The Institute of Management Accountants is examining the possibilities of artificial intelligence in the accounting profession and how it will affect finance jobs now and in the future, as the organization itself recently went through a second round of staff cutbacks.
The IMA did not disclose the number of people laid off in February. The organization had an earlier round of reductions in force about two years ago.
"IMA recently implemented a strategic restructuring, which did impact headcount," said a spokesperson. "Our focus is on positioning IMA for the future — aligned with the needs of our global members. IMA remains committed to our collective growth, and continues to invest in opportunities to advance our organization and profession."
The IMA released a
"Generally speaking, when people talk about AI, it tends to be very theoretical and high level, and what we have found is our members —those that are working in businesses and working with day-to-day processes and procedures and people — really want to understand what's the practical implication of this new technology on the work that they're doing," said IMA president and CEO Mike DePrisco.
For the report, the IMA talked to about 40 finance leaders from around the globe to understand from their perspective, the main challenges, concerns and opportunities related to leveraging AI and emerging technology into finance and accounting.
"We did a number of focus groups with this group of leaders, and they represent every region of the world," said DePrisco. "A number of challenges surfaced that were really categorized around four areas: the human aspect, the technology data aspect, operational aspects and ethical and governance aspects."
One of the worries about AI is the potential for layoffs. "I do think that is probably the biggest concern that many practitioners and organization leaders have as it relates to AI, and that is job displacement," said DePrisco. "That's another reason why stakeholders are sometimes hesitant to adopt AI technology in the workplace because of that. Everything that we see and hear suggests that AI will augment and not replace accounting and finance professionals, but the role of what people will do is different in the future than it is today."
The most cited concern among 38% of the respondents to the IMA survey was the human aspect of working with AI. "The human aspect really is about getting the attention and support from top leadership to invest in and implement AI is a key challenge and a key opportunity for organizations," said DePrisco. "Those organizations that have full support from leadership — those individuals that control the funding and the allocation of resources to certain projects — those organizations that have that support and alignment have a better chance of getting AI projects implemented successfully. The lack of that support, buy-in and alignment from top leadership was cited as a concern."
Another concern relates to the skill gaps of individual employees who are required to work with AI. "Many individuals in accounting and finance may not have had exposure to this type of technology, and the challenge therefore in implementing these projects is how do you help upskill finance and accounting professionals and practitioners?" said DePrisco. "How do you give them the tools, skills and knowledge they need to work with the technology individuals and data scientists in the organization, so they are leveraging and building these algorithms, that they're being built on practical applications or outcomes that the business needs to achieve."
There's also a challenge around stakeholder buy-in, with employees accepting the idea that AI and machine learning are going to add value to the organization and not take away control or displace jobs.
"Getting that buy-in is a critical challenge and an opportunity," said DePrisco.
There are also operational challenges with implementing AI, including cross-functional collaboration. "Implementing AI projects in an organization requires your finance and accounting business people working with your data people and your IT people to ensure that the data going into the machines represents the practical real-world scenarios that accounting and finance individuals are facing and what they need help in, so that when the machine spits out the information and data, it's useful, reliable and suitable for the needs of the business," said DePrisco. "Resource management is always a challenge and concern. Do we have enough resources to help ensure that this project is successful? It can't be something that is just added to someone's plate as another thing that they need to do and manage. AI projects are pretty complex projects. They're time-consuming projects. Create space for your team to dedicate time to a successful implementation."
Organizations may need to reengineer their processes to get good use out of AI. "If your processes are not good, layering in AI on top of bad processes is not going to get you a successful outcome," said DePrisco. "The first step in implementing any AI project is to look at your processes, and to re-engineer processes in a way that's going to be added value once you begin to implement the AI technology on top of it. Making sure that you're rooting out bad processes, reengineering those processes, and taking the time at that point to do it is really the best practice as it relates to that."
Choosing the right AI technology can also be a challenge. "It takes a lot of investment to bring in AI technology," said DePrisco. "You have to look at what kind of technical depth you have. What's needed from an integration perspective before you start making purchases, and starting to think about how you implement AI on top of that?"
Data integrity and maturity are important considerations as well. "Many organizations have data siloed throughout the organization," said DePrisco. "It's structured data and unstructured data. How are you bringing all that together and integrating that data and making sure that it's reliable, clean and trustworthy, so that it can be leveraged and used to develop algorithms?"
Another challenge uncovered by the research centered around ethical and governance concerns. "These concerns are what you hear most about in mainstream media, the importance of data security," said DePrisco. "How does AI technology impact an organization's ability to maintain data security and data privacy? How are you governing the AI in your organization? Many organizations that implement these types of projects need to set up an AI Center of Excellence, for example, to ensure that people throughout the organization have visibility into how the AI is being used. What business outcome are you driving toward? What is the cost of implementation and maintenance? And data integrity. Is the data free of bias? Is it reflective of the business problems that you're trying to solve?"
To help accounting and finance professionals adjust to the far-reaching changes emerging from AI, the IMA is planning to provide more training. "We need to ensure that we're providing education, knowledge and certification training for practitioners who are moving to new roles," said DePrisco. "These can be roles like compliance analysts, individuals that utilize AI to ensure the finance operations are adhering to laws and regulations. There are probably going to be new roles in risk assessment and management, that merge financial expertise with AI proficiency, for example, roles that identify bias in data and mitigating that bias."
He noted that the IMA has long said that accounting and finance professionals are strategic business partners. "The more work is automated, the more opportunities individuals have to step away from some of those manual routine administrative types of tasks that accountants have done over the last 100 years and into that strategic business partner role," said DePrisco. "That's so critically important these days to help organizations achieve their outcomes."
Many accountants are not sure whether it's a good idea to trust AI systems yet with their clients' data since programs like ChatGPT have a reputation for "hallucinating" or making up plausible-sounding information that turns out to be partly or wholly fictitious.
"You need knowledgeable accounting and finance people to question the data that comes out of the machines to ensure that it reflects the real-life scenarios that happen day to day and that reflect data that's correct, accurate and with integrity." said DePrisco. "That becomes an important role of accounting and finance people. That's on the back end, but you also need that capability on the front end. And that's why when I talk about the collaboration, you need experienced, qualified accounting and finance professionals to work with data scientists to build the algorithms that are being used to automate processes and automate a number of these financial processes that are going to create financial statements and other things that the organization is going to rely on. Making sure that the data that's going in there is accurate, free from bias, and represents both unstructured and structured data that may exist in the organization. It's the job of the accounting and finance professional to ensure that those algorithms are being built with the proper data. That's how you mitigate the risk around hallucinations or information coming out that's half baked."
AI can be used for tasks like data analytics, to spot patterns and red flags, but it still requires the professional skepticism that an accountant can bring.
"The machines are proving to be very powerful technology that is creating new value, improving efficiency and productivity overall," said DePrisco. "Like any new technology, there needs to be a healthy dose of skepticism and rigor applied to ensure that we're not just relying on what a machine spits out, that we're actually applying critical thinking, bringing our experience, judgment and curiosity to any data that becomes available through a machine. We've seen this throughout the years as new technology is adopted. There's a maturity curve, and we're still in the early stages of that maturity curve with AI. There will be a lot of learning that happens over time."
