EY, KPMG upgrade AI capacities for auditors
May 21, 2024
Big Four firms EY and KPMG announced recent upgrades to the AI tools used by their auditors, reflecting both firms' embrace of the new technology.
KPMG announced upgrades, specifically, to KPMG Audit Chat, the firm's proprietary generative AI tool, and to KPMG Clara, its global smart audit platform. These upgrades coincide with the deployment of Copilot for Microsoft 365 to all KPMG partners and professionals.
While KPMG declined to share information about all of the upgrades, they include the ability for audit professionals to: record and analyze discussions with management to understand a firm's processes, risks and more; automatically create process maps from a recorded walkthrough meeting or transcript; more effectively review their audit documentation compared to the auditing standards and KPMG methodology via an enhanced gap analysis capability; extract, identify and summarize relevant contract terms for analysis and in year-over-year comparisons; and assess financial statement disclosures against peer company disclosure requirements.
These changes come as the firm's auditors increasingly use AI in the course of their work. Since October 2023, when KPMG launched Audit Chat, partners and professionals have engaged in more than 600,000 Audit Chat conversations to more effectively search databases, extract and summarize data from contracts, analyze documents and aid with research on complex accounting and auditing topics. A KPMG spokesperson said all of the firm's technology, including its generative AI solutions, are deployed within the requirements of its established software audit tool testing policies and procedures.
Audit Chat was piloted and fielded after extensive review. Its prompt-engineering process fielded and evaluated potential prompts and engaged reviewers with different skill sets to drive quality outputs. Its prompt library now includes over 100 prompts for auditors. Auditors using the solution require upfront training in responsible use of AI, and all final decisions and documentation are people-validated. The spokesperson said that, in fiscal year 2023 alone, audit partners and professionals spent 23,000+ hours collectively learning about the connection between AI and ethics.
"We continue to transform our audit globally to be AI-powered and people-driven, building on our existing AI transaction scoring capability," said Thomas Mackenzie, global audit chief technology officer for KPMG. "While we've moved quickly to test, pilot and deploy AI in our audits, where we are today is just the tip of the iceberg. As our technology teams and auditors gain more experience with generative AI, we expect to accelerate innovation across our audits."
EY also announced generative AI upgrades to the assurance service line last week for its 9,000 technology risk professionals. The updates include the release of more than 25 new EY Assurance technology capabilities, the fruit of a multiyear investment of more than $1 billion into generative AI.
Like KPMG, EY declined to provide a list of all of these new capacities. A spokesperson said that, in general, the latest releases and pilots are clustered within three domains: 1) data and analytics; 2) intelligence and automation; and 3) experience and collaboration. Within each of these domains, EY has released new technology including AI-enabled capabilities; new guided workflow capabilities in EY Canvas (EY's proprietary cloud platform); and new capabilities within EY's cloud-based analytics.
New data and analytics capacities include the ability to evaluate IT process-level access rights and configurations as part of financial audit IT procedures, as well as to share data analytics with clients. Regarding intelligence and automation capabilities, professionals now have access to AI-enabled capability to support financial statement tie-out procedures as well as generative AI enabled capability to search and summarize technical auditing and accounting content. Finally, regarding experience and collaboration capabilities, EY professionals have expanded capabilities for audits of non-complex entities, as well as new capability to enable audits of complex entities.
This announcement was paired with the release of an AI Assurance Framework, which provides internal guidance for EY professionals as they address AI technology developed by the companies that the firm serves, according to an EY spokesperson. The framework is viewed not as the final word, but as a starting point for EY Assurance considerations. It will evolve over time based on the development of auditing standards, evolving laws and regulations, the development of organizational guidelines, and the experiences of audit teams as they encounter AI applications during their audit.
This AI Assurance framework builds on EY's Responsible AI principles, and guides EY professionals to assist in matters pertaining to identifying the risks of material misstatement arising from management's use of artificial intelligence applications within the financial reporting processes, considering the types of management controls expected to address the risks introduced by the use of AI, determining the nature, timing and extent of EY teams' responses to the identified risks of material misstatement. Risks considered through the use of the AI Assurance Framework include design risks, data risks, algorithmic risks and performance risks. The Assurance AI Framework also addresses considerations related to bias, resiliency, explainability, performance and transparency.
"The impact of AI is creating new risks and opportunities for organizations," said Paul Goodhew, global assurance innovation and emerging technology leader for EY. "EY teams are increasingly leveraging the potential of this technology to support the provision of assurance services and utilizing that experience with EY clients to provide confidence in their responsible use of this emerging technology."
[Accounting Today]
