AI Is Co-Writing Financial Reports. Here’s Why That Matters
Oct. 31, 2025
The technology’s use in producing annual and quarterly filings—even with the requisite human oversight—raises questions of transparency and trust
Companies are increasingly relying on generative AI to help draft their financial reports, aiming to save time and money, but also raising questions on trust and transparency in executives’ efforts to communicate with their investors and regulators.
ON Semiconductor, a maker of chip components, has ramped up its use of AI to compile a draft of its financials over the past year, especially in sections that aren’t number-heavy, CFO Thad Trent said.
AI is “getting better and better” at writing the management discussion and analysis section, with ever less editing from the company’s regulatory-reporting manager, Trent said. Previously, the company would ask AI to write one paragraph in a particular section. “Now, we’re getting to the point of: Write the whole section for me,” Trent said.
The efforts are the latest sign of finance executives’ growing ease with AI for public-facing work that was long handled solely by humans. Companies are already using AI to prepare for questions that analysts might ask on earnings calls.
Financial reports are a crucial proving ground for AI. The Securities and Exchange Commission requires all U.S. public companies to file them quarterly and annually, including a balance sheet, income statement and management discussion and analysis. Investors rely on the filings to inform their decisions, and false or misleading statements can create both civil and criminal liability.
‘Much better than humans’
Besides providing text for financial reports, AI is also seeping into the number-compilation effort. In particular, it’s become helpful in tracking down information for journal entries, such as accruals for revenue and expenses, that are eventually tallied up on the balance sheet, ON Semiconductor’s Trent said.
AI’s assistance in financial reporting has helped the Scottsdale, Ariz.-based company reduce the time between closing the books and reporting quarterly results to roughly eight days from 10, with plans to pare that to six in the next year, Trent said.
“It’s doing it much more efficiently across the entire financial reporting cycle than what we’ve done in the past,” he said.
Another company, Hewlett Packard Enterprise, which has used its large language model for earnings-day prep, is getting ready to use it to formally produce a first draft of financial statements, possibly for the quarter ending in January, said Chief Financial Officer Marie Myers.
Myers aspires to have both the financial and non-financial parts of the company’s SEC filings written by AI at some point, she said, with people continuing to vet the work.
“You want to pick up everything that’s in your current earnings and make sure it’s all well reflected, and that’s what the AI is good at,” Myers said, referring to AI’s ability to mirror and aggregate data. “It’s much better than humans at doing that.”
The Spring, Texas-based enterprise-tech company currently uses an AI platform it co-developed to help its investor-relations team not only predict the questions analysts might lob on earnings day, but analyze their executive’s response for tone and delivery, Myers said.
AI’s role in the accounting that details a company’s financial position and performance in an SEC filing is minimal right now—but likely to grow, corporate advisers say.
Companies sometimes ask AI to calculate a journal entry to address a particular accounting issue, for instance, creating a sales reserve in compliance with French accounting rules, said Ash Mehta, a senior director analyst at Gartner. The finance team then checks if the suggestion would work—but AI generally isn’t posting an accounting entry to the books, he said.
At present, people are said to be “in the loop” on AI, keeping a close eye on it. But, Mehta said, “There is this idea that’s emerging of humans ‘on the loop,’ which basically means AI operates autonomously and humans step in only to check exceptions or override certain decisions.” He added, “This is a little while away from becoming commonplace.”
For now, executives seem to be approaching gingerly. In an impromptu survey in April, 28% of attendees at a Financial Executives International event said they use generative AI to help prepare external reporting, such as 10-Ks, compared with 16% who said they use it extensively and 57% not at all.
Others express greater confidence in the technology. At the beginning of the year, Goldman Sachs CEO David Solomon commented that 95% of an S1 filing, which a company submits to the SEC before going public, could be completed by AI in a few minutes.
Expanding uses, possible flaws
New tasks for AI in assembling financial reports are being found, but there could be drawbacks, some of them subtle. For one: Its robotic voice could change the tenor of how companies communicate with the public. Indeed, there is a risk that trust between management and investors could erode over time, said Keren Bar-Hava, head of the accounting department at the Hebrew University Business School.
“If disclosures begin to feel formulaic or emotionally empty, investors may disengage—not because they don’t care, but because they no longer believe management disclosures tell them anything real,” Bar-Hava said. “The danger is when AI replaces critical thinking and honest managerial voice.”
The new uses come as President Trump has called for an end to quarterly reports, saying that reporting every six months instead would save companies time and money. Last month, SEC Chair Paul Atkins said he supports proposing a rule change to do so, though the agency hasn’t put one forward yet.
Any shift away from quarterly reporting could further spur companies to replace finance people with AI to help save money, said John Coffee, a law professor at Columbia University.
The accounting firms tasked with auditing companies’ financials are also stepping up their use of the AI in that work. But, Coffee warned, “AI is not foolproof and they don’t have the same depth of knowledge at least initially as the Big Four.”
Public companies are expected to disclose the use of AI when it is material to their business—but aren’t specifically required to tell whether or how the technology assisted in compiling financial reports.
Still, companies could burnish their services to investors if AI freed up time for companies to issue custom financial reports intended for different types of shareholders in the next several years, said Dilshoda Yergasheva, head of financial services at Writer, an enterprise AI startup.
“Imagine you have to do a hundred or a thousand different versions to meet your shareholders where they are. But with AI, I can see it being done,” Yergasheva said. “A new way of reaching your shareholders could be how it’s done in the future.”
[Wall Street Journal]
