According to a survey by Tokyo Shoko Research Yamagata, only 16% of companies in the prefecture stated they are “actively promoting” the use of generative AI. This figure is not unique to Yamagata. It reflects the reality that many companies headquartered in regional areas are hesitant to adopt generative AI.
On the other hand, the same survey shows a certain number of companies are “considering adoption.” In other words, there is a significant group that is interested but has not yet taken concrete action. This article introduces specific methods to break through this “16% barrier.”
Why Does Adoption Stall at 16%?
There are three main reasons why generative AI adoption is not progressing.
First, there is a vague anxiety of “not knowing what to use it for.” Even if people know the names of tools like ChatGPT or Claude, they can’t visualize how to apply them to their own business operations.
Second, there is the concern of “security worries.” Especially for companies handling customer information or confidential data, there is resistance to inputting information into AI.
Third, there is the difficulty of management decision-making due to “unclear implementation costs and man-hours.” For small and medium-sized enterprises that cannot afford to assign a dedicated person, even the time for trial and error is precious.
These challenges are actually based on major misunderstandings. Implementing generative AI is neither as difficult nor as expensive as it seems.
Specific Strategies to Start for Under 0
In my own company, we are generating value equivalent to approximately $53,000 per year with a monthly operating cost of about $150. This is not an isolated case; there are many tasks where an initial investment of $700 and a monthly cost of $150 to $200 can produce sufficient results.
As a specific way to start, I recommend the following three steps.
Step 1: Start with Tasks That Don’t Handle Non-Public Information
To minimize security risks, it is safe to introduce AI starting with tasks that only handle public information. Examples include analyzing competitor press releases, summarizing industry news, and rewriting your company’s past articles.
In fact, a long-established infrastructure company reduced work hours by 60% by automating inquiry responses based on public information (as reported on mbp-japan.com). Sufficient results can be achieved without handling confidential information.
Step 2: Conduct a PoC with Free or Low-Cost Tools
There is a wealth of tools you can start with zero initial cost, such as the free version of ChatGPT, free trials of Claude, and Microsoft Copilot (starting at about $22/month for business). First, try focusing on a specific task for one week.
Based on my experience, the following tasks are particularly effective.
- Drafting emails (saving 3-5 minutes per email)
- Summarizing meeting minutes (reducing a 10-minute task to 1 minute)
- Generating internal manuals (reducing a half-day task to 30 minutes)
Step 3: Measure Effectiveness and Decide on Expansion
Evaluate the PoC results based on “time saved” and “changes in quality.” For example, if using AI for email creation saves 30 minutes a day, that’s 10 hours a month. At an hourly rate of $20, that’s $200 of value per month. If the tool costs $15 a month, you can immediately decide to expand its use.
Targeting Technique That Achieved 5x the Order Rate
The true value of generative AI is realized not just in operational efficiency, but in tasks directly linked to revenue. In a case published on PR TIMES, a company achieved a 5x increase in order rate by combining a company list, public information, and generative AI.
The specific method is as follows.
- Have the AI analyze your existing customer list to identify common attributes (industry, revenue size, challenges).
- Based on the analysis, generate a list of new target companies (using public information).
- Use AI to create optimized proposals for each company.
The key point of this method is that it is completed using only public information. It uses no personal information or confidential data from business partners, reducing security risks to zero.
I have also applied this method to automate the creation of sales lists for clients. A task that used to take a salesperson 30 minutes per company can now be completed in 5 minutes using AI. In many cases, the accuracy surpasses human effort, and AI’s advantage is particularly notable in generating hypotheses like “this proposal will resonate with this company.”
Checklist for Overcoming Adoption Barriers
Finally, here is a checklist for executives and CTOs considering generative AI adoption.
- ☐ Have you identified one person in your company who “wants to try it”?
- ☐ Have you listed at least three tasks that can be completed using only public information?
- ☐ Have you planned a one-week PoC using free tools?
- ☐ Have you decided on metrics for measuring effectiveness (time saved, quality score, etc.)?
- ☐ Have you secured a budget of $35 or less per month?
If you answered “yes” to all five, you should start now. The 16% barrier can be broken simply by changing “reasons not to start” into “reasons to start.”
Generative AI is not just for large corporations. On the contrary, regional companies struggling with labor shortages stand to benefit the most. The cost starts at a few tens of dollars per month, and the implementation period starts at one week. Everything begins with starting small and experiencing the results firsthand.
Comments