One year ago this month, PricewaterhouseCoopers auditors first began using the firm's own enterprise-led generative AI tools, designed to accelerate the process of putting together audit and other assurance reports, according to PwC’s Shawn Panson.
Since that time, it’s the fast pace of the technology’s change that has been most remarkable for Panson, the Big Four Firm’s U.S. assurance and transformation leader who has led the initiative and worked with engineers and data scientists to develop large-scale AI tools.
“What surprised me the most is just the pace of change and how much even our tools that we’ve built have evolved over a one year period,” he told CFO Dive. “All in a good way but certainly I think it provides just one more challenge as you’re running a business and trying to build technology in the ever-changing landscape.”
For 2026 year-end audits, there are a number of areas that PwC will be using AI to reduce manual work and accelerate the process.
For example, PwC’s Simpified Audit - Execute tool uses AI to read and review supporting documents and flag missing information. Another tool analyzes audit committee communications to surface any issues that need to be handled sooner in the audit preparation cycle, and an advanced walkthrough assistant documents business processes to help auditors prepare for discussions.
The assurance initiative is one piece of the firm’s broader AI strategy. After genAI exploded on the scene in 2022, PwC and other accounting firms, including KPMG and Deloitte, began racing to ensure they don’t fall behind in AI with investments designed to expand their capabilities.
In PwC’s Global Annual Review 2025 letter, Global Chairman Mohamed Kande said the company was accelerating nearly $1.5 billion in investments to grow its next-generation AI capabilities. A spokesperson pointed to the report as the latest available total for PwC’s AI investments.
The company’s AI initiatives have included acquisitions as well as the development of its agent OS, which is a command centre that connects AI agents with workflows, according to the letter. In addition, PwC also rolled out its own GenAI chatbot named ChatPwC.
Prior to last year’s launch of the enterprise-led AI in assurance, Panson also said assurance practitioners have been given AI training along with access to large language models they could then use for certain pre-approved work such as administrative tasks.
Since the program’s launch, Panson said there are some things that have changed with regard to auditing and assurance practice, and others that haven’t.
For example, the final product of the audit reports still looks the same, he said.
“We’re still subject to regulation and professional standards. What comes out the back end is still comfort on the financial statements or whatever we’re giving comfort on…An audit is what it is,” he said.
Where AI is having an impact: The pace of the work is quicker and the quality is more informative, he said. AI is also particularly good at pointing out to auditors where risks may lie in a company, he said.
But the full efficiencies of stripping time out of the process have yet to be realized, according to Panson. Accounting’s storied busy season, which typically runs through much of the first quarter as the audits are readied for 10-Ks, is so far only seeing a modest shift in terms of the time needed for the work to be completed, he said.
Asked if three months of work have been cut down to one month, he said the AI-related time savings have been more of a shift and less dramatic.
“It’s not that drastic yet,” he said. “I think we’re certainly seeing some time savings, but it’s an evolution and we would expect to continue to see that in subsequent years both from the improvement of the technology and as we get the investment and adoption behind us.”
