The software provider of a unified finance and HR platform announced on Tuesday it plans to acquire Evisort, a platform that uses AI to review documents and extract key data with a simple prompt. This takes aim at the pain points of analyzing unstructured data, like legal documents.

Evisort was founded in 2016 by Jerry Ting, Jake Sussman, and Amine Anoun. The three met while attending Harvard Law and MIT. Among Evisort’s clients are Microsoft, Keller Williams, and Workday (a CFO Daily sponsor).

To learn more about the strategy behind the acquisition, I sat down with Terrance Wampler, group general manager for the office of the CFO at Workday, during the company’s annual customer conference in Las Vegas this week.

Wampler discussed Evisort’s AI-powered tool that can read and analyze contracts, policy documents, and invoices across the organization. He noted the tool can also automate tasks like uncovering any anomalies in contracts, locating a specific section of legal documents, and finding patterns within the data that can inform accounting decisions and make recommendations.

An example? Wampler said that since the platform uses large language models, you can type a query such as: “‘Can you give me a summary of clauses that may be different? Can you go tell me if this rule exists?” The AI can analyze more than just formal type-written contracts—it can also read handwriting and complex tables, he said. Like with any automation tool, it’s viewed as an assistant to an employee, as Workday practices keeping a “human in the loop,” he added.

There are other document intelligence solutions out there besides Evisort. But a differentiator for Workday is that there has to be trust when handling customer data, Wampler said. “They’ve built that trust in the customers that they have today,” he said.

CFOs want to be able to move away from a lot of manual transaction processing and review. But 80% to 90% of business data is unstructured in the form of documents or even video and audio, according to analyst estimates. That means finance and accounting professionals may often need to manually search through company documents. According to Gartner, there are four top use cases for AI in finance: automation for information processing, anomaly error and detection, analytics, and operational assistance. 

Workday, a Fortune 500 company, said the acquisition of Evisort is expected to close in the third quarter of its fiscal year 2025, ending Oct. 31, subject to the satisfaction of customary closing conditions. The financial terms of the agreement were not disclosed. In 2022, Evisort announced it raised $100 million in capital. Workday’s most recent acquisition was HiredScore, an AI-powered talent acquisition platform, in April. Other acquisitions include VNDLY in 2021, Scout RFP in 2019, Adaptive Insights in 2018, and Platfora in 2016. 

By adding Evisort to the mix, Workday has provided a strong signal once again that the role of AI in the workplace is only going to grow.

Sheryl Estrada
sheryl.estrada@fortune.com

Quick note: We have a few seats left for the CFO Collaborative Dinner: Strengthening the CFO-CISO Alliance on Sept. 26 in Washington, D.C., in collaboration with Workday and Deloitte. I will host the event along with Fortune’s executive editorial director Diane Brady. We have lined up a great audience and speaker. Diane will interview J. Michael Daniel, president and CEO of Cyber Threat Alliance, and former cybersecurity coordinator of the White House Executive Office.

For more information, please email cfocollaborative@fortune.com

The following sections of CFO Daily were curated by Greg McKenna

Leaderboard

Xavier Heiss will retire as CFO of Xerox (Nasdaq: XRX) after over 30 years with the office equipment giant, effective Feb. 1. Heiss has served as CFO since 2021. He will be succeeded by Mirlanda Gecaj, who joined the company in 2022 and currently serves as vice president and chief accounting officer. Her previous stops include stints at PwC and chemical company Element Solutions, respectively.

Mitchell Chan was appointed EVP and CFO of REGENXBIO (Nasdaq: RGNX), a clinical-stage biotechnology company, effective immediately. He will succeed Vit Vasista, who is stepping down after serving in the role since 2009 and will remain at the company as an advisor through Jan. 3. Chan previously served as CFO of Viela Bio, where he oversaw the company’s IPO and eventual acquisition, and has held investor relations leadership roles at AstraZeneca and Genentech-Roche.

Big Deal

Over half of C-suite executives expect deepfake attacks targeting their organizations’ financial and accounting data to increase in the next 12 months, according to a new report from Deloitte. Out of more than 1,100 executives polled, roughly 26% said their organizations had experienced at least one such incident in the past year.

Just 46.5% of those surveyed said they are confident in their organizations’ ability to manage deepfake financial fraud threats. Executives whose companies faced such an incident in the past year were most likely to increase employee education about new threats.

“Deepfake financial fraud is rising, with bad actors increasingly leveraging illicit synthetic information like falsified invoices and customer service interactions to access sensitive financial data and even manipulate organizations’ AI models to wreak havoc on financial reports,” said Mike Weil, Deloitte’s digital forensics leader. “The good news is that concern about future incidents seems to peak after the first attack, with subsequent events tempering concerns as organizations gain more experience and become better at detecting, managing, and preventing fraudsters’ deepfake schemes.”

Going deeper

How Large Language Models Could Impact Jobs is a new article from the business journal of the University of Pennsylvania’s Wharton School. It highlights a recent paper from Wharton professor Daniel Rock, two OpenAI staff members, and an AI governance expert who covered U.S. government data on 21,000 tasks across 1,016 occupations over three years, starting in 2020.  

They found roughly 1.8% of jobs could have over half their tasks affected by large language models, with that share jumping to 46% after incorporating likely future software developments to complement such models. According to Rock, however, alarmist rhetoric about these models replacing jobs is overblown.

Overheard

“Powell choosing between 25 and 50 basis points can be likened to a driver descending a winding mountain road going 60 mph when the speed limit says 25. Common sense says he should slow down immediately, not slowly decelerate to 55 as the road gets bumpier.”

— Jeremy Siegel, emeritus professor of finance at the University of Pennsylvania’s Wharton School and chief economist at WisdomTree, wrote in a Fortune opinion piece about why Fed chair Jerome Powell should consider deeper rate cuts to avoid risking a recession.