[Moneyverse] A Useful Hiring Question in the AI Era

AI Is Still Struggling to Find Omissions
Duplicates Stand Out, but Omissions Don't
The Key to Getting Work Done Without a Hitch: "Leaving Nothing Out"

"People say you can feel at ease when you assign him a task. When he plans and prepares work, he takes care of every related detail, leaving nothing out, and gets the job done without a hitch. That's why his supervisors feel at ease."

I once heard this assessment of the work skills of an acquaintance who had become a senior public official. Work skills can be divided into creative planning and execution, coordination, and driving initiatives forward. Among these, the ability to think of everything without overlooking anything is especially important during the planning and execution stages.

“Leaving nothing out” does come up in training on work skills. The consulting firm McKinsey includes it in the concept of “MECE” (Mutually Exclusive, Collectively Exhaustive). Literally, this means “mutually exclusive and collectively exhaustive.” Even the translated term is difficult. I suspect that people at consulting firms came up with such a complicated name because they need to showcase their intellectual assets. My guess is they wanted to flaunt their “Isseobiliti” with a new buzzword. Put simply, it means “no overlap and no omissions.”

McKinsey recommends using MECE to analyze complex business problems and develop solutions. It explains that when breaking a key question down into subproblems, they should be structured so that they do not overlap and collectively cover the entire problem without leaving anything out. It also advises applying MECE when preparing communications materials such as reports, memos, and presentations to reduce logical overlap and omissions. MECE is mentioned in many self-help books.

A high-quality report meets the MECE standard. Which is more problematic: overlap or omission? Even if a report contains some repetition, the work can proceed without incident as long as nothing is missing. But work cannot go as planned if the report leaves out important information. As the example of the public official's meticulous work showed earlier, carefully ensuring that nothing is omitted is essential to getting work done smoothly.

Four cases plotted based on overlap and omission. Quadrant 1 has neither overlap nor omission, indicating a high level of completeness. Quadrant 3 has both overlap and omission, making it the worst case. Comparing Quadrant 2, which has overlap, with Quadrant 4, which has omission, Quadrant 4 should be avoided more.

Removing duplication from a document is relatively easy. If you look closely, duplicated sections stand out. By contrast, finding omissions is extremely difficult. You need to extend your attention to areas not covered in the document, but readers tend to focus only on what is written there. The five Ws and one H were devised as a way to convey information without omissions, but there are many types of important omissions beyond those six categories. In particular, most are “atypical” and do not fall into any set category.

It is even more challenging to create a “no omissions” question: provide a passage and ask readers to find what is missing. Perhaps that is why such questions are rarely covered in workplace training, such as report-writing courses.

We live in an era when artificial intelligence (AI) can readily investigate situations and organize the findings into documents. Delegating work to AI has become a standard part of the working environment. In response, companies are conducting in-depth interviews with applicants to select talent, assessing their ability to use AI proficiently and verify and evaluate its output. I propose adding “find the omissions” questions to recruitment tests. While preparing and delivering a writing course on using AI, I learned that AI is still not good at finding omissions. “Finding omissions,” a task that cannot be outsourced, is also an essential subject for building the capabilities of existing employees.

“Baekseolibulyeoilrye.” It means that a hundred explanations are no match for one example; I coined the phrase. Let us identify the key information missing from the following passage. It appeared in an exhibition at a public museum. The aim was to convey information, but the omission meant it failed to do so.

"Charles Babbage, Ahead of His Time (title). Charles Babbage is considered the first person to develop a mechanical computer. His “Analytical Engine” was designed to process all possible calculations automatically, and had a structure and functions very similar to those of today's digital computers. Following the emergence of ENIAC after World War II, the era of digital computers began in earnest."

Previously, AI failed to identify the key omission. Curious about what answer it would give this time, I asked again. AI replied, “The most important information missing from this passage is whether the Analytical Engine designed by Babbage was actually built and completed.” It went on to explain, “Readers could interpret this account as saying that Babbage built a working computer,” and supported its answer by noting that “Babbage's Analytical Engine was designed but never completed during his lifetime.” As a second omission, it pointed out that “there was a gap of about 100 years between Babbage's design and the emergence of ENIAC, but the passage says nothing about how computer technology developed in the meantime.”

I asked AI when Babbage came up with the Analytical Engine. It answered, “1837.” I said, “I think this is a key omission,” and AI agreed with me.

“Leaving nothing out” is even more important for today's office workers. If you have drafted a report, I recommend asking AI not only to proofread it but also to check whether anything is missing. Sometimes, AI can identify key omissions.

Baek Woojin, Economic Columnist

This content was produced with the assistance of AI translation services.