- feature
- TECHNOLOGY
How finance can get started with gen AI
Follow these steps to make sure you are on track with this game-changing technology.
Related
‘We’re still the thinkers’ — a reminder for tax pros in the AI era
Summing up economic sentiment and concerns about inflation and tariffs
Business outlook brightens somewhat despite trade, inflation concerns
When paired with talented staff, generative AI can improve efficiencies across a business and open new opportunities to create business growth.
But incorporating generative AI in a way that aligns with controls and governance can be daunting, particularly in small and midsize businesses that have little experience with the technology.
“Most CFOs and finance executives understand that generative AI is something that they need to embrace,” said Paul Parks, CPA, CGMA, director–Management Accounting at AICPA & CIMA. “However, they recognize the importance of adopting the technology in a disciplined manner. There are risks associated with using gen AI, with data privacy and security being at the top of the list.”
HOW TO GET STARTED WITH GEN AI IN FINANCE
CFOs and finance teams must address several concerns to integrate generative AI into the finance function responsibly and unlock the value of generative AI tools, according to the AICPA & CIMA-led Future of Finance Leadership Advisory Group (FFLAG). This group is composed of about 70 senior finance leaders from large corporations.
Robust governance, controls, and processes must address data quality, manage risks, and ensure the ethical deployment of generative AI tools, FFLAG members suggested. Controlled experimentation can help develop the necessary guardrails.
Organizations can begin by taking the following steps:
STEP 1: GET ACQUAINTED WITH GEN AI
Val Orekhov, chief technology officer and partner at accounting and advisory firm Withum, works with business clients of all sizes to incorporate more generative AI into their processes.
AI technologies such as machine learning and predictive analysis have been around for decades. Organizations likely already use such traditional AI systems to aid with data entry and account reconciliation. But the introduction of generative AI now allows people up and down the corporate ladder to use tools that improve their day-to-day efficiency while also unleashing new insights and analysis to inform strategic decisions. CPAs can use generative AI to create presentations and craft staff memorandums, and also for more complex work, such as analyzing years of invoices and purchases to make strategic recommendations.
Traditional AI relies on automated rules and algorithms to analyze and identify patterns within data. It generates results by following pre-determined processes, such as running inquiries into vendor histories.
But generative AI “is different from anything people have experienced before,” Orekhov said. While any form of AI analyzes data to detect patterns, generative AI can create entirely new content by synthesizing patterns learned from large sets of data, rather than being programmed to do specific tasks, as with traditional computer programs.
What follows for businesses are great advantages, along with risks and challenges that need to be understood.
STEP 2: START SMALL WITH GEN AI INTEGRATION
Many finance departments may be reluctant to realize the full spectrum of generative AI because of their instinctive focus on risk.
Those that are using it are just scratching the surface in terms of potential and haven’t explored the full potential of the technology, said Carsten Poulsen, global finance technology lead at professional services company Accenture. “Finance by and large are using traditional AI, like machine learning for their predictive capabilities and their deep quantitative work,” Poulsen said. “They have yet to fundamentally reinvent the way their people do their jobs and to leverage the full potential of AI in general and generative AI in particular.”
Parks suggested starting out by identifying areas where generative AI can add the most value, defining objectives, and getting started.
“Start small with pilot projects,” he said. “Smaller projects provide insights into costs and benefits, making it easier to project ROI on full-scale projects. Knowledge gained through using the technology leads to more ideas and innovation on future projects.”
Examples of where to start could include using generative AI to develop slides for presentations to investors or corporate leadership, from the data analysis to the actual production of the slides. Then as you and your team become more comfortable, you can incorporate generative AI tools to assist in other areas.
STEP 3: ESTABLISH A GEN AI VISION AND ROAD MAP
A road map should outline the steps leading to the adoption and use of generative AI. It should include a strategy aligned with business goals and cover establishing project governance, updating data governance policies, providing employee training programs, creating or updating acceptable use policies, selecting tools and platforms, and developing a generative AI adoption policy.
Adopting generative AI tools goes hand-in-hand with the finance function becoming a strategic adviser, a data storyteller, and a value driver, according to FFLAG. This transformation requires leadership and finance team members who are data-literate and have the expertise to embrace the new technology. It’s important that the finance, information technology, and human resources teams are closely aligned (and allied) as they work with leaders across the business to define use cases and deploy the technology.
STEP 4: PROMOTE A COLLABORATION MINDSET
It’s important that finance executives work closely with their counterparts in technology and operations when launching strategic initiatives around AI, Orekhov said. Upfront collaboration creates buyin across business functions and can generate insight from different stakeholders to maximize the efficiencies that AI, including generative AI, can deliver while ensuring investments align with genuine business needs.
Many large companies have the business, finance, and tech capabilities in-house to scrutinize potential generative AI applications and to ensure they comply with data regulations and achieve strategic goals. Smaller companies, if they don’t have those people on staff, can look to consultants to ensure that they can tap into the efficiencies that generative AI can offer while com- plying with existing regulations and protocols.
“Leverage your ecosystem of vendors and advisers and create a dedicated team in your organization,” Poulsen advised. “In order to scale gen AI, you need the right digital core, a willingness to change your processes and ways of working, reskill and upskill your people, and build new capabilities around responsible AI.”
STEP 5: SET REALISTIC EXPECTATIONS FOR GEN
AI Don’t expect generative AI to work flawlessly, Orekhov cautioned. Doing so can result in writing off generative AI prematurely as an inadequate tool or failing to maintain reviews and audits to ensure the analysis is in line with business objectives.
He strongly suggested verifying results and being on the lookout for possible errors or deviations in the output, just as you do when humans are doing the work. By its very nature, AI learns by finding and averaging patterns across many examples, which means individual outputs may not always be precise or consistent.
“You have to have processes and checks and balances that look at a result and apply professional judgment, regardless of whether produced by human or machine,” Orekhov said.
STEP 6: ADDRESS DATA QUALITY AND QUANTITY
One of the biggest hurdles to adopting generative AI in the finance function is making sure the right data is fed into the application to realize value in business settings, Poulsen said.
“Many [finance executives] are challenged with finding and shaping use cases that drive value at scale” because of continued issues with data, he said.
Generative AI works well when it has vast amounts of data to analyze and learn from. But because many companies don’t have easy access to their data — it may be spread out across many sources in disparate locations or of poor quality with a lack of detail — the generative AI can’t perfect its analysis, especially when the generative AI is working on a closed system that only takes in information provided by a company. Smaller pools of data could lead to biased or inaccurate outputs, such as failing to account for past revenue trends when creating forecasting models.
Poulsen suggests making it a strategic priority to ensure finance data is accessible and usable, so that it can also be deployed to train generative AI products and create more efficiencies in the future.
Incorporating generative AI may seem a daunting task, but taking steps now to set up procedures and protocols for using the advanced technology will go a long way in ensuring success as well as preparing your organization for the near and far future, Parks, Poulsen, and Orekhov all said.
“Don’t sit on the sidelines,” Orekhov said. “The shift is happening now.”
About the author
Sarah Ovaska is a freelance writer based in North Carolina. To comment on this article or to suggest an idea for another article, contact Jeff Drew at Jeff.Drew@aicpa-cima.com.
LEARNING RESOURCES
Generative AI use is growing, and ChatGPT is leading the way. But what exactly is ChatGPT, and how can it help you increase your productivity? This seminar will show how you can use ChatGPT to streamline your workload, save time, and boost efficiency. Earn one CPE credit.
Dates and times: Feb. 13, 1–2:20 p.m. ET; March 13, 1–2:20 p.m. ET; May 19, 1–2:20 p.m. ET
WEBCAST
This is a self-study course on the topics covered in the webcast of the same name (described above). Earn one CPE credit.
CPE SELF-STUDY
Ethics in the World of AI: An Accountant’s Guide to Managing the Risks
This webcast will discuss the current uses and risks of artificial intelligence in business and provide practical suggestions to address these risks effectively. Earn two CPE credits.
Dates and times: Feb. 4, 9–11 a.m. ET; March 4, 9–11 a.m. ET
WEBCAST
Ethics in the World of AI: An Accountant’s Guide to Managing the Risks
This is a self-study course on the topics covered in the webcast of the same name (described above). Earn two CPE credits.
CPE SELF-STUDY
Seasoned CFOs from across the country will gather at the La Quinta (Calif.) Resort & Club to provide the latest innovations and anticipated trends to keep you on the cutting edge.
April 23–25
CONFERENCE
Generative artificial intelligence is on the agenda at the biggest event in the accounting profession, AICPA & CIMA ENGAGE 25, to be held at the ARIA in Las Vegas. Don’t miss it!
June 9–12
CONFERENCE
For more information or to make a purchase, go to aicpa-cima.com/cpe-learning or call 888-777-7077.
AICPA & CIMA MEMBER RESOURCES
Articles
“More Organizations Are Turning to AI in the Finance Function,” JofA, Sept. 16, 2024
“Gen AI Could Save Companies Money, Improve Customer Relations,” FM magazine, Aug. 14, 2024
“4 Considerations for Finance Teams About Gen AI,” FM magazine, April 19, 2024
“Gen AI for Accountants: 10 Prompt-Writing Tips,” JofA, April 1, 2024 Podcast episodes
“How AI Can Drive a Proactive Approach to Corporate Finance,” JofA, May 2, 2024
“How Accountants Can Appropriately Rely on AI,” JofA, April 25, 2024
Reports
