Public accounting firms face pressure from multiple directions. Clients expect more value. Talent is constrained. Regulatory scrutiny remains high. Engagement economics are changing. At the same time, client financial environments are becoming more automated and complex.
In this Vision 2026 executive conversation, Jeff Kovacs, President of Citrin Cooperman, joins Sarah McGinnity, Audit and Assurance Practice Leader at MindBridge, to discuss how AI, automation, and full-population financial analysis are changing the economics and execution of audit while professional judgment remains central.
Kovacs describes the current period as one of the most significant transformations he has seen in more than 40 years in the profession. Previous waves of technology often focused on making existing audit processes more efficient. Generative and agentic AI create an opportunity to go further by rethinking the work itself.
That transformation begins with standardization. As a highly acquisitive firm operating across geographies and industries, Citrin Cooperman first worked to create greater consistency in its workflows, workpapers, methodology, training, and audit platform. Kovacs summarizes the logic simply: firms cannot scale without automation, and they cannot automate what they have not standardized.
This foundation also changed how the firm evaluates technology. Citrin Cooperman historically leaned toward buying tools rather than building them internally, but over time found that disconnected point solutions did not always improve the audit lifecycle as a whole. The firm now takes a build-and-buy approach, partnering where outside technology offers differentiated value and building more bespoke workflows where its own methodology and processes matter most.
Across the audit lifecycle, Kovacs sees AI affecting ingestion, risk assessment, substantive testing, quality review, and client communication. He points to risk assessment as one of the most important areas because better visibility into client data can help auditors identify where attention is needed earlier in the engagement.
He specifically discusses Citrin Cooperman’s work with MindBridge in full-population financial analysis. By analyzing transaction populations and risk signals, auditors can gain deeper visibility into complex client activity and use that information to inform their risk assessment and audit response.
Revenue testing is one example. Kovacs says the firm has developed policies for using MindBridge risk scoring in revenue testing, aligned with audit standards and relevant assertions. Rather than relying only on limited views of the data, auditors can examine the full population and use the resulting risk information to focus their work.
Citrin Cooperman is also exploring agentic workflows in substantive testing. Kovacs describes agents preparing initial workpaper content based on the firm’s methodology and procedures arising from risk assessment. In that model, professionals spend less time on manual preparation and more time reviewing outputs, investigating anomalies, assessing outliers, and determining what should happen next.
The firm is also working to move quality review closer to where the work is performed. Instead of allowing issues to surface later in the engagement, Citrin Cooperman is exploring diagnostics and review workflows that can identify potential problems earlier.
Kovacs is clear that this does not remove the need for professional judgment. Risk signals are not audit conclusions. AI can surface information, automate work, and help professionals move faster, but auditors still need to interpret the evidence, challenge results, apply skepticism, and determine the appropriate response.
That distinction becomes more important as AI takes on more work. Kovacs describes Citrin Cooperman’s staffing philosophy with the equation “same equals more”: the same number of professionals can potentially deliver more value when technology reduces mechanical work.
The economics of audit may change with that shift. Kovacs estimates that human capital may represent roughly 70% of the cost of delivering a traditional audit today and uses a future scenario in which that could move toward approximately 40% to 50% as technology assumes more of the work.
He also uses a 400-hour private-company audit as an example. Based on Citrin Cooperman’s modeling, Kovacs estimates that automation could eventually remove roughly 20% to 35% of the hours across the audit lifecycle, potentially reaching 40%. He notes that those savings are offset in part by the cost of technology and presents the figures as forward-looking estimates rather than realized results.
The broader significance is not simply fewer hours. A different cost structure can support a different engagement model, while stronger technology can improve risk assessment, execution, quality review, and client insight. Kovacs expects this shift to challenge traditional time-and-materials billing and move firms further toward fixed-fee models.
The conversation also explores what this means for auditor development. Historically, junior professionals learned through repetition: performing manual tasks, receiving review comments, and gradually building judgment. As technology performs more of that work, firms need new ways to develop technical understanding, skepticism, and professional judgment.
Kovacs expects training to place greater emphasis on audit standards, assertions, evidence, methodology, risk assessment, and the ability to recognize when AI has produced an incorrect or incomplete answer. Senior auditors may spend less time correcting manual work and more time developing how younger professionals reason and apply judgment.
For firms deciding where to begin, Kovacs recommends starting with one or two meaningful pain points rather than attempting to redesign everything at once. Identify where teams spend the most time, where clients experience friction, and where risk assessment or workflow execution could improve. Then test, learn, and expand.
The session presents audit transformation as a combination of people, process, and technology. AI can change how work is performed and how engagements are structured, but the profession’s enduring value remains human: technical knowledge, judgment, skepticism, critical thinking, and the ability to understand and serve clients.
Chapters
00:55 Transforming audit and introducing Jeff Kovacs
01:50 Why this period of audit transformation is different
04:03 Quality, talent, economics, and the pressures driving change
07:01 Standardizing audit workflows before automating them
10:46 Build vs. buy and creating a connected audit technology strategy
14:39 Where AI is changing the audit lifecycle
20:15 Where AI in audit is mature and where it is still early
23:33 Full-population analysis in risk assessment and revenue testing
26:41 Deciding what technology to build and what to buy
32:30 Professional judgment and the changing role of the auditor
38:44 How technology changes audit quality and engagement economics
44:14 Regulation and the adoption of AI in audit
45:47 Developing junior auditors as repetitive work changes
53:24 Where accounting firms should start with audit transformation
56:15 What the auditor’s role could look like five years from now
