Artificial Intelligence and Internal Audit

08/05/2017 03:32 am ET
Copyright Swiss Banking Advisory

Auditing is about analyzing, being able to collect information around the audited subject and understanding its connections to other relevant subjects or areas. Going forward, auditors will not only uncover issues and errors, but will also provide solutions. This means that the reports from internal audits will not only list errors and process flaws, but also potential solutions to issues in collaboration with the experts from the audited area...

In most cases, the issues addressed by the auditors are known to the audited area, but in the day-to-day context of business activities, these issues are mostly not considered as urgent. In the future audit needs to adopt its approach to generate a benefit for the audited area too. The most important task of an internal auditor is to be able to analyze the collected information, while the question part of an audit can be done by a junior auditor. By repeatedly asking “why?”, an auditor can collect large amounts of information which helps to understand the entire landscape around a subject. It enables the auditor to evaluate the facts and make assessments.

When I was working as an internal auditor, I was involved in a project that was searching for an early warning system using available technology. We were tasked to ask simple questions and to evaluate the collected facts about a particular subject. In contrast to Eliza which became famous decades ago for being a revolutionary IT solution, the operating system featured in the movie “HER” did not provide a revolutionary new insight into the latest AI (Artificial Intelligence) technology.

Humans can differentiate between a conversation with a human and a conversation with a robot. Furthermore, they are able to make jokes, write poetry or recognise a lie by a person’s voice or mimic. Can AI such as Eliza take over the activity of internal auditing? As mentioned, by asking the simplest questions, almost everybody can access information. How intelligent must AI become in order to be able to act as an internal auditing system, and what would be the role of humans in this process? I think that by even having simple artificial interaction software in place, the interview part, as well as the structuring of the collected information, can be taken over by a computer. The collected information can help clarify subjects and to draw conclusions about the problem.


Here is an example of how a simple question can be asked to collect all needed information about a particular issue: The issue is that I get up early in the morning.

AI: Why? Answer: Because I like the early morning energy and silence.

AI: Why do you like silence? Answer: Because, if it’s quiet, I get into a different state of mind with little effort.

AI: Why do you need a different state of mind? Answer: Because in a different state of mind I can see ordinary things from a new perspective.

AI: Why do you need a new perspective? Answer: Because different perspectives can reveal new solutions to problems.

AI: Why? Answer: Because new solutions will give me the ability to better solve outstanding issues.

AI: Why? Answer: Because by better solving my issues I have a better day ….

In practice, questions can be chosen in a way that the person being interviewed does not discover that the questions are asked without the answers being listened to.

After the financial crisis, the area that would benefit the most from the introduction of AI into its processes is internal audit. This way, existing resources can be used in a more efficient way and it would be possible to audit more areas in shorter period of time. In the future, internal audit will use software like Eliza to interview experts from all areas almost monthly and will be able to collect information by setting up an early warning system for reporting by searching for critical words.

Areas with the most critical words will be audited with the most urgency. The interview can also include hidden checks to ensure that the person understands the answers being given and validates the truth. This will help to control more efficiently, without any additional resources. Furthermore, issues can be found more quickly, potential losses can be detected much earlier and before they cause damage to the organization...

Source: Banks of the Future, by Ella Thuiner; Published by Springer, © 2015

This post was published on the now-closed HuffPost Contributor platform. Contributors control their own work and posted freely to our site. If you need to flag this entry as abusive, send us an email.