We extended the use of our Clara analytics AI Transaction Scoring (AITS) solution to over 55 audit engagements this year. KPMG launched AITS globally in 2023, in alliance with MindBridge, a third party specialising in financial risk discovery and anomaly detection.
AITS, a world-first AI-enabled solution, allows our auditors to better address risks of error related to revenue and operating expenses and to more accurately identify high-risk, or outlier, journal entries, for further testing.
By analysing 100% of transactions, our audit teams can significantly reduce the reliance on traditional audit ‘sampling’ techniques to obtain sufficient appropriate audit evidence.
For one new client, AITS has been successfully used to analyse revenue from sales across more than 600 petrol stations. Previously, a reconciliation of cash to revenue would have been performed – but the promised usage of AITS, underpinned by the audit team’s previous experience using the solution on other retail clients, was key to KPMG’s approach and proposition.
The audit team sat with the client and reviewed the graphs and analyses that AITS generates and was able to challenge and investigate outlier transactions. Both sides were able to better understand some of the nuances in the client’s revenue streams. This is a far more robust approach than the traditional method of performing a cash-to-revenue reconciliation, or sampling transactions, which may involve limited, physical interaction with the client.
Using AITS, our auditors may see, for example, 85% of their revenue transactions scored as having a ‘lower’ risk. The remaining 15% of transactions will be spread across ‘medium’ and ‘high’ risk groupings because these have been identified by the AI solution as having more unexpected or anomalous characteristics, based on the individual data and analytic techniques run.
These scores require auditors to assess what specifically differentiates these higher-risk transactions. Based on that understanding, auditors may perform further traditional testing to get sufficient appropriate audit evidence over those items.
This prompts meaningful conversations with clients, with KPMG auditors better able to explain why certain transactions are being selected for evaluation. In several cases, by isolating these transactions AITS has provided valuable insights for clients by revealing potential weaknesses in internal company processes.
AITS is system-agnostic, meaning that it is accessible to more audit teams than other data and analytic routines, which can require the development of bespoke routines.
It is important to appreciate that the scores provided by AITS are not used in a vacuum or with undue reliance on AI; auditors understand the scores and use them to help determine what further audit procedures may be required. The human element is still crucial.
