Share This

I had the opportunity to meet with and provide a perspective on this topic to several dozen auditing and compliance professionals at a recent conference in Rochester, NY. Interest levels were very high, and leaders from several major firms continued our discussions afterwards. Here is a recap with examples of breakthrough results, as well as an industry leader’s perspective on the impact that AI will have in elevating the technology skills that auditors will need to work with and analyze new, very large datasets.

The Perfect Storm for AI

The dramatic rise of AI is the result increased Processing Power, Analytics Advances and Big Data. Today’s advanced artificial intelligence technology has 3 transformational capabilities:

  • Natural Language Processing
  • Understands unstructured data
  • Self-learns from new data

AI Value Chain – Audit, Risk & Compliance

The AI Value Chain looks like this:

AI Value Chain

Breakthrough Results – Examples

    • User Experience – Augmenting the Auditor
      • We’ve seen that machines can do some things much more efficiently than people, such as using NLP for document text analysis. This can be 3x more accurate and 2x more efficient.
      • Innovation is enabling auditors to deliver powerful insights that simply weren’t possible before. For example, using the latest technologies, auditors can analyze complete data sets rather than samples. Advanced tools can be applied to all of a company’s contracts related to an area of audit interest, or to metadata about an automated key control. This can reduce audit risk by making it less likely for an unusual transaction to slip through the cracks.(1)
    • Automation – Streamlined Compliance & Auditing Automation
      • Natural Language Processing (NLP) for Document Review – Identifying change controls in hundreds of thousands of legal documents used to take dozens of Deloitte employees a half a year. Now, team of six can use an AI system to complete in less than a month.(2)
      • Machine Learning for Anomaly Detection – EY’s Invoice Fraud Detection machine learning system has a 97% accuracy and has been rolled out to over 50 companies.(2)
    • Analytics – Cybersecurity & Threat Detection
      • Threat detection is certainly a main focus of today’s AI and machine learning push. Some are beginning to focus less on sealing borders from outside threats and more on sensing bad behavior inside as it happens—when it can be stopped. They use machine learning to define what “normal” looks like for any network and all its devices and then report on deviations and anomalies in real time.(3)
      • “The complexity and scale of our systems are too overwhelming and too volatile. Traditional rules-based systems overwhelm our security analysts with tons of noise that allows breachers to get past them because they are simply not able to look at them all. Machine learning is the key to finding high fidelity signals that are comprehensive in nature, that are not based off of one event stream but a collection that all together signify a breacher being active on the system.” Mark Russinovich, CTO, Microsoft Azure.(4)
      • “In Microsoft’s Azure system, a former rules-based system scored 28% of logins as suspicious. One billion logins per day = 280 million “suspicious” logins. After applying Machine Learning, the rate dropped to 0.001%.” Mark Russinovich, CTO, Microsoft Azure.(4)

AI’s Significant Impact on Auditors

Today, AI’s natural language processing, machine learning capabilities and image recognition advances are rapidly being deployed across a wide range of business functions. This has created new challenges and opportunities for those who provide the audits, and manage the risks and compliance related to these new capabilities. Jon Raphael, CPA, partner and audit chief innovation officer of Deloitte & Touche LLP included this assessment in his article.(1)

“Auditors don’t necessarily need to be technology development experts or computer programmers; however, they do need practical knowledge, experience, and a high level of comfort using cutting-edge, rapidly evolving technology to manipulate and analyze data.