The Inconvenient Truth About Why Generative AI Isn't Taking Root in Companies and How to Deal With It

Doesn't the content discussed in the PIVOT video below perfectly describe what is happening in many companies?

I think the "inconvenient truths" raised here are truly insightful. Furthermore, the solutions provided are also very helpful.

I felt that approaches like "involving the front lines to make it their own" and "creating success stories in small organizations that agree with the purpose and express active commitment, then rolling them out to other organizations" are highly versatile methods applicable to any case of introducing new systems or services, not just AI.

The "Inconvenient Truths" and "Hope" Discovered Through Practicing AI-Driven Management: An Interview with Kenji Sudo, CEO of Kaizen Platform

Kenji Sudo, CEO of Kaizen Platform, explained in a PIVOT video the inconvenient truths and hope that have become visible through the practice of his book, "AI-Driven Management." We will share, along with concrete examples, how AI is changing work and business models, and what is required of us as individuals.

■ The Current State of AI Utilization and Major Challenges

Incorporating AI into business is a key point that has the potential to increase sales and profits without increasing headcount. However, the path is not smooth.

According to Mr. Sudo's actual experience and observations, a problem is occurring in many companies where AI is not being used much within the company. This trend is also seen in organizations with many people who have high tech literacy.

The main reasons for this are as follows:

• It is difficult and troublesome to choose and use appropriate AI tools. There are a vast number of types of AI tools on the market.

• The psychological hurdle is high. You can fall into a vicious cycle where you don't use it, so no ideas are born.

• Since current AI is not general-purpose AGI but rather narrow AI specialized for specific tasks, it is necessary to use them selectively according to the use case. To use an analogy, it is not a mall where everything is available, but rather like a shopping street with many specialty shops like fishmongers and greengrocers. Since the performance of individual shops (AI tools) has improved significantly, productivity will improve more by choosing and creatively using what is available rather than waiting for a mall to be built.

Due to these challenges, the current situation is that even if AI tools are introduced within a company, the actual usage rate has not increased much.

■ Breakthroughs in AI Utilization: The Keys are "UX" and "The Front Lines"

So, what should be done to advance AI utilization? Mr. Sudo and his colleagues say they have found several breakthroughs.

One of the cases where AI usage grew most dramatically was incorporating AI agents into the tools used daily (such as Slack). For example, they introduced a meeting minutes agent into Slack, making it possible to automatically create minutes, respond to questions about the content, and draft emails for the next steps simply by pasting a link to the meeting recording.

This is the result of saving users the trouble of "writing prompts themselves" and enabling them to use AI with the same feeling as a regular work request, just by mentioning it and pasting a URL. In short, a simple and natural experience where AI can be used just by turning a faucet is what is important. From this experience, they realized that UI/UX design is extremely important for the success or failure of AI utilization.

Also, what was important in spreading AI utilization within the company was an approach that involves the people working on the front lines. Through AI hackathons and the like, they solicit ideas from the front lines for tasks that would be made easier by AI, and together they create prototypes and introduce them into operations starting with those that have a high impact. Because the ideas came from the employees themselves, feedback like "I want to do it more like this" becomes active, creating a virtuous cycle where better things are born. A market-in approach tailored to front-line needs, rather than product-out, is the key to success.

Furthermore, approaching from small organizations or specific departments is also effective. By automating the majority of the work in a specific department, they create a success story and then roll it out horizontally to other departments. This is similar to marketing theory where early adopters within a company naturally spread utilization, and it is more effective than forcing everyone to do it uniformly. Ultimately, the goal is not just for specific scenes, but to run the entire workflow with AI.

■ The "Inconvenient Truths" That Have Become Visible in the AI Era

As the adoption of AI progresses, some "inconvenient truths" have also become apparent.

1. Changes in the SaaS business model. Until now, SaaS has provided an interface and a database as a set, charging based on the number of users or usage volume. However, with generative AI, interfaces are becoming diversified (allowing AI to be called from Slack, Teams, LINE, etc.), and as AI begins to perform tasks on behalf of humans, the number of users may decrease. As a result, there is a possibility that the model will shift from per-user billing to a value-based billing model based on sales or profits. This can be described as a shift from SaaS that helps humans process operations to SaaS where AI performs the tasks.

2. Resistance from the front lines. When introducing AI, it is not uncommon for frontline employees to become a force of resistance due to the feeling that "AI will take their jobs."

3. Shifts in who benefits and challenges for junior staff. The ones who benefit from AI first are actually junior staff with little experience. Even someone who knows nothing can suddenly improve their quality by working according to AI's instructions. However, a challenge also arises where it becomes harder to master skills by relying too much on AI.

4. Polarization of talent and the importance of verbalization skills. In the AI era, only a small top tier will remain, and among them, those who can verbalize will be at a significant advantage. Especially in white-collar work, the ability to clearly verbalize one's own thoughts and instructions to AI becomes crucial.

5. The strength of AI for amateurs. Interestingly, there are cases where operations go better when an amateur who knows nothing is paired with AI than with a professional accustomed to existing methods. Professionals may try to correct things in their own way, but amateurs execute AI's instructions without doubt, resulting in shorter lead times and lower costs. This suggests that when automating a task, it is effective to run a completely new AI-driven line in parallel with existing methods to compare them.

■ Hope in the AI era and required capabilities

AI is not an enemy. It is important to have the preparedness to be beaten by AI while considering how to master AI and how to collaborate with it. Beyond just streamlining work, it is possible to use AI to reconstruct the work itself or engage in unprecedented creative activities (e.g., a salesperson who composes music). Here, creativity becomes extremely important.

The following points are cited as human-specific capabilities that will become particularly important in the AI era.

• The ability to ask questions. AI is good at providing answers, but setting the questions—such as what to have AI do in the first place or how to think about it—is a human role. Whether you can ask good questions becomes important.

• The ability to evaluate. You need the skills and expertise to evaluate whether the answers or ideas produced by AI are of good quality and close to what is required.

• Verbalization ability. The ability to clearly put your thoughts and instructions for AI into words is essential for mastering AI. Especially in white-collar work, it is important not just to learn by watching, but to verbalize things so they can be understood.

• Communication skills. Communication between humans is essential, but communication with AI is also becoming important. However, communication with AI has different characteristics, such as not needing to 'read the room' as you would with humans. You need the ability to use these skills appropriately depending on the situation.

• Ambition and curiosity. If you lose these, you risk becoming a slave to AI. The drive to think and learn new things is crucial.

While simple routine work and operational tasks are highly likely to be replaced by AI, creativity, ideation, and coordination skills—jobs that require a certain degree of creative elements—will be handled by humans.

AI may also change how we master skills. It is entirely possible that we will grow in ways different from before through the massive amount of feedback provided by AI.

■ Summary: How to Survive the AI Era

AI has the potential to dramatically improve work efficiency, but it also brings inconvenient truths such as the transformation of SaaS business models, resistance in the workplace, and the polarization of talent. However, rather than viewing AI as an enemy, mastering AI as a weapon and honing human-centric abilities like creativity, the power to ask questions, the ability to evaluate, and above all, verbalization skills, communication skills, and ambition is the key to surviving the AI era. In particular, the cycle of asking questions—such as why we do things and how we think—and engaging in dialogue to evaluate and improve AI's answers is vital. The era has arrived where we leave simple, routine operations to AI and focus human efforts on more creative and essential tasks.

This video was highly thought-provoking, capturing the reality of the changes brought about by AI while providing hints on how to face them and carve out the future.

■ Click here for the digest version.

■ Click here for an explanation of the PIVOT tie-up video's business model.