AI in audit: from vision to practice

The meteoric rise of Artificial Intelligence (AI) within the auditing profession is causing enormous change. Not only in how audits work, but just as much in the role of auditors and the organisations they work for. According to KPMG experts Marcel Boersma and Anastasia Priklonskaya, AI offers enormous opportunities, but as with any technological breakthrough, these opportunities do not come without challenges. According to Marcel: “The questions raised by many organisations include: How can we concretely get started with AI? What are the benefits for us? What is the role of directors and, most importantly, what are the steps needed to effectively integrate AI into audit processes? How can we bring new challenges such as the data needed for ESG reporting into this story?”

The benefits of AI in audit are tangible, especially in the areas of quality and efficiency. For example, recent research by KPMG found that making better data-driven decisions and being able to predict trends are seen as key benefits by organisations. Where auditors used to work with samples, AI now allows them to analyse entire data sets. This provides a more complete and accurate picture of an organisation's financial situation, allowing them to make strategic adjustments as needed. Anastasia explains: “AI rapidly detects patterns and anomalies that would otherwise go unnoticed, increasing the quality of audits and enabling timely corrections. Through predictive AI tools, organisations are continuously 'in control' of their financial integrity.”

The deployment of AI is not only changing the way audits are conducted, but thus the role of the auditors themselves. Previously, time-consuming, routine and repetitive tasks were the norm. But using AI for that type of work – which AI is much better at anyway – allows them to focus more on identifying risks and making connections beyond the numbers. This is where the major difference between human intelligence and artificial intelligence immediately becomes apparent. AI ‘thinks’ based on algorithms and always gives the ‘most logical answer’ based on that. Human intelligence is all about analytical ability, emotion, creativity and intuition. Within the audit profession, employing that very combination is imperative. In short, leadership style, vision, courage, experience – these are important ‘data points’ for auditors that are less likely to be seen by AI, but for which there is more time to consider thanks to the deployment of AI.

One thing is clear: AI is only as strong as the data it runs on. For organisations, this means that ensuring the quality of their data is essential to the success of AI-driven audits. This starts with digitising financial records and ensuring that all relevant data is available in a uniform and accessible format. Anastasia says: “Without high-quality, well-organised data, it is difficult for AI to deliver the desired results. In addition, organisations need to understand that AI tools can only be effective if the underlying processes are clearly defined and consistent.” This requires a change in mindset within many organisations, to increase their focus on the importance of standardised processes and data integrity. Marcel adds: “Too often, there are exceptions where standard processes are deviated from. It must become clearer internally why this uniform approach is so important and what the benefits are. Incidentally, AI can also help with that, by questioning employees when they deviate from the process.”

To successfully integrate AI into audit, an important task lies with the auditor. For example, research shows that more than half of organisations (64%) see auditors as the driving force behind applying AI in audits. For optimal results, though, it is important that the systems of organisations and the AI capabilities of their auditors be maximally aligned. This can range from simply digitising documents to implementing sophisticated data integration systems that can share real-time data with the audit firm's systems. The more alignment, obviously, the higher the quality and effectiveness that can be achieved.

To successfully implement AI in audit processes, there are technical and organisational challenges to overcome. The aforementioned KPMG study shows that organisations experience difficulties in collecting relevant and consistent data. Marcel recognises these challenges and explains: “Many organisations have old systems that are not easy to integrate with modern AI solutions. The quality of the data itself is also regularly a major challenge, especially in organisations where processes are not yet fully digitised.” And while many organisations do realise the benefits of AI, a structured plan to effectively integrate this technology is often lacking. The role of directors in implementing AI is crucial in this regard, but not without challenges. According to Anastasia: “Directors can struggle with the rapid pace of technological change – especially around the question of what are long-term developments and what are temporary trends.” All in all, successful deployment of AI requires an open attitude and a willingness to take risks and experiment with new technologies.

The question from many organisations right now is, 'where do we start?' Marcel says: “The good news here is that not everything has to be changed immediately. At the same time, it is important to take steps quickly so that as an organisation you are well prepared for what is to come.” Here are the most important actions to start with:

1. Where do you stand as an organisation?
   The first step is clear: Examine where you are in terms of ‘digital maturity’. Insight is extremely important to get a good picture of what is still needed. On the basis of those findings, create a clear roadmap outlining the ambition for the next three, four or five years and what steps are needed to achieve that.
2. Start with small experiments
   For organisations that want to get started with AI in audit processes, it is wise to start with small, defined experiments. This can be done, for example, by selecting a specific audit function where AI can add value and implementing it step by step. By starting small, organisations can assess the effectiveness of AI without taking major risks.
3. Invest in data quality and process optimisation
   As discussed earlier, data quality is critical to the success of AI. Organisations must therefore invest in digitising and organising their data. It is also important to optimise existing processes so that they are better suited to the capabilities of AI. This can range from standardising data collection processes to implementing advanced data integration systems.
4. Training and culture change
   AI requires new skills and a different mindset from many people within the organisation. This means investing in training staff so they can use the new technologies effectively and, just as importantly, interpret the results correctly. It is also important to foster a culture of experimentation and learning so that AI implementations can be successful.

AI offers enormous potential for improving auditing processes, but its success depends on the willingness of organisations to invest in the right technologies, processes and people. By starting with small experiments, focusing on data quality, and embracing a culture of innovation, organisations can get the best results from AI in auditing.

Want to learn more about this topic or have questions? Download the full research report on AI in auditing here, or contact Marcel Boersma or Anastasia Priklonskaya.