In November it will be two years since ChatGPT dropped, bringing the power of generative artificial intelligence (GenAI) into the hands of anyone with a smart phone and setting records as the fastest growing app in history.
But how is AI faring generally and what is it doing for businesses?
It’s a question a recent KPMG report, ‘AI in Financial Reporting and Audit, navigating the new era’ goes some way towards answering, charting both AI’s impact to date and its potential for the future.
The report is based on a survey of 1800 financial reporting executives around the world spanning six industries, including approximately 100 from Irish companies. It finds we are on the cusp of a genuine financial reporting revolution, moving from the ‘digital age’ to the ‘AI age’.
As with other business processes, it finds AI is transforming financial reporting and auditing, which traditionally involves time consuming and repetitive tasks.
The benefits we are poised to reap are enormous, including insight generation, ability to predict trends, real-time insights to risks and outlier detection, and better data-enabled decision making. All will combine to make financial reporting more robust and informative than ever.
It’s already happening. 6 out of 10 survey respondents are either piloting or using AI in financial reporting. What’s more, in nearly three years nearly all respondents expect to be piloting or actively using AI.
Companies are putting their money where their mouth is. Already AI accounts for 10 per cent of respondents’ IT budgets, a figure set to rise significantly in the next 12 months.
That’s because 65 per cent expect to reap the benefits of better prediction of trends and impact, 60 per cent see it providing real time insights into risks, 57 per cent see it enabling better data-enabled decisions, and, as well as increased data accuracy.
And they are right. AI should be seen as a tool to enhance capabilities and drive organisational success.
All this is likely coming, and more. But for the organisations keen to embrace it, getting it right will take time.
Hype cycle and managing expectations
Like any new technology, AI, and GenAI in particular, is subject to a “hype cycle”. From the enormous excitement of its launch, moving to a peak of inflated expectations, followed by a trough of disillusionment, as adoption and financial returns are slower than expected.
But rest assured, the remainder of the hype cycle is playing out too, as use cases emerge and move mainstream. Within the next two to three years, as these become embedded, businesses will start seeing massive upsides.
Indeed, benefits are already crystalising in the world of financial audit.
GenAI-enabled products, such as Microsoft’s Copilot, are already helping in myriad ways, from automating tasks to summarising documentation, preparing presentations or undertaking research.
The big step-change from previous technologies is that GenAI can “read” documents for you and respond to queries in an uncannily human manner.
But there is a drawback. For now, GenAI is not unlike an enthusiastic new colleague who doesn’t want to disappoint you, so they respond with answers that feel really professional, sound right, and are delivered with confidence. You still have a responsibility to ensure the responses are indeed correct, so you need to double check to make sure.
There are other hurdles too, notably around data integrity, security, privacy and sustainability. Therefore, it is important that senior leadership and board members assess the strategies and AI frameworks they have in place.
Taking the lead
At KPMG we have taken GenAI solutions and made them secure and are now safely deploying GenAI to answer queries about methodologies and accountancy standards, thereby further enhancing the quality of our audit work.
We are also leveraging machine learning and other AI technologies, including defined algorithms, which we can apply to our clients’ financial data sets. The technology allows us to apply multiple tests across all transactions to identify “higher” risk transactions and anomalies for further investigation. This leads to earlier focus on potential audit issues and reduction in substantive procedures when compared to a traditional audit approach.
The value of machine learning and algorithms, as opposed to GenAI, is that the procedures are repeatable and explainable. With GenAI the same queries can prompt different answers, sometimes incorrect ones. That’s the challenge.
With machine learning and algorithms, you don’t get hallucinations. Once you load up the data, if you ask it a question today, you’ll get the same answer tomorrow.
Reaping the benefits
Clients are impressed at the level of granularity AI provides, enabling us to spot patterns and anomalies that may not be uncovered through sample testing alone.
Because AI is more data driven, it takes away any biases that management may have, and uses the data to get directly to the facts of the matter. It also allows us to ask the important questions about any business much earlier in the audit process. Clients are seeing the benefits of that and are impressed by it.
For clients adopting and trialing new AI technologies they should consider the potential risks and barriers to its successful adoption within the business and financial reporting process.
At KPMG we help clients by assessing where they are in terms of their AI maturity level, ensuring that they have appropriate governance structures around data and security and investigating the potential business cases for the usage of AI within their business.
When it comes to AI we are all in learning mode. Some applications may work for you, some may not. It’s about having an open mindset and the ability to invest the time required, because it probably won’t work for you first time. But the more you try, the better it, and you, become.
We are already seeing clients reap the benefits of smart note taking and legal document reviews. These practical use cases are key, because the more organisations do, the more they understand how to effectively apply AI across its business.
All this takes work, and time. AI will be a completely transformational technology, just not in the timeline initially expected. It’s a slower burn than we all anticipated when ChatGPT dropped. But make no mistake, it will catch fire.