Technology in Accounting (Part 2): Generative AI and the Future

There is a lot of press about generative artificial intelligence (GenAI) and a subset called large language models (LLM). As someone who makes a living in public accounting, I find this topic very interesting. I lived through the changes that paperless audits, mobile devices and the internet brought to the profession. My first audit room had one phone we would unplug when someone needed to access email by dialing onto the internet. My middle schooler’s mobile phone is faster than the computer I used back then.

Our profession has been dealing with new technology for awhile. How will the adoption of GenAI and LLM be different? Let’s first compare the changes coming from GenAI to previous changes. Then, we can play the part of “futurist” in envisioning how those differences might play out.

Changing Technology in the Accounting Industry

The two most significant differences I see are the pace and degree of change. The pace of change seems to be much faster in the world in general, particularly in this emerging technology. The change we experienced living through various versions of Windows, faster processors, faster internet, etc., took years to fully permeate the profession. It felt like a gradual change in many aspects of what we do. Our ability to spread work across a larger geographic area increased when messaging applications (i.e., Skype then Teams) brought the Earth’s four corners together. That said, fully remote work didn’t happen until a pandemic forced the issue when Zoom became an almost overnight sensation. Getting to 100 million users took the telephone about 75 years. Smartphones took about 15. ChatGPT reached 100 million users in two months, faster than any other application in history.

In addition to the pace of adoption, the degree of change post-adoption seems much more acute. Instead of a tool that helps a seasoned professional complete a task more quickly (i.e., Excel to recreate a loan amortization vs. hand-drawn schedules), we have a tool that can completely replace that seasoned professional. A friend told me recently he asked ChapGPT to write a business apology email that was sensitive to two very different cultures, and he barely had to change any wording after reviewing the email with his internal customer service and legal teams. In an article by Ethan Molick, a noted voice in the field of LLM and its impact on the world, he made an analogy between GenAI and a knife. To paraphrase, you can choose to give your organization a “dull knife” (GenAI with less power but more guardrails) or a “sharp knife” (GenAI with full power but few checks). The implication is that this tool can significantly alter the future for the individual user and the organization. There is real danger involved in choosing to use (or not to use) this technology.

Implementing New Technology in Accounting

What does it mean if this technology is more disruptive than previous technologies? The implication is that there will be winners and losers. The obvious losers will be those that have no guardrails or strategy (no “rules of the road” for employees, no incentives for entrepreneurial thinking, etc.) and those who choose to act as if there is no barbarian at the gate (i.e., blanket bans on the use of GenAI or LLM). The winners will be in between those two extremes. They  will quickly assess the landscape and pick a path with GenAI that best matches their organization’s needs. Those on the front line must know clearly where no chance can be taken (i.e., safeguarding the most precious asset of every accounting firm, public perception) vs. where there are no sacred cows (i.e., just because a senior partner says that’s the way it’s always been done, doesn’t mean it is the best way).

The implications of managing a firm during this period of disruption are endless. The first step will be to know what questions to ask. Some questions will be generic for all organizations. For example, is the goal to replace humans at performing an organizational task or make humans even better at performing that task? Some questions will be specific to public accounting. For example, how do you deal with the “black box” of LLM answers when complying with auditing standards? Another imperative is an organization’s realization (and even celebration) that people will make mistakes. Our tendency to lean towards perfectionism will require new, experimental and exponential mindset changes with technology. Finally, if you follow private equity investment in public accounting and you read the annual reports, there is a shift in our profession away from “compliance” work to “advisory” work. The types of people who are value drivers in our firms have changed, and GenAI will only serve to compound this shift. Instead of the individual with the most significant knowledge on a topic, the most valuable person may be the person who can write the best prompt balanced with the ability to spot AI hallucinations and communicate the results to a consumer for whom the answer matters.

Where Do We Go from Here

I once heard a story that modern ATMs were developed when someone asked a bank branch manager what he wished for. He said that he wished customers could perform simple teller services (deposits and withdrawals) outside the bank so he could focus on his customers who really needed his help inside the bank. I assume he enjoyed this work more, and the bank made more money doing this work. I believe now is another time in history when we should ask ourselves, “What do we wish for?”

If you have any questions about the use of technology or Generative AI in your organization, our Assurance Team can help. Contact us for more information.