Even a Small Company Can Make AI Its Weapon

Yamazaki: We're a real-estate company of about 15 people, handling everything from property acquisition to sales. Our long-standing challenge was how dependent on individuals and inefficient our operations were. Back-office work in particular relied heavily on manual effort, and we were forever scrambling to keep up with the day-to-day.

Then this year I had the chance to attend South by Southwest (*1), and I was struck when I heard people from tech companies all over the world saying, in one voice, "No matter how small your company is, you'll fall behind if you don't use AI." The energy felt just like the dawn of the internet, and I became convinced that AI can create a competitive edge regardless of company size.

After returning to Japan, I took on the role of CAIO (Chief AI Officer) myself, and we started by building a culture of AI use within the company. At first we were only using tools like ChatGPT in a piecemeal way—AI adoption was below 1% of our overall work—but under the rallying cry of "Everyone, use AI every day," we drove it home through study sessions and small projects, and now almost all of our employees use it as part of their daily routine. Some even devote a third of their work to collaborating with AI.

Yamazaki: It all started over drinks (laughs). We were talking about AI trends, and in that flow they held a Cursor study session for us during Golden Week. The first time I actually tried it hands-on, problems we could never get past even with GAS and the like were solved all at once. With complex system integrations, fixing one part would cause something else to break—and I was amazed to see Cursor fix that automatically.

The deciding factor in asking for their support was Takumi Nakazono himself. He has an uncanny knack for calibrating the difficulty of each task to your skill and understanding, and when you pull it off, he praises you with everything he's got. The way he keeps you motivated feels great, and it makes you comfortable asking questions. Time and again, things that would have taken hours on my own were solved in just five minutes.

Yamazaki: They gave us broad, hands-on support on things that tie directly into real work—the initial Cursor setup, the design philosophy behind the MVC model (*2), choosing the right language models for the job, configuring voice-input tools, and more. On top of weekly coaching, it was reassuring that I could ask questions by message whenever I needed to. It wasn't just learning how to use the tools—it felt like they installed in us how to craft effective prompts and how to organize the points where people tend to get stuck.

Yamazaki: The biggest result is that we automated everything from scraping to analyzing real-estate data, and we can now collect and evaluate about 50 records every day. From the properties our own criteria flagged as "undervalued," we actually purchased one and generated a profit in the tens of millions of yen. By our calculations, including work we could never tackle before, this amounts to operational efficiency equivalent to about 1.5 full-time roles.

On top of that, for work we used to outsource, we can now assign internal team members whose own roles became more efficient—which ended up cutting costs. The fact that AI could cover the work of a position we'd planned to hire for is another big financial benefit. And beyond that, a positive mindset of "try it first" and "ask for help when you get stuck" has spread across the company, and we're settling into a rhythm of forming AI teams like task forces to deliver results and then rolling those out internally.

Yamazaki: The biggest thing is that even when I'm faced with a new tool, I've come to think, "Let me just try it first." If you try it and it doesn't fit, you can stop; if it works, the work gets better. Not clinging to sunk costs and repeating small improvements is a mindset that applies not just to AI but to every kind of business improvement.