After Doctors and Architects, Accountants? "Why AI Can't Replace Accountants" [K-INVESTORS]

Self-taught in Python, evolving into a 'data investigator'
"Data analysis: not an option, but a necessity"

"An accountant is not just someone who verifies the accuracy of numbers, but someone who reads the stories hidden within them. This is an area that artificial intelligence (AI) cannot replicate."

Amid the current surge of AI, waves of change are sweeping through accounting firms as well. AI is now widely used for simple, repetitive tasks such as journal entry input and financial data analysis. The Bank of Korea has already identified accountants as one of the professions most likely to be replaced by AI, following doctors and architects. However, Gyubin Koo, an accountant at Deloitte Anjin LLC who plans and develops AI solutions, offers a counterargument to this outlook.

In a recent interview with Asia Economy, Koo emphasized, "Accountants are not a profession that will be replaced by AI," and added, "On the contrary, accountants will be able to use AI as a tool to focus on more advanced tasks such as communication and complex case analysis, thereby enhancing their expertise."

Gyubin Koo, Senior Manager at Deloitte Anjin LLC, is being interviewed by Asia Economy. Photo by Dongjoo Yoon

Self-taught in Python, evolving into a 'data investigator'

Koo passed the Certified Public Accountant (CPA) exam in 2012 and spent about ten years performing traditional accounting audits and financial statement reviews before expanding his career into data-driven AI solution development. At the end of 2019, he joined the 'AI Asset & Analytics Group' within Deloitte Anjin's audit division, where he develops practical, data-based tools that can be applied to internal controls, financial analysis, and audit automation.

Koo recalled, "After spending a long time handling repetitive tasks in the audit division, I wanted to change the way accountants work by leveraging data analysis technology. I believed that by creating tools that combine accounting knowledge and data analysis skills, we could enhance both the accuracy and efficiency of accountants' work."

Having majored in business administration at university with no prior IT knowledge, Koo taught himself programming skills such as Python and SQL to expand his career. Initially, he struggled to run even a single line of code due to unfamiliar syntax, but he gradually built his capabilities by understanding the structure from small units and applying them to real work. Currently, half of the members of the AI Asset & Analytics Group, including Koo, are from non-IT backgrounds. He said, "If I had thought of it as simply learning a language, it would have felt vague, but by approaching it as learning a tool to do my job better, I was able to pick it up much faster."

Based on this, he developed the 'International Standard Computer Language (XBRL) Disclosure Automatic Analysis Tool.' By simply uploading electronic disclosure files, this tool automatically generates XBRL disclosure Excel templates, dramatically improving the traditional manual data entry and verification process and significantly boosting work efficiency.

Recently, he also participated in developing the 'Lighthouse' solution, which detects signs of fund accidents and embezzlement within companies. The solution can detect risk signals based on inconsistencies between purchases, sales, and account flows, or repetitive journal entry patterns, and is a practical tool that internal audit and accounting teams can immediately use. Many companies are now considering its adoption. Koo remarked, "Through data analysis, I became convinced that accountants can go beyond simple consistency checks to interpret the structure of incidents and uncover the truth, taking on the role of a 'data investigator.'"

Gu Kyobin, Senior Manager at Deloitte Anjin Accounting Corporation, is being interviewed by Asia Economy. Photo by Dongjoo Yoon

"Data analysis: not an option, but a necessity"

Koo predicts that data analysis capabilities will soon become essential for accountants. As corporate financial information grows ever more vast and complex, the ability to structure and interpret data is necessary to quickly extract meaningful insights. He explained, "In the AI era, domain knowledge and problem-design capabilities are more important than technical implementation skills. While the scope of what AI can do has expanded, deciding what to build and what standards to use for judgment ultimately remains a human decision. An accountant's analytical ability is a critical asset that makes those judgments more accurate."

Koo advised junior accountants to start by adopting an attitude of wanting to understand technology. He said, "There is no need to develop AI yourself. Begin by learning the syntax of simple automation tools, and gain experience by applying tools such as Excel macros, SQL, and Python to your work." He added, "It is most effective to first become a domain expert by understanding the structure and logic of your field, and then layer technology on top of that foundation."

He now aims to go beyond accounting data to connect unstructured data and build a platform capable of comprehensive risk diagnosis and prediction. Koo stated, "Whereas traditional audits have focused on reviewing the past, now is the time for a predictive audit platform designed for 'future response.' I want to lead this trend, and especially play a role in helping companies that lack data infrastructure."

Gu Kyobin, Senior Manager at Deloitte Anjin Accounting Corporation, is being interviewed by Asia Economy. Photo by Dongjoo Yoon