There’s a notable shift in the perception and acceptance of artificial intelligence unfolding within the accounting profession, as a growing number of firms explore the various applications of AI and generative AI, and the improved efficiencies and competitive edge to be gained from them.
“I think firms realize how much technology has transformed their practice over the past decade. We are in a very seminal moment here with this next game-changing capability, and I think firms of all sizes are very interested in understanding how they can incorporate it into their business and into their services to their clients,” said Erik Asgeirsson, president and CEO of CPA.com.
From search engines, to virtual assistants like Alexa, to self-driving cars, AI has become a part of our everyday lives, and the momentum shows no signs of slowing. Now throw into the mix generative AI, which can create various types of content such as text, images, music and code on demand.
In November 2022, generative AI ignited a frenzy with the debut of ChatGPT, a sophisticated, next-generation chatbot developed by OpenAI. It was swiftly hailed as the fastest-growing consumer application in history and attracted more than 1 million users in just five days. Today, according to recent statistics from Exploding Topics, ChatGPT has more than 180 million users.
Many accounting firms have taken a wait-and-see approach, especially as it relates to generative AI. However, there now appears to be a momentous shift underway as a growing number of firms are taking a deeper dive into AI and are eagerly exploring how it can fuel greater productivity and profitability within their practices.
The next frontier
While the various applications of AI within the profession are still playing out, what is clear is that the innovative technology will have a profound impact on the industry and on daily life, much like the invention of the PC, the internet, and cloud computing.
MangKangMangMee – stock.adobe.com
“I think we are at an inflection point. This [AI] will change the industry in a material way, just like the computer did,” said Bill Armstrong, chief innovation officer for Top 100 Firm Moss Adams in Seattle. In fact, a recent Moss Adams-commissioned survey of 500 corporate tax and auditing professionals found that a large majority (79%) of accountants believe it is beneficial to have AI assist them with their job. Furthermore, well over half (67%) said they expect their companies to boost investments in AI in 2024 through expanded implementation, partnerships with AI software companies, or further research. According to the survey, respondents said they interact with AI through work devices (50%), apps and tools (43%), and AI models like ChatGPT (50%).
Furthermore, the “Tax Reimagined 2023: Perspectives from the C-suite” survey from Big Four firm KPMG revealed a strong willingness among C-suite executives to embrace AI, with 99% saying that AI is the “next frontier” for corporate tax departments. More than half (59%) of the C-suite leaders surveyed said they are already using AI in their tax or finance departments to improve workflow efficiency and ease the strain on existing staff. And of the 41% who are not yet using AI, all are interested in doing so, according to the survey of 500 C-suite executives at organizations with $1 billion or more in revenue.
However, that’s not to say that reservations on the use of AI within the profession no longer exist. As revealed by the Moss Adams research, 89% of accountants expressed at least one concern about AI. Top-cited concerns included:
- Work quality (42%);
- Data accuracy (41%);
- Fiscal costs (31%);
- Undetected bias (26%); and,
- Ethical issues (21%).
Therefore, it is critical that firms think about best practices and have the proper AI guidelines and policies in place, as well as factors such as privacy, transparency and fairness.
Jin Chang, CEO and co-founder of Fieldguide, a provider of automation and collaboration software for assurance and advisory firms, said, “The firm leaders and practitioners that I have spoken to are very receptive to AI, which is consistent with research on the industry as a whole. We know from recent studies that virtually every CPA firm out there is looking for ways to close the talent gap with emerging technologies, such as AI and cloud. Of course, they want to do it in a safe and responsible way, but they recognize that the value and leverage AI delivers far outweigh the risks.”
Applications of AI
So how are firms currently using AI? The applications of AI within the profession continue to evolve as technology providers increasingly implement AI into their software solutions, while firms — typically larger ones — build and deploy their own AI-powered tools. AI is fundamentally transforming how business gets done, and generative AI is currently taking the spotlight. Some of the commonly cited applications of AI that are currently in use include:
Internal virtual agents: Leveraging the technology to serve as virtual agents that function as internal resources for a firm’s staff.
Said Asgeirsson, “I think what you’re seeing with a lot of firms is they are first leveraging it for supporting some internal capabilities, internal needs. A lot of larger firms that have different policies and procedures related to how they operate have put in place virtual agents to answer those questions.”
He continued, “One of the powers of generative AI is to have that virtual agent, to put that virtual agent on top of some source of information. I know a number of firms … who have launched generative application internal tools to sit on top of their HR platforms. You can ask the virtual tool how many vacation days you have left, or what are the policies related to a particular matter, or just even simple questions like, ‘How do you get your PC updated?’ — just questions that you would’ve asked via email or picking up the phone and asking somebody.”
Client communication and automation: Saving staff time to focus on higher-level work by assisting with such tedious tasks as drafting memos and other documents, responding to emails, summarizing meeting information, etc.
“Overall, the professionals are very eager to get their hands on some of the AI tools that are out there. What we’ve seen, anecdotally within our own firm, there’s been a lot of interest in products such as [Microsoft 365] Copilot, things that would help them do their jobs. … Helping writing memos, responding to emails, what is in their email box, what should they be responding to, those sorts of things,” said Russ Ahlers, chief information officer for Top 10 Firm BDO USA.
Unveiled in 2023, Microsoft 365 Copilot combines large language models with business data and the suite of Microsoft 365 apps to help organizations improve productivity. Excel, Outlook, Word, PowerPoint, Teams and other apps are all integrated with the AI assistant to help users perform a variety of tasks.
Echoing the sentiment, FieldGuide’s Chang said, “Most firms are using generative AI to help with drafting documents and summarizing information. However, there is a real need and desire to use AI for higher-value tasks, such as creating test plans and evidence requests, automating manual processes, and querying documents — all based on a firm’s methodologies, processes, and specific language models.”
Data analysis and anomaly detection: Analyzing vast amounts of data to identify risk anomalies and possible instances of fraud to improve the efficiency and accuracy of audits.
KPMG’s 2023 “AI in Audit Survey” found that most U.S.-based business decision-makers “expect auditors to use AI and GenAI to challenge their assumptions and to identify what they are overlooking through anomaly detection.”
Said Chang, “AI algorithms can analyze large volumes of data quickly and accurately. This includes extensive data sets that were previously difficult to access across spreadsheets and documents. As a result, AI that is designed for the advisory and audit profession enables practitioners to gain profound insights, easily recognize patterns in the data, pinpoint anomalies, and improve the identification and assessment of risks.”
Added Danielle Supkis Cheek, vice president of strategy and industry relations for financial risk discovery and anomaly detection platform provider MindBridge, “Our technology, when a firm is implementing it, they are usually implementing it as part of their audit process or part of some kind of internal control process. What we do is we use AI to look at full populations of data, like full registers, full ledgers, and tell you how different every single transaction is from each other. … Execution at scale is actually surprisingly complex, and so you need AI to do that scalability piece of it. For example, our product goes up to 500 million records, and we’ve analyzed over 100 billion transactions.”
As it relates to data, Cheek also noted that greater attention is being placed on leveraging AI to turn unstructured data into structured or semi-structured data: “When you get into generative AI concepts, you are going into the world of unstructured data, and there is so much unstructured data that is in the accounting profession. The real value that people are going to start to see, and why they like it so much, is not just for the efficiency plays [i.e., drafting emails, etc.] … but the real transformation is going to start to come when you can take … unstructured data and move it into a place of semi-structured or structured data. So, that is going to start to be the use cases related to summarizing contracts, looking for differences in contracts, or looking for getting a helping hand to find what are the key provisions of contracts. Revenue contracts with customers and leases have become so important to accounting professionals on both sides, industry as well as auditors, that those are going to be profoundly important for understanding your daily operations.”
Tax and client services: Helping clients to make more informed business decisions, as well as conducting fast and accurate tax research to keep pace with legislative and regulatory changes and ensure compliance.
“Every technology provider that we are interacting with in this space is currently incorporating some of the generative AI capabilities into their solutions. In the tax area, there are a number of things that you can do. One, you can use it for basic research. It is the next-generation research tool. The next phase of that is to put a particular [client] scenario into the generative AI tool … and then get recommendations,” Asgeirsson said. “When you talk about serving your clients, there are going to be efficiency gains. There’s going to be better insights because you’ll be able to leverage these virtual assistants to provide those better insights. And then there’s going to be really interesting scenario planning capabilities as you take a particular client situation and you put it into a large language model and you get all kinds of different options that you can discuss with your client. It could be related to how they are driving their overall business from an accounting standpoint or potentially related to their tax areas and their tax services.”
What the pioneers are doing
Among the firms experiencing the benefits of AI-powered capabilities firsthand is BDO. One of the AI initiatives underway is the firm’s in-house generative AI solution.
“The primary functions of what we use it for is we allow a space for people to use the native ChatGPT technology to ask questions. Those questions could be a variety of different things from helping to improve grammar, checking the tone of an email, doing brainstorming, summarizing notes, summarizing tasks. It is really used for a variety of different things,” said senior director of technology and innovation Mike Gerhard. “We were able to build something that keeps our data within the BDO walls while still allowing our users to take advantage of that type of technology.”
Gerhard said the firm has also built in personas, which are essentially custom GPTs designed for a specific task. These personas leverage internal resources like corporate manuals, code of conduct, employee health benefits, and the help desk knowledge base so employees can quickly get answers to direct questions. Questions could include, for example: What is my 401(k) match? Or, what does this error message mean?
While BDO is building generative AI capabilities into internal applications, it is also building the technology into solutions used for client services. What is key is leveraging the technology to augment, not replace, the role of its professionals. “Areas where we are going in the future, in future developments with different applications, whether within tax or audit, specifically in the generative AI area, are focused on automations, efficiencies. For example, also streamlining the usage of data analytics. Another part of that is really trying to personalize the customer experience,” Ahlers said.
Moss Adams is also looking for ways to increasingly implement AI-powered capabilities. For instance, the firm has developed its own private version of ChatGPT and built specific embedding models around it to help facilitate client work (i.e., draft, review or edit documents). However, the real game-changer, said Armstrong, is within Moss Adams’ use cases.
“I find discrete tasks, like I need to convert a trial balance into a financial statement, and then what I do is I highly narrow down the problem space, and I build an embedding that is for that — or fine-tuning, depending on the case, and by the way, that may actually be a pipeline as opposed to a discrete [task] — and I build that into my existing workflow,” Armstrong explained. In short, Moss Adams strives to identify time-consuming tasks and bottlenecks within the firm, and develops automation capabilities to improve efficiencies, often leveraging machine learning or LLMs to achieve its goal.
Meanwhile, solution providers are increasingly implementing AI innovation to help firms drive efficiencies and better serve clients. One such example is FieldGuide, which launched FieldGuide AI. As explained by Chang, “FieldGuide AI enables firms to automate tests and requests based on a firm’s unique methodologies, generate tailored engagement deliverables, and intelligently analyze client documents.”
The newest features include the ability to automate any type of manual task or activity, the ability to query documents, and the ability to develop firm-specific plans. Because FieldGuide AI incorporates a firm’s specific methodologies, policies and language, practitioners can generate consistent test procedures. This can prove especially beneficial for firms that outsource portions of their engagement or have junior resources.
Looking to help firms tear down silos and unify data to improve efficiencies, Aiwyn, a provider of practice automation solutions, has added Relay to its product portfolio. The data ecosystem is designed to connect, synchronize and automate actions across a firm’s entire tech stack, and features smart automation for tasks triggered by specific events.
“It is basically an intelligent orchestrator of the firm, and there’s as many use cases as you can imagine with it. That is something we are really excited about because our vision is that firms in the future are going to want their tech stack to be best of breed. And there’s not going to be a single vendor that is the best of breed at everything, so whatever a firm wants to use, at the very least they should be able to have the data flow and speak to each other,” said Justin Adams, CEO of Aiwyn.
Mindbridge is also bolstering its offerings, and has rolled out a new product series on its MindBridge AI platform with four new products:
- Vendor Invoice Risk Analytics;
- Payroll Risk Analytics;
- Company Card Risk Analytics; and,
- Revenue Risk Analytics.
“One of the things we’ve done is we have converted our general ledger individual product into an entire platform where you can load up additional analysis of different types of data,” said Cheek. “So, we’ve built four additional pre-built products. We have capability to do more, and you’ll see more from us in the future of different types of products. We started with the ones that have the most common data sets.”
Furthermore, CPA.com announced in 2023 that it is delivering a series of tools and resources to help build awareness and greater understanding of generative AI and its potential impact on the accounting profession.
For instance, in January 2024, CPA.com and the AICPA convened subject-matter experts, technologists, practice leaders, and solution providers for the first AI Symposium. And CPA.com also released a downloadable generative AI toolkit that includes use cases and a strategic, multistep plan for constructing a tailored AI strategy.
Said Asgeirsson, “What we were trying to do a lot in 2023 was to demystify [generative AI]. To help frame the discussion. To help firms begin to get their arms around what this capability is and, more importantly, how it can help them evolve and drive their client services forward.”
