KPMG is rolling out new AI features to Clara, the internally-developed software platform its auditors use to automate repetitive parts of their work.

Several of the additions are AI agents. An AI agent is a neural network that a company adapts to a specific use case, usually by putting it through additional task-specific training. Clara is also receiving a machine learning tool called FRA that will help auditors prepare financial disclosures.

KPMG created the first iteration of Clara in 2017 through a partnership with Microsoft. Initially, the focus was on providing an interface through which KPMG staffers can provide clients with updates about the status of an ongoing audit. The Big Four firm has enhanced Clara multiple times since, including through a 2024 update that added the platform’s first set of AI features. KPMG’s newly announced upgrade extends those features. 

Auditing a company’s books involves a significant amount of manual work because there is a large number of transactions to review. The new AI agents included in the latest Clara release are designed to automate some of that work for KPMG’s audit teams. The firm employs more than 95,000 auditors worldwide. 

According to KPMG, one of the tasks that the new AI agents will streamline is expense vouching. This is the process of double-checking the expenses listed in a company’s books for accuracy. It’s often one of the most time-consuming aspects of the audit workflow.

The system in which a company keeps accounting data is known as the general ledger. Typically, the general ledger organizes each corporate expense in a separate spreadsheet row. This spreadsheet row might, for example, include the sum that a company has paid for Microsoft 365 licenses and the date of the purchase. Performing expense vouching on such a hypothetical ledger entry would involve checking its accuracy by comparing the price and purchase date against the recipient that Microsoft issued for the transaction. Other types of documentary evidence besides invoices may be used as well.  

According to KPMG, its new AI agents will also help its auditors sort out accrued expenses and unrecorded liabilities. An accrued expense is the term for a situation in which a supplier has provided goods or services to a company, but the company hasn’t yet paid for them. This might happen when the supplier doesn’t send over its invoice before the company closes the books on the quarter in which the deal occurred. An unrecorded liability, meanwhile, is essentially the opposite of an accrued expense. The term describes situations where a company has paid for something but hasn’t yet added the transaction to its ledger. Fixing such gaps is a major focus of financial audits.

The new AI agents are rolling out to Clara alongside another machine learning feature called the Financial Report Analyzer, or FRA. It helps auditors work with required disclosure checklists, documents that specify how a company should go about structuring financial disclosures such as earnings reports.

“AI is freeing up resources for our auditors to spend more time on the areas of highest risk, sector-specific risks and challenges,” Thomas Mackenzie, KPMG’s U.S. and Global Audit Chief Technology Officer, told Boardroom Insight. “The integration of AI agents into KPMG Clara, along with our FRA AI engine, will empower our professionals to continue enhancing quality and uncovering deeper insights for audit committees and management, all while exercising professional skepticism and maintaining a human-in-the-loop approach.”

KPMG plans to integrate more AI agents into Clara over the next year. The additions will focus on easing tasks such as testing financial controls, the internal process through which a company ensures that its financial statements are accurate.  

The Clara enhancements tie into a broader effort on the firm’s part to integrate AI into its business operations. “These AI upgrades are part of our ecosystem of solutions at KPMG and are core to our broader firmwide commitment to enhancing delivery with AI across audit, tax and advisory,” a KPMG spokesperson explained. 

The Big Four firm previously developed an AI tool called Advisory Content Chat to help staffers find internal data more quickly. The software is powered by Microsoft’s public cloud. KPMG has also partnered with Google, one of Microsoft’s top competitors in the cloud market, to support its machine learning initiatives. The Big Four firm plans to use AI features in Google Cloud to help its newly launched legal services division streamline tasks such as contract reviews.

text

text