[AI Error Note] Ford's Accounting Department Fails to Lay Off 500 Employees

④ "AI, Just Implementing It Won't Solve Everything"
Ford Tried to Boost Productivity with PCs but Failed
Later Innovated Workflow through 'Reengineering'
AI's Effectiveness Is Limited with Simple Task Automation
"Requires Fundamental System and Process Redesign"

Executives at the American automobile company Ford were shocked while reviewing materials about the Japanese automobile company Mazda. Mazda's accounting team, including all accounts payable processing and major accounting tasks, was handled by just five people.

How was Ford's accounting department? It employed as many as 500 people.

Image generated by ChatGPT

In the 1980s, the American automobile industry was facing severe management difficulties due to fierce challenges from Japanese car manufacturers. Not only were they lagging in technology, but the inefficiencies in bureaucratic organizational workflows were beyond imagination.

Inventory management, including parts stock, was a mess. The stock of necessary and unnecessary items fluctuated wildly. Supplier management was also unstructured. In terms of quality control, they lagged behind their Japanese competitors.

The problems in the accounting department were particularly severe. Processing a single payment required going through 14 complicated steps. At each step, document handling and verification tasks were repeated. None of the steps were processed quickly. Bottlenecks occurred everywhere. Document loss and errors were frequent.

Ford was greatly stimulated by Mazda's unbelievable efficiency. They pinned their hopes on cutting-edge technology of the time: computers.

Ford's New Technology to Surpass Japanese Cars: Computers

On April 2, 1987, an IBM marketing executive is announcing IBM's new PC at the IBM building in New York, USA. AP

At that time, computers were considered revolutionary new technology, and many companies besides Ford rushed to computerize their offices. However, as seen earlier (in the 3rd installment), most companies did not experience the dramatic productivity improvements they expected.

Ford was no exception. Ford's accounting department was still entangled in a complex and inefficient process of manually matching purchase orders, delivery confirmations, and invoices. Lack of communication between departments caused delays and high error rates, showing no signs of improvement.

Despite huge investments in new equipment, the effects were minimal. The large computers filling the offices were treated as "money-eating hippos."

"Don't Automate. Obliterate!": The Reengineering Technique

Amid this, a management innovation technique called ‘Reengineering’ emerged.

James A. Champy and Michael Hammer published an ambitious paper in the 1990 Harvard Business Review titled ‘Reengineering Work: Don't Automate, Obliterate.’ In the paper, they argued that "the power of modern information technology (IT) should be used to fundamentally redesign business processes and dramatically improve performance."

The answer to the question "Then how should we redesign?" is explained in the subtitle. It is not simply to "automate (Don’t Automate), but to obliterate (Obliterate)" the processes.

Champy and Hammer's core argument was that no matter how much modern IT is applied without eliminating the complex work processes existing within a company, it will not be effective. At that time, most companies regarded computers and IT systems only as tools for "automating tasks."

However, in Champy and Hammer's view, the true power of these technologies was not that. It was not merely about making old processes more efficient, but about breaking the old organizational rules and creating new ways.

In this respect, reengineering differed from restructuring in its approach. While restructuring focused on partial reorganization within existing organizational structures, reengineering meant redesigning business processes from scratch.

Restructuring often aimed at staff reductions and financial cost savings themselves. In contrast, in reengineering, staff reductions and cost savings appeared as results of business process improvements. If unnecessary processes were found during reengineering, they would naturally be cut, but conversely, if new processes were deemed necessary, staff numbers could even increase. Reengineering is not about reducing staff but about improving service quality.

Ford applied this reengineering approach to its accounting department. First, they clearly defined the purpose and goals of the work. Then, they analyzed the processes, eliminated unnecessary steps, and redesigned them to focus on core tasks. The previously complicated 14-step payment process was simplified to just 3 steps.

They built an online database to automate document processing and introduced an electronic payment system with suppliers. They established an electronic data interchange system between the purchasing department and suppliers to automatically process purchase orders, delivery confirmations, and invoices.

What was the result? Ford was able to reduce the accounting department's workforce by as much as 75%. Despite the reduction in staff, the accuracy and efficiency of work processing greatly improved. This was an achievement gained through fundamental redesign of business processes, not simple automation.

The Impact of AI on Business Processes: Changing Importance by Stage

On the morning of the 16th of last month, a participant is looking at the seating chart at an academy in Seoul. Photo by Yonhap News

Let's look at a more recent example. University regular admissions applications will begin at the end of this month.

The core business processes of universities during the admissions season can be broadly divided into three: university promotion, selecting successful applicants, and retaining admitted students (who might go to other universities). Suppose a university uses AI. It quickly identifies outstanding talents based on various data such as applicants' undergraduate grades, experience, and recommendation letters.

Traditional admissions screening was a high-cost process requiring many personnel and time. Therefore, indiscriminately accepting many applications was not the best strategy. There was a high risk of missing truly important talents in the pile of documents. However, if AI conducts the first screening, the story changes.

Being able to select successful applicants faster and more accurately means the cost (personnel, time, risk, etc.) of the ‘selecting successful applicants’ process is reduced. If this process no longer incurs high costs, it becomes possible to accept more applications in the future. There is no need to worry about whether an applicant has a real chance of admission, is a false positive, or is likely to choose another university even if admitted. The goal becomes to collect as many applications as possible.

Then this stage turns into an opportunity to increase application revenue. The nature of the process itself changes. More investment can be made in marketing costs going forward. The function and importance of the ‘university promotion’ stage completely change.

Also, since successful applicants can be selected quickly, the university has more options compared to others. Furthermore, if it becomes possible to accurately predict which admitted students will stay, the importance of appropriately calculating scholarships and various incentives increases. New personnel may be introduced in this process, or new organizations may be created. Saved time and resources can be reinvested in the areas where they are most needed.

Thus, with the introduction of AI, the existing process of 'university promotion → selecting successful applicants → retention' changes the tasks to be done at each stage, i.e., the goals change. The importance of each stage also changes.

The AI Revolution Brings Changes to Existing Organizations and Structures

Getty Images Bank

AI can simplify business processes or even eliminate unnecessary ones. In the process, it can create new tasks and roles with newly emerging importance.

CEOs or organizational leaders aiming to introduce AI into their work should keep in mind Ford's failure and the case of reengineering. Introducing the new technology of AI is not simply about inserting AI into existing tasks.

It requires more than just strengthening the IT department or introducing new systems. It means reviewing organizational goals and fundamentally redesigning business processes.

"The greatest danger in times of turbulence is not turbulence itself, but to act with yesterday’s logic."

This is a quote from the legend of management, Peter Drucker. It is an insight to keep in mind even in the AI era.

Next Series Preview

⑤Why Do We Waste Time at Airports? (December 7)
⑥"Let's Make Something with That AI Too," Says the Executive Director (December 8)
⑦The Seeds of the AI Revolution Sown by Steve Jobs (December 16)

Editor's NoteIt has become an atmosphere where it seems impossible to do anything without artificial intelligence (AI). Almost every product and service promotion includes AI. Office workers are also struggling. Because of easy demands at work like, "Hey, let's make something with AI too." Does having AI guarantee success? Not at all. In fact, the world of AI has been a series of countless failures. Companies that have hit the jackpot with AI are only 0.1% of cases. 99.9% of companies and services have failed. However, one of the fastest paths to success is studying failures. The same applies to AI. Examining 99.9% of failures is the shortcut to 0.1% of success. The 'AI Error Note' series explores failure cases related to AI products, services, companies, and individuals.