“Reversible Management” in the Age of AI: Is It the Employees’ Fault? No, the Problem is “Half-Baked Tools”

“My employees aren’t skilled enough, so work doesn’t move forward.” I often hear this concern. But is that truly the case? Perhaps the problem lies not with the “people” but with the “tools.” This time, I’ll introduce a concrete first step for “reversible business process reform” utilizing AI.

Hello. This is Hodaka Goto.
We engage in daily conversations with SME leaders on the theme of “Reversible Management.” Recently, I’ve been feeling something particularly strongly. It’s that advances in AI are making “reversible decisions” possible.

In the past, implementing a major system was an “irreversible decision.” It required significant investment and a long time. But now it’s different. By leveraging AI, you can start small and make immediate corrections. This is the strength of modern “Reversible Management.”

1. The Leader’s Honest Feeling: Is It the “Unskilled Employees”?

The other day, a business owner consulted with me, saying:
“Projects aren’t moving forward at all because of unskilled employees.”

Upon hearing the details, the situation was as follows.

• Sales data is managed in Excel.
• Customer information is in a separate cloud tool.
• Progress reports come in sporadically via a chat tool.
• “People” are manually consolidating and analyzing all this.

In other words, it’s a state where several half-baked tools are used across the board, with several half-involved people engaged. It’s no wonder the work becomes complicated. I realized the problem wasn’t “employee capability” but “the very design of the work process.”

2. A Shift in Thinking: Use AI to Create a “Custom Tool for Your Company”

The traditional solution was this:
“Introduce a high-functionality integrated tool (expensive) and train everyone on it.”

However, this takes both time and money. Moreover, it tends to become an “irreversible” decision. What we proposed was the complete opposite approach.

“Let’s use AI to build a tool perfectly suited to our company from scratch.”

Specifically, the steps are as follows.

1. Write out all current workflows.
   (Who is doing what, using which tool)
2. Extract entire sections that “could be handed over to AI.”
   (Data transcription, aggregation, simple analysis, report drafting, etc.)
3. Use no-code tools or AI APIs to create a minimal tool (prototype).
   (It’s okay to start with just one task.)
4. Try using it, and if it doesn’t work, fix it immediately.
   (This is the core of being “reversible.”)

The important thing is to incorporate AI into the tool as an “actively working mechanism.” Not just a database, but a design where the AI proactively suggests, “This data has decreased by 20% compared to last month. Shall we analyze the cause?”

3. Identifying Tasks “Done by People” That Can Be Handed to AI

So, what specific tasks should you delegate to AI?
Leaders, first try categorizing your company’s tasks into these three types.

① Tasks that absolutely must be done by people (Judgment, Creation, Empathy)
Examples: Management decisions, critical negotiations with customers, brainstorming new product ideas.

② Tasks currently done by people but can be delegated to AI (Routine, Repetitive, Aggregation)
Examples: Input and transcription of various data, formatting reports, creating regular reports, drafting template responses to simple inquiries.

③ Tasks where collaboration between AI and people multiplies effectiveness (Analysis, Preparation, Expansion)
Examples: Analyzing market data and extracting trends (AI prepares → people judge), drafting proposal outlines (AI creates the skeleton → people add substance).

In many SMEs, there are cases where the more skilled employees are the ones whose time is consumed by type ② tasks. By shifting these to AI, people can focus on ① and ③. As a result, both productivity and work quality improve.

4. A Concrete Approach to “Reversible” AI Implementation

There’s no need to AI-fy all processes at once. In fact, that’s risky. From the perspective of “Reversible Management,” I recommend the following flow.

Step 1: Decide on a Pilot Project (2 weeks)
Start with one task in one department, not company-wide. The goal is “learning,” not “perfection.”

Step 2: Visualize the Current Process (1 week)
Write out the “As-Is” of that task in detail. You will inevitably find wasteful steps.

Step 3: Create a Prototype of the AI Tool (1-2 weeks)
Create a trial version using no-code tools (offered by many cloud services). Keep initial investment minimal.

Step 4: Test Small and Fix Immediately (Ongoing)
Have people actually use it and immediately improve based on “this is inconvenient” feedback. If it doesn’t work, decisively change direction.

Step 5: Expand Successful Patterns (Ongoing)
If it works well, expand it to similar tasks. This repetition leads to sustainable AI integration.

With this method, there are no major failures. You can advance while always maintaining a “reversible” state.

5. What Leaders Should Do and Should Stop Doing

Finally, let’s clarify the leader’s role in this period of transformation.

[What You Should Do]

• Cultivate the habit of thinking “it’s the tool’s fault.”
  When work stalls, review the process and tools before suspecting the people.
• Permit small experiments and praise failures.
  Great learning comes from “reversible” small failures.
• Do a “preview” yourself of what AI can do.
  Leaders themselves should try AI tools a bit to personally feel the potential.

[What You Should Stop Doing]

• Continuously searching for an “all-powerful AI tool.”
  The perfect tool for your company doesn’t exist off-the-shelf. It’s something you cultivate yourselves.
• Declaring a company-wide “Great AI Transformation Campaign.”
  That exhausts the frontline and creates immense “irreversible” pressure.
• Expecting “immediate labor cost reduction” from AI introduction.
  The goal is “to free up human capability for higher-value work.” Reduction is a result, not the objective.

The problem of “unskilled employees” is often actually created by “work process designs that exhaust skilled employees.” AI becomes the best leverage to break this vicious cycle.

Start with one step. With the relaxed mindset that you can always go back even if you fail, why not begin creating your company’s own AI tool?