How Accounting Firm Owners are Thinking About AI's Potential for Auditors

AI’s impact on the workplace and society quickly became one of 2023’s hottest topics and one ripe for speculation. Into the new year, while the advancements of AI’s capabilities in certain fields are already being felt, the story of how this new technology will impact the working lives of auditors and accountants is still one that remains largely yet to be written. As firms position themselves as experts in the field, ready to sell lucrative advisory services to clients, everyday auditors wonder what the future of their profession holds in a landscape of generative AI. AccountingToday asked firm leaders what they anticipated will transpire in 2024 with regard to AI for the industry. Not surprisingly, many responses included, either explicitly or otherwise, mentions of the ability of AI to take on the menial tasks and free accountants to be more productive in other areas or perhaps even achieve greater work-life balance.

‍

How ready are both large and small firms to deliver on the vision of AI they’re predicting? We wanted to take a deeper dive into what industry leaders are saying about AI and how they are positioning themselves to lead the charge of these innovations. First, returning to what leadership, including several from the Big Four told AccountingToday, the transformations will allow auditors to focus on the most valuable components of their jobs and free up capacity to deliver a better experience for their clients. Taking a step further, it's clear that in some capacity leadership sees AI as a means to allow more parts of firms to serve in an advisory role by shifting some of the compliance responsibilities and tasks onto AI. Putting aside independence concerns, the potential for AI to allow members of firms to focus on higher margin projects would certainly be particularly impact firms struggling to maintain margins in core assurance practices.

‍

Coinciding with these efficiencies exists is the component of AI capabilities that firms look to sell externally. Often by partnering with a large LLM developer, the initiative to sell solutions to clients who are looking for solutions to complex accounting topics using LLMs specially trained on financial reporting standards has already been launched by a few firms, including Marcum (which is launching their service first with Marcum employees, integrated into Microsoft Teams). While this service may prove beneficial in scaling some amount of financial accounting advisory projects, it’s difficult to gauge how receptive clients are to taking the recommendations of a generative AI onto their financial statements. Not surprisingly, the discomfort with a real or perceived lack of human oversight ranked highly in CFOs reservations about using AI. Ultimately however, where and how much oversight auditors and accountants have over AI will be determined by what tasks it ends up performing for firms. Revisiting the potential to automate repetitive tasks, despite being a major focus of industry leadership’s prediction, there has been less public roll-out of such capabilities to date. Perhaps an LLM that has mastered the nuances of US GAAP disclosures is more impressive than one that can trace or vouch client data, but there appears, at least from an outsider’s perspective, to be a disconnect between what firms are predicting and what is being brought to market.

‍

So where are firms spending their time and efforts with this technology? Are they rushing to be first to market with an AI analyst that can spot trends in data and make recommendations no human could possibly notice or generate? Or are they putting the investments into their own employees, hoping that the automation of the menial will unleash productivity (and, perhaps less overtly stated, reduce headcount). As mentioned with Marcum, firms certainly do see the impact of the technology within their own operations. Microsoft itself has even launched a finance specific instance of its Copilot product, trialed by their own finance team. Interestingly, the functionality is focused on those in industry, particularly with a lens on internal reconciliation and other month-end related tasks. Within assurance however, the question remains as to how exactly auditors will automate menial tasks when most at the moment are using an in-house or third party large language model. A chatbot is certainly useful for drafting certain wording but absent the ability to integrate seamlessly with the data provided as part of the audit, it will be difficult to automate even some of the most basic audit procedures. For assurance practices to realize the full potential of generative AI, and to deliver on the speculation of automating repetitive tasks, they’ll need to move beyond the input and output of these chatbots and think of ways in which AI automations can be built into existing audit procedures.  

AccountingToday also outlined how firms, particularly smaller firms, may be able to implement such functionality without developing it themselves from scratch. In particular, they mention the advantages of partnering with developers to develop prompting that can further refine these models to perform specific tasks. At Tellen, we believe that by working with our partners, we can develop refined prompt engineering within our applications to unlock the potential that so many in the audit industry are anticipating. We’re developing the tools to integrate our automations into every phase of the audit lifecycle and have already partnered with several firms who will help us develop and customize these applications to suit their needs. If you’d like to learn more about our process and see if Tellen could be a good partner for your firm, contact our Head of Growth, Jason Jones at jason@tellen.ai.