{"id":1893,"date":"2026-03-09T17:41:03","date_gmt":"2026-03-09T08:41:03","guid":{"rendered":"https:\/\/digirise.ai\/?post_type=chaen&p=1893"},"modified":"2026-03-09T17:41:52","modified_gmt":"2026-03-09T08:41:52","slug":"20260220-wework-honmachi-opening","status":"publish","type":"chaen","link":"https:\/\/digirise.ai\/en\/chaen\/20260220-wework-honmachi-opening\/","title":{"rendered":"[WeWork Honmachi Presentation Report] Supporting 400 companies with AI reveals the crucial difference between "growing companies" and "stagnating companies""},"content":{"rendered":"
Talk session at the WeWork Honmachi opening party<\/p>\n<\/figcaption><\/figure>\n

Appeared at the opening of WeWork Osaka Honmachi<\/h2>\n

In February 2026, I spoke at the opening party for WeWork Honmachi (Osaka).
\nThis was WeWork's first new location in the Kansai region in about six years. The event featured a talk session on the theme of AI and a networking event with over 100 business owners and startups.<\/p>\n

I will summarize the content of the talk session on the day in as much detail as possible.<\/p>\n

\n

GMO AI training results for 6,000 people: utilization rate goes from 6% to a shocking 90%<\/h2>\n

In Digirise, in the past year and a littleOver 400 companies<\/strong>We have supported the introduction of AI in<\/p>\n

Among them, the most symbolicGMO Internet Group<\/strong>Examples of:
\nHere are the results of AI training for 6,000 employees.<\/p>\n

    \n
  • AI utilization rate:6% \u2192 Approximately 90%<\/strong><\/li>\n
  • Work hours reduced per person:38.4 hours per month<\/strong><\/li>\n<\/ul>\n

    38.4 hours per month meansAlmost 5 full days' worth<\/strong>This has an effect when multiplied by 6,000 people.<\/p>\n

    The purpose of introducing AI is simple:"Hand over tasks that don't need to be done by humans to AI, and focus on the tasks that humans should be doing."<\/strong>.That's it.<\/p><\/blockquote>\n

    \n

    AI has made work incredibly fast. But why has it made me busier?<\/h2>\n

    This is my own experience, but since I started using AI,On the contrary, I've become busier<\/strong>.<\/p>\n

    Minutes, emails, document preparation - post-meeting administrative work can now be completed almost instantly.Meetings keep popping up on the calendar<\/strong>As efficiency increased, we were able to do more and work more.<\/p>\n

    Another big one isSpeed of research<\/strong>.
    \nWhen a new industry project comes up, we quickly research it using ChatGPT's Deep Research, put it into documents, check it with robots, and then discuss it with executives. This process is incredibly fast.<\/p>\n

    however,Whether or not the speed of decision-making has increased is another matter.<\/strong>.<\/p>\n

    The time from research to discussion has become much faster.
    \nbutIn the end, it's humans who decide<\/strong>This place remains unchanged.
    \nLarge companies in particular have multiple layers of approval processes, and decision-making is slow even with AI.<\/p><\/blockquote>\n

    \n

    What is the difference between "growing companies" and "stagnating companies"?<\/h2>\n
    \"WeWork\u672c\u753a
    Talk session<\/figcaption><\/figure>\n

    Having looked at over 400 companies, there is a clear difference between companies that are well suited to AI and those that are not.<\/p>\n

    Characteristics of growing companies<\/h3>\n
      \n
    • The top management is working hard<\/strong>(Not only by command, but also by touching it himself)<\/li>\n
    • Paid accounts are distributed to all employees<\/strong>(It also helps to combat shadow AI)<\/li>\n
    • The stance is "Let's work together" rather than "Use it"<\/strong><\/li>\n
    • At the field levelSmall success stories<\/strong>are being loaded<\/li>\n<\/ul>\n

      Characteristics of companies that are stalling<\/h3>\n
        \n
      • Just by installing the toolNo operational design<\/strong><\/li>\n
      • The person in charge of promotion is isolated<\/li>\n
      • Management thinks that "AI is scary" and "it's too early"<\/li>\n
      • IT vendors themselves are not using AI<\/strong>(This is quite common)<\/li>\n<\/ul>\n

        The last point in particular was met with a lot of sympathy from the audience.
        \nIf your IT vendor doesn't use AI,The implementation location is unlikely to go well<\/strong>.<\/p>\n

        \n

        In an age where AI writes code, the value of "engineers who can speak" is skyrocketing<\/h2>\n

        This was another topic that got everyone excited at the event.<\/p>\n

        In an age where AI can write code,The relative value of good coders is declining<\/strong>.
        \nInstead, what's going up isPeople who can handle upstream processes<\/strong>value.<\/p>\n

          \n
        • What the client wants to doHearing<\/strong>do<\/li>\n
        • Structuring the problemrequirements definition<\/strong>do<\/li>\n
        • Between the field and managementBridging<\/strong>do<\/li>\n<\/ul>\n

          In short"Engineers who can speak" are the best<\/strong>The era of.<\/p>\n

          AI can write the code.
          \nBut what really gets people to decide "what to make" isOnly people who can speak<\/strong>.
          \nThere is an overwhelming shortage of personnel who can handle upstream processes (interviews and requirements definition).<\/p><\/blockquote>\n

          Incidentally, the hiring criteria for Digirise have also changed."Can you speak?"<\/strong>We place importance on this.<\/p>\n

          \n

          New businesses are entering an era of "making 100 and hoping one hits the mark"<\/h2>\n

          There are also big changes on the business side.<\/p>\n

          Until now, the standard procedure has been to analyze the market and carefully launch a business.
          \nIt's different now. AI can create prototypes quickly,Make 100 and focus on one<\/strong>This approach is becoming more realistic.<\/p>\n

          The speed of the PDCA cycle is crucial<\/strong>An era of division.
          \nCompanies that think too much and fail to act will be overtaken by companies that think while acting.<\/p>\n

          \n

          Things you should never do in the age of AI<\/h2>\n

          Don't just give out AI-like output.<\/h3>\n

          Send the Gantt charts and documents you created with ChatGPT directly to your clients.This is a big no-no<\/strong>.<\/p>\n

          When one of our members submitted consulting materials that looked too much like AI, a client pointed out to him, "You're using AI, aren't you?" It's perfectly fine to use AI.Carefully check your output to make it look handmade.<\/strong>If you neglect this, you will lose trust.<\/p>\n

          Don't just accept AI answers<\/h3>\n

          The AI's answer is 100%, which meansNot likely<\/strong>.
          \nWhat is most needed in today's world isThe ability to judge whether the answer given by AI is correct<\/strong>.<\/p>\n

          Get into the habit of drawing your own conclusions.
          \nBefore asking AI, first have your own opinion<\/strong>.
          \n"When I asked AI, this is what it said" is no longer valid.<\/p><\/blockquote>\n

          \n

          Stories about how businesses are actually being born from the WeWork community<\/h2>\n
          \"WeWork\u672c\u753a
          Commemorative photo with the speakers<\/figcaption><\/figure>\n

          Switching gears a bit, we also talked about the power of the WeWork community.<\/p>\n

          When I started my business in Shanghai as a student, my first office was a WeWork. I believe in the power of place so much that visiting WeWork locations around the world has become a highlight of my business trips.<\/p>\n

          In fact, even in Digirise,Employees interacted naturally within WeWork and won a development project worth 10 million yen<\/strong>It wasn't an intentional sales pitch, but rather trust was born through daily face-to-face contact, which led to business.<\/p>\n