Originally, I worked in advertising, then became a freelancer, consulting for various companies. When ChatGPT came out in 2022, I was truly surprised. Since then, I’ve been absorbed in studying “generative AI.”

I felt that in Japan, conversations about AI were often focused only on topics like “the performance of the newly released ChatGPT 4.1 is awesome,” “how to develop it,” or “how to utilize generative AI in work.” On the other hand, seeing “AI Agents” being born in the US and rapidly used in business settings made me think, “This is amazing!”

Honestly, I think discussions about which AI has the best performance are like a hobby for enthusiasts. Even if a Ferrari engine is amazing in an F1 race, it’s not directly related to daily shopping or commuting, is it? What everyone truly wants is a “practical car (AI Agent)” that helps them finish work faster or enables them to do new things. (Of course, I understand that such discussions can be interesting as entertainment, though).

Furthermore, I believe it’s a big mistake to talk about Japan’s unique AI development or Japan’s winning strategy. After all, 3-4 companies are developing with annual costs in the trillions of yen, and there are services available that also support Japanese. There’s no way Japan can win against that. I feel that the current situation, where Japan is not keeping up with the evolution of AI Agents that link with major global AIs (like Google and OpenAI), could lead to another “Galapagos syndrome,” where Japan is left behind again.

The main reason I started Agentools is that I realized the value that AI Agents bring to business: “fast, cheap, and good.”

“Fast, cheap, and good” means that implementation and getting started are very fast (preparation takes only tens of minutes to a few hours), the price is low with free trials or available from a few thousand to tens of thousands of yen per month, and because they are used by many people worldwide, they are good tools that are genuinely useful for work.

Moreover, AI Agents get power from rapidly evolving AI (LLMs) and become easier to use as more people around the world use them. If you build something from scratch yourself, it’s hard to keep up with this evolution, and it can even become difficult to use. If an AI Agent doesn’t work out after trying it, it’s easy to switch to another. Therefore, instead of suddenly building something yourself, the shortest path to success is to first try out various AI Agents to see what fits your company.

At Agentools, in order to find AI Agents that can achieve this “fast, cheap, and good,” we spend a great deal of time thoroughly investigating and selecting only those that are truly useful for work from the more than 25,000 AI Agents said to exist worldwide, one by one. We honestly evaluate them as “This is good!” or “This is a bit…” and try to explain the basis of our research as clearly as possible.

By doing this, I would be happy if the readers of this article can find the perfect AI Agent for themselves and immediately start making their daily work more efficient and easier.

I want to continue working hard so that the amazing potential of AI Agents becomes commonplace in various work settings in Japan.