AI in Auditing: Traditional vs Prudent AI-driven Audits



Challenges to the Audit Profession


In the real world, Accountants are true domain experts with in-depth knowledge in finance, audit and taxation. When an auditor with tons of experience examines a transaction, he can immediately determine whether it is suspicious or not. All organizations will have suspicious transactions and as domain experts, auditors challenge themselves to find such transactions. However, with too much data, too little time, extreme dependency on experienced audit professionals, and limitations inherent in the internal controls of clients, it becomes impossible for auditors to review all such transactions in an organization and find the needle (suspicious transactions) in the haystack.


Hence, using their gut and rich experience, and by adopting traditional sampling techniques, they try to identify such transactions and misstatements. This cannot provide absolute and guaranteed assurance on their findings. Auditors have been operating under the philosophy of reasonable assurance for decades and use their professional judgement to fill in the gaps. Regulators are also turning the heat as the quality and reliability of audit has come under the spotlight on too many occasions. In fact in a recent report submitted by Sir Donald Brydon , author of the report into the effectiveness and quality of audit in the UK, he highlights the changing public-interest responsibilities of business including their corporate reporting and is challenging auditors to play a pivotal role to meet these changing needs and restore public trust.



Artificial Intelligence can transform Audits


AI empowers auditors to mitigate all the components of the audit risk – Inherent, Control and Detection risk. With AI, the growing amount of client data and time constraints and years of experience for an auditor are less of a concern as the work is offloaded to the application to test the complete ledger data by going beyond traditional sampling techniques to identify misstatements based on risk analysis rather than traditional audit rules. Experienced auditors are extremely valued as their brains have been constantly learning the various patterns over the years. Similarly, AI draws on the knowledge of real-world professionals and practices to identify unusual patterns.

The power of AI is that it is constantly learning about and identifying patterns in data. As these systems learn more about client transactions, they are able to analyze secondary data and cross-correlate hundreds of variables to identify rare items using empirical methods, to determine what is usual versus unusual. These methods can detect outliers in transactions without bias or history which goes beyond what a human is able to detect.



Prudent AI gives Auditors Superpowers


Prudent AI has developed an expert system by working with audit practitioners to understand suspicious and risky transactions, based on hundreds of account interactions and their associated concerns. It unlocks the superpowers of the auditors by using their rich experience and judgement only to focus on key risks and issues that matter rather than executing manual mundane tests to arrive at such insights. From gut, prior experience and expertise, we’re now shifting towards consistent data-backed, technology-enabled auditing process which ensures that potential loopholes are identified very early in the audits.


Prudent AI can be embedded into both planning and execution phase of the audit process. During the planning phase, the audit teams remotely gain key business insights, critical risks and outliers by scanning complete client’s data. Using this analysis, the audit teams execute razor sharp focused audit tests to determine if the potential risks have materialized. Auditors gain more comfort as 100% of transactions are tested during execution phase instead of sample-based tests. Since control failures are detected early, the auditors can flag concerns and discuss with the management to take prompt actions.



A sample Use Case


COVID-19 has affected the growth and operations of all businesses. While understanding a business and its environment, the auditors identify that market conditions have deteriorated in the last 3 months and will continue to get worse in the upcoming 6 months.


So, the auditor identified the presence of an inherent risk in the business during his planning process – Customers might not have paid the outstanding invoices. He instructs his team to investigate and determine the following points during the field work:


      • Has the management appropriately considered the market deterioration while creating provision for doubtful debts?
      • Will trade receivable valuation have higher inherent risk assessment?
      • Should detailed procedures be applied to find out any evidence of misstatement in trade receivables accounts?

During the audit, the auditor will perform audit checks and procedures on the data, derived from sampling techniques, to identify any material misstatement present in the period-end balance and collect substantive evidence. Also, the audit team will examine a handful of invoice transactions to identify any revenue transactions which have been recorded during this period.


Overall, this process is highly labor-intensive, expensive, and leaves significant audit risk of an incorrect opinion by the auditor.



How do we do it?


Prudent AI runs a deep scan on 100% transactional data instead of relying on sampling to perform following main tasks:


  1. Detailed Analysis of ledgers and data-points: Prudent AI expert system combines machine learning with traditional rules-based algorithms to slash the time, efforts and risk involved in otherwise labor-intensive examination and comparison of data.

Prudent AI’s General Ledger (GL) Module examines the journal entries and trial balance to provide significant analysis of data points in balance sheet, income statement, cash and bank accounts, significant ledgers, accounting ratios and co-relation between two or more accounts.


Similarly, Revenue and Receivable Module monitor various activities pertaining to revenue and receivable process which includes customers review, critical path of revenue process understanding, transaction type analysis, matching of invoices to deliveries, ageing analysis, allowance for doubtful debts analysis, etc. It presents the key insights through trending charts and other visual graphs for the auditors to understand key business insights and outliers.


  1. Assignment of risk score to prioritize high-risk items: Prudent AI assigns a risk score to each transaction and customer divides them into high, medium and low buckets from where the auditors can attend the riskiest transactions and customers first followed by others. Presented in an interactive tabular form, risk score is backed by an explanation to enable the auditor to understand the possible control point failure.

  1. Filtration through custom criteria: Prudent AI is a powerful enabler, not replacement. Based on insights delivered by the platform, auditors can apply their rich professional experience to narrow down the risk by applying custom criteria like materiality, normative flow between accounts etc to eliminate the high audit risk present in transactions and processes.

  1. Concise briefing to audit teams: Even before initiating his fieldwork, using the systems’ and custom analysis, the auditor is remotely able to reasonably identify the areas with high risk of material misstatement in trade receivables. It enables them to give laser-focused guidance to the audit teams to investigate, analyze and determine if the risk has materialized in provision of doubtful debts. This drives a more focused audit, allows Auditors to adjust the timing of their audits and address areas that require attention well before year-end time pressures.

  1. Perform In-depth and faster substantive tests: While Prudent AI platform facilitates a top-down, risk-based approach, it is also used to support control testing as well as substantive procedures in the execution phase. Since it covers the entire population of testing the entire process of invoice to collections and results in performing In-depth assessments that focus on higher risk transactions and customers, it creates higher quality evidence for an auditor. Given it is automated through Prudent AI platform, it is faster and improves consistency, efficiency and quality of procedures. It also allows Auditors to provide early insights to optimize processes and controls and discuss key issues with management early, iterate, and finalize.


Conclusion


Auditors can deliver sharper audits consistently with Prudent.AI. In summary, Prudent AI Driven Audit, enables an auditor to significantly reduce audit risk and deliver a higher quality audit efficiently by:


confidence
feedback
trends
analysis
time

Providing greater confidence in reporting as it helps identify hidden patterns and trends in data to provide better insights.

Providing relevant feedback and insights during the audit, to help optimize business processes and controls.

Identifying trends and anomalies in processes and controls, to help direct investigative effort in the right areas.

Allowing a deeper understanding of business by analyzing larger, relevant data to present a fuller picture and identify the main risks.

Saving valuable time and effort through globally integrated data capture and extraction tools.

*****

Interested in knowing more? Request demo of the product.