms have struggled to find AI systems and applications that fit into a cohesive platform that could address the variety of industries, clients, processes, and transactions they handle.

“We were always in a position where there was variability, and that variability made it harder for auditors to benefit from tools and systems that required more consistency in execution,” explained Danielle Supkis Cheek, CPA, senior vice president of AI, analytics, and assurance at Toronto-based software company Caseware. “Auditors are expected to do repetitive tasks just slightly differently enough that automation can be hard. That is why the current and next wave of intelligent automation capabilities are so exciting to practitioners, as we can now automate even with [the] variability that auditors are accustomed to.”

The development of generative and agentic AI technology has brought audit automation closer to reality. Here is how CPAs are implementing these technologies and how to examine the benefits and risks of AI in the audit.

GEN AI IS CHANGING HOW AUDITORS WORK

Generative AI, which burst into the public eye when OpenAI unveiled ChatGPT in 2022, can streamline workflows in many ways. For example, CPAs can use generative AI to speed up research by employing its linguistic and analytic capabilities to deliver citations from a firm’s knowledge base, navigating context and nuance in a way a traditional search could not.

Auditors already are using generative AI tools for basic audit tasks such as:

• Reading and summarizing contracts.
• Writing, or suggesting language for, communications such as memos.
• Processing large volumes of supporting documents and extracting or summarizing the most relevant information for audit workpapers.
• Assisting with basic data analysis and helping to create visualizations such as charts or graphs.

Some firms are investing in specialized generative AI audit tools designed to support specific audit functions. For example, there’s Caseware’s new “digital assistant,” AiDA, which is integrated into the Caseware Cloud platform to help streamline workflows, ensure compliance, and provide context-aware responses. AiDA retrieves information from a firm’s own body of documentation as it answers questions, analyzes documents, generates memos, and more.

The idea is to provide the flexibility of an app like ChatGPT with the detail of a firm’s own policies and memos and the data protection of the firm’s secure network.

“I want to automate the first draft of a lot of content for our users, but we’re going to do it in a way that safeguards and preserves human judgment,” Supkis Cheek said.

Caseware is also working with the AICPA and CPA.com on the Dynamic Audit Solution (DAS), a cloud-based audit platform that integrates real-time data ingestion, analytics-driven risk assessment, adaptive workflows, and automation to streamline the audit process.

Emily Remington, CPA, CGMA, director of audit product management for CPA.com, emphasizes the importance of using fit-for-purpose solutions that align to the auditors’ workflow.

“The real value of generative AI in audit isn’t just about speed; it’s about shifting the focus of the practitioner,” she said. “By automating documentation, assisting with summarizing, and providing preliminary analysis of data, we can re-allocate auditors’ focus from low-value, repetitive tasks to the high-value areas that truly require their professional judgment, critical thinking, and skepticism. This fundamentally elevates the quality and insight of the audit.”

AGENTIC AI COULD BE A GAME CHANGER

Agentic AI is touted to be the next leap for AI technology. Instead of just answering questions and generating other content, agentic software can take action (also see, “Agentic AI Poised to Change the Way CPAs Work,” JofA, June 1, 2025). For example, ChatGPT’s agent mode can create and carry out multistep plans, including gathering information from third-party services, downloading and manipulating files, and making purchases (also see, “Creating an AI Agent in ChatGPT,” JofA, Nov. 1, 2025).

Agentic AI can perform advanced audit tasks easier and with less human input than generative AI. Consider these examples:

• Compliance workflows: An AI agent may perform a series of tasks, such as analyzing data, verifying controls, escalating issues, and generating reports, without manual input.
• Reconciliation: An agent could access, for example, the general ledger, subledgers, and bank feeds to perform reconciliation in real time, flag mistakes with explanations, and generate draft adjustments for human approval.
• Continuous monitoring: An agent may monitor data from disparate systems in real time and detect anomalies, such as duplicate payments or control failures, and potential fraud.

Tech companies in the audit and accounting space are racing to launch products with agentic capabilities. “That’s going to be a huge game changer,” said Jessie Kanter, CPA, a partner at New York City-based Citrin Cooperman, where she is laying the groundwork for agentic AI software as the leader of the firm’s audit innovation and methodology initiatives.

For example, she said, a set of agents at her firm could soon:

• Access the client’s general ledger data to fill out a workpaper.
• Fill and send a cash confirmation form to the bank.
• Review the bank’s response and compare it to the workpaper, asking for human intervention if a discrepancy is detected.

“It takes us a lot of time to just look over the information that is received by an auditor,” Kanter said.

Agentic AI is at the heart of intelligent process automation, a more flexible and adaptable evolution of technologies like robotic process automation (RPA), said Amy Pawlicki, vice president—Assurance & Advisory Innovation at the Association of International Certified Professional Accountants. “The difference between RPA and agentic AI is the intelligence factor,” she said.

Byron Patrick, CPA/CITP, senior product manager at software maker Karbon and co-founder of generative AI training provider TB Academy, describes the differences between RPA and agentic AI as follows: RPA is a linear technology that follows a specific recipe or set of steps. If something in that recipe changes, RPA will fail without human intervention. Agentic AI can understand context and goal setting. It can evaluate the result at each step along the way and then decide how to approach the next step, though human review of the results is still necessary.

“It’s like walking down a hallway blindfolded and just counting your steps to decide when you turn versus being able to see and decide when to turn based on what you are seeing,” Patrick said.

ORGANIZATION BEFORE IMPLEMENTATION

How does a firm start implementing AI technology? Often, the first step is getting organized. For example, a customized AI tool is only as good as the data it’s accessing.

“We know that the backbone of transformation is standardization,” Kanter said. “You’ve got to clean out your closet in order to help you find things easier.”

Doing so at Citrin Cooperman is a “huge” undertaking that has stretched for more than a year, requiring participation not just from technologists but also from subject matter experts, managers, and the firm’s quality and standards teams.

“We have a whole project team that is just developing the standardized workpaper, the layout, the functionality,” Kanter said. Standardizing the firm’s knowledge base and processes makes it easier for AI to surface the right information or apply the appropriate firm processes to a client’s specific needs. In other words, standardization provides the structure AI systems need to recognize patterns and scale automation across varied engagements.

Firms may also need to get clients involved in implementation. In a recent survey of audit firms by CPA.com, a lack of access to client data was one of the top obstacles to implementing AI and other technology.

Managing all this technology can become a job in and of itself. Firms should think strategically and look for chances to simplify, rather than complicate, their tech stacks as they implement AI.

“We see success when firms move beyond viewing technology implementation, inclusive of AI, as a side project and rather integrate it into a cohesive, firmwide strategy,” Remington said. “The goal is not only optimizing individual audit processes, but to elevate the entire client experience.”

KEEPING THE HUMAN IN THE LOOP

It’s critical to think about how AI will fit into existing workflows, Supkis Cheek said. Whether they use generative or agentic AI tools, firms should ensure humans review and more broadly take responsibility for the output — similar to how staff output is reviewed. For example, an internal generative AI platform should be configured to provide easily verified citations and references. Software vendors “need to put out more and more tooling to fact-check the content as fast as possible,” Supkis Cheek said.

The goal is to fight “automation bias,” or the instinct to simply trust the output of technology. And auditors, she said, are well positioned to provide the human oversight and judgment needed to make the most of emerging technologies, because the profession’s standards encourage independent thought and questioning.

“Our code of ethics should be the envy and model for every other profession to use AI,” Supkis Cheek said. “Don’t subordinate your judgment. Keep your skepticism.”

FROM EFFICIENCY TO TRANSFORMATION

With the current wave of AI implementation, firms seem to be aiming for efficiency above all else, according to the inaugural AICPA and CPA.com Audit Transformation Survey, released in December.

“We saw firms are overwhelmingly prioritizing time saving, trying to reduce the time and really seeing that as a motivator to invest in technology,” Remington said.

But she and others urged firm leaders to think bigger: How will humans work with the new technology to do their jobs better, driving quality and offering more value? “The future of the auditor is not one of AI replacement, but of upskilling and role redefinition. We are moving from a historical, compliance-based mindset to a data-driven, technology-enabled, insights-led partner.”

While firms initially gravitate toward technology as a time-saving measure — the true opportunity demands a larger vision. Leaders must focus on how this technology will augment the human element, elevating the quality and value of services.

AI and the modern audit, Remington said, “is here. It is now. It’s about adapting with an agile approach as our access to high-quality data increases, constantly refining your process, your thinking, and how you execute.”

---

About the author

Andrew Kenney is a freelance writer based in Colorado. To comment on this article or to suggest an idea for another article, contact Jeff Drew at Jeff.Drew@aicpa-cima.com.

---

LEARNING RESOURCES

AI-Driven Audit Dynamics: Navigating the Standards in a New Era

The pervasiveness of AI provides an opportunity to leverage technology for a higher-quality audit. Join our discussion for the latest on AI, its use in audits, and the pros and cons of emerging technology.

Feb. 3, 1–2:25 p.m. ET

WEBCAST

Futuristic Audit Trends: The Future Is Now

You’ll learn how to take advantage of cloud technology, AI, machine learning, and blockchain to streamline your workflow and utilize remote applications and also how to utilize ecosystem applications to create more profitable engagements and a better work environment for your teams.

June 5, 11 a.m.–noon ET

WEBCAST

AICPA ENGAGE 26

Artificial intelligence and audit will be on the agenda as the biggest event in the accounting profession celebrates its 10th anniversary at the ARIA in Las Vegas.

June 8–11

CONFERENCE

Certified Information Technology Professional

The AICPA’s Certified Information Technology Professional (CITP) credential represents the intersection of technology and financial reporting. CITPs understand how information technology, accounting, and finance interrelate. Choose from three pathways to earn the credential. For more information, click on the headline above.

CREDENTIAL

For more information or to make a purchase, go to aicpa-cima.com/cpe-learning or call 888-777-7077.

---

MEMBER RESOURCES

Articles

“AI and the Audit: Finance Leaders Strongly Support Forward-Thinking Firms,” JofA, Nov. 19, 2025

“Writing an Effective AI Prompt for an Audit,” JofA, Nov. 17, 2025

“AI Promises More Efficiency but Won’t Replace Auditors,” JofA, Oct. 9, 2024

“A&A Focus Recap: The Ethics of Using AI in an Audit,” JofA, July 22, 2024

“What AI Can Do for Auditors,” JofA, Feb. 1, 2024

JofA Artificial Intelligence (AI) Coverage