AI Takes Calls for Local Businesses 24/7, and Also Searches for Investors—A List of ‘Zero Labor Cost Operations’ and Their Pitfalls

A Phone Operator’s Salary Reduced from 300,000 Yen to 50,000 Yen

To get straight to the point, starting this month, local small and medium-sized enterprises can have operations with ‘zero labor costs.’ Phone answering, investor list creation, and simple website production are now at a stage where they can be entrusted to AI.

However, it is not simply a matter of replacing everything with AI. There are operations that can be replaced and those that absolutely should not be. Misjudging this boundary can lead to losing customers instead of cutting costs.

In this article, we will list the ‘operations that can be replaced starting this month’ and organize their actual costs, risks, and outcomes numerically.

—

List of Replaceable Operations

1. 24/7 Phone Answering—From Two Part-Time Workers to 50,000 Yen

AI phone answering services are rapidly reaching practical levels. A representative example is AI phone answering systems like ‘Clara.’

Traditionally, when local small and medium-sized businesses tried to handle phone calls outside of business hours, they had two options:

• Hire part-time workers: 150,000 to 300,000 Yen (Calculating at 1,000 Yen/hour for 8 hours a day over 20 days amounts to 160,000 Yen. Night and holiday work incurs additional costs.)
• Use a phone answering service: 30,000 to 100,000 Yen (However, there is a limit on the number of calls; exceeding this incurs additional fees.)

AI phone answering can provide 24/7 service for 20,000 to 50,000 Yen per month. It also has a forwarding function to human operators. In other words, it is now possible to design a system where ‘AI handles initial responses and only forwards cases that require human judgment.’

This is significant for local small and medium-sized enterprises. Resources that were tied up with one employee dedicated to phone answering in a five-person company can now be freed up. That one person can now focus on sales. This creates not just cost savings but also the capacity to increase sales. This is the essence of the matter.

#### Pitfall: The Cost of Incorrect Responses is High

However, clear risks exist. Research data shows that if AI provides an inappropriate response, the customer attrition rate can reach about 30%. This means that 3 out of 10 people feel, ‘I will not call this company again.’

Particularly dangerous is the case where complaint handling or urgent inquiries are left to AI. There is nothing that adds fuel to the fire more than an AI responding with a template to an angry customer.

The solution is simple. Set the ‘scope of AI responses’ narrow from the beginning. Let AI handle only these three tasks: business hours guidance, appointment scheduling, and accepting callback requests. Everything else should be immediately forwarded to a human. Aim for ‘zero missed calls’ rather than a perfect response.

—

2. Automatic Creation of Investor and Sales Lists—From Days of Work to Hours

A developer is conducting an interesting experiment. They created a CLI tool that automatically collects email addresses of investors likely to be interested in their app using an AI agent and sends personalized emails automatically.

The result: 43 emails sent, 6 replies. A response rate of 14%.

Anyone who has done sales will understand this figure. The average response rate for cold emails is said to be between 1% and 5%. A 14% response rate is exceptionally high. The reason is that AI analyzes the investment tendencies of each investor and customizes the email content accordingly.

Now, let’s consider this in the context of local small and medium-sized enterprises. For example, suppose a manufacturing company wants to develop new business partners. Traditionally, the president or sales representative would attend trade shows, collect business cards, and write individual emails. If that task took 10 hours a week, an AI agent could achieve equal or better results in just 2 to 3 hours.

#### Pitfall: ‘Automated Sales’ are Disliked

The problem lies in whether the recipient realizes it. If a large number of emails that are clearly templates arrive, they will be treated as spam. In the worst-case scenario, the domain could end up on a blacklist.

What matters is not ‘quantity’ but ‘accuracy.’ Twenty precise emails will yield better results than 100 poorly crafted ones. Let AI handle the list creation and draft preparation. Before hitting the send button, a human should review the emails. This extra step can prevent accidents.

—

3. Simple Website Creation—From 1,000,000 Yen to 50,000 Yen

There has been buzz around a developer who earns 1,000 dollars (about 150,000 Yen) per hour using AI for development. They reportedly create marketing sites in a short time using AI tools and earn between 1,000 to 2,000 dollars (15,000 to 30,000 Yen) per site.

The essence of this case is not that developers are making money. It is that the market for website production is collapsing.

When local small and medium-sized enterprises request a homepage from a production company, the market rate is between 500,000 to 1,500,000 Yen. Now, if you can hire someone who can use AI tools, it can be done for 50,000 to 300,000 Yen. Furthermore, with the combination of no-code tools and AI, the barrier has been lowered to the point where someone in the company can create it themselves.

#### Pitfall: ‘Being Able to Create’ and ‘Delivering Results’ are Different Matters

Websites created with AI may look reasonably well-designed. However, the design of ‘who to communicate with, what to convey, and how to prompt action’ cannot solely come from AI. Understanding what local customers are seeking and how to differentiate from competitors is a matter of management judgment and human work.

In other words, while AI has dramatically lowered the ‘production cost’ of websites, the ‘design cost’ will not be zero. Rather, as production costs decrease, the value of design is relatively increasing.

—

Operations That Can Be Replaced and Those That Should Not Be

Let’s summarize the examples so far.

Operations that can be replaced (i.e., routine tasks with limited impact from errors):

• Initial phone responses outside business hours
• Creation of prospect and business partner lists
• Drafting of standard emails
• Simple website creation
• Automatic generation of invoices and estimates
• Creation of meeting minutes
• Drafting of social media posts

Operations that should not be replaced (i.e., those requiring trust and judgment):

• Complaint handling
• Negotiations with important clients
• Job interviews
• Decision-making related to management judgments
• Building relationships with local communities

The boundary is clear. ‘Operations that can be recovered from failures’ can be entrusted to AI. ‘Operations that would lose trust if failed’ should be handled by humans. Remembering just this is sufficient.

—

So, What Should We Do?

When local small and medium-sized enterprises introduce AI, the biggest mistake is to ‘mimic large companies.’ Large companies invest tens of millions of yen to build AI systems. Small and medium-sized enterprises do not need to do that.

There are only three things to do.

1. Start by replacing just one operation. Whether it’s phone answering or email creation, introduce one tool that can be tested for under 50,000 Yen per month.

2. Use it for a month and analyze the numbers. Look at the number of responses, the number of errors, and the time saved. Make judgments based on numbers, not feelings.

3. If it works well, expand to the next operation. Don’t try to change everything at once.

With AI phone answering at 50,000 Yen, you can save 150,000 Yen in labor costs, allowing that person to focus on sales and generate 500,000 Yen in revenue. This kind of ‘domino effect’ structure is a strength of small enterprises with few employees. In large companies, different departments make this impossible.

Now that the cost of AI has dramatically decreased, the ‘risk of not using it’ has become greater. Competitors in neighboring towns are already implementing AI, answering calls 24/7, automatically creating sales lists, and refreshing their websites for 50,000 Yen per month. Meanwhile, what is your company doing?

Start with just one. I hope you will try it by the end of this month.