Audit and Artificial Intelligence

By July 08, 2016 0 Comments

While auditing can hardly be coupled with the word ‘entertaining’, companies and regulators take audits very seriously.

Auditors nowadays face the challenge of reviewing thousands of documents, scrutinising hundreds of contracts, reviewing the implementation of regulatory changes and undertaking investigations on ambiguous transactions. While auditors spend considerable time flipping pages and opening individual digital files to review them, it is arguable that artificial intelligence could help on some of those tasks.

AI is definitely not the right vehicle for very complex analysis, but for those repetitive, semi-defined tasks, a well-tuned system will start counting items, spot patterns, and flag anomalies to a level that meets and exceeds the requirements of a corporate exercise.

Cognitive systems can emulate structured processes and pre-defined tasks that normally are undertaken manually, collect information from thousands of documents and structure them in a way that brings anomalies faster to light. Further audits have historically involved taking samples of populations. With the power of big data and analytics, the auditor can choose to analyse all items in certain populations, not just a sample of them.

“With big data and analytics, auditors can choose to analyse all items, instead of just a sample”

AI can read thousands of contracts and extract what the auditors are after without the need to read them. Systems can also analyse entire sets of expenses, read the entire text, categorise the expenses and potentially expose non-authorised claims.

Systems can identify indications of fraud, just by reviewing changes in patterns of invoices over time or unusual invoiced items, and quantities based on previous history. Auditors have some leads to identify mismanagement and save millions to their companies.

“AI introduces the idea of continuous auditing and proactive analysis”

In addition, AI introduces the idea of continuous auditing and proactive analysis. Investigations can be programmed on a regular basis and auditors will have the opportunity to focus to more advanced analysis. AI can apply risk indicators to large datasets to detect risks that would otherwise remain hidden.

For the audit professionals, this can be translated both as an opportunity or a threat. Association of Chartered Certified Accountants has predicted that over the next two years automation will simplify processes such as bookkeeping and transaction coding, enabling accountants to focus on advisory services and other higher-value work.

To achieve this, auditors need to enhance skills such us data management and machine learning, something that universities do not currently provide. It is those professionals that understand auditing methodologies along with technologies that will show us the road to automation.

It is an exciting path!

This article has been first published in Digital Catapult Blog.