What Happens in a World Where a Phone Operator Costs 200,000 Yen a Month, but Now Only 5,000 Yen?

Hiring one part-time employee as a phone operator costs between 150,000 and 200,000 yen a month. Even then, they can only handle calls during business hours.

Calls received at night or on holidays are missed. By the time you call back, the caller may have already contacted a competitor. Small and medium-sized enterprises in rural areas have likely experienced how this “missed call” directly impacts sales.

However, now an AI phone agent can handle calls 24/7 for just 5,000 yen a month. This is about 1/30 of the traditional cost. This is not just a matter of “becoming more convenient”; the economic structure of phone handling has fundamentally changed.

This article focuses on three tools: the phone AI “Clara”, the AI kanban board “QodFlow”, and the offline memory tool “Engram”. Together, they cost a total of 5,000 yen per month. By combining these three, the flow of “receiving calls → converting to tasks → recording information” can start operating without any human intervention.

We will create a “self-ending” business this week. Here’s how to do it.

—

1. Clara Phone Agent — The “Phone Operator” Disappears for 5,000 Yen a Month

What Changes?

Clara is an AI phone agent designed for local businesses. It handles calls in Japanese 24/7, managing reservations, business hour inquiries, and frequently asked questions on behalf of humans.

Let’s look at the numbers.

• Part-time phone operator: 150,000 to 200,000 yen per month (including social insurance), working 8 hours a day.
• Clara: 5,000 yen per month, available 24/7.

There’s a 30 to 40 times difference in simple cost comparison. Moreover, humans can fall ill or leave; Clara does not.

The important point is not just that “labor costs for phone handling are saved”. The person who used to be a phone operator can now focus on generating sales. In small and medium-sized enterprises, the biggest waste is the time capable employees spend tied to the phone.

What to Do This Week: Setup Steps

1. Create an Account (10 minutes): Sign up for a monthly plan on Clara’s official website. You can start immediately if you have a credit card.
2. Set Up Call Forwarding (15 minutes): Forward calls from your company’s phone line to Clara’s number. If you use NTT or Hikari Denwa, this can be completed in just a few clicks from the management screen. It’s recommended to set up conditional forwarding that transfers calls if not answered within three rings. Start by having a human answer first, and only route to AI when they can’t.
3. Input Business Information (30 minutes): Enter business hours, service details, pricing, and frequently asked questions in text format. You don’t need to aim for perfection at first. Writing down ten things that customers ask at least five times a week will cover 80% of the calls.
4. Test Call (15 minutes): Call Clara from your smartphone. If there are any odd responses, correct them on the spot.
5. Start Operation & Improve (5 minutes daily): Check Clara’s call logs every morning. If there are unexpected questions, add the information. The accuracy will significantly improve within a week.

Time Required: 1 hour on the first day, then 5 minutes daily.

Points to Note

Clara is not omnipotent. Complaints and complex inquiries should be handled by a human. That’s why it’s best to start with “conditional forwarding”. AI can handle initial inquiries and escalate to a human if necessary. Once this separation is established, 70-80% of phone handling can be in a “self-ending” state.

—

2. QodFlow — AI Kanban Board Eliminates “Progress Check Meetings”

What Changes?

“What’s the status of that project?”

How many times does this question fly around the office in a day? Meetings just for progress checks, confirmations via chat, verbal reports — all of these arise from the lack of centralized information.

QodFlow is an AI-powered kanban board. Unlike regular kanban boards (like Trello or Notion), AI automatically updates the status of tasks. The AI reads the content of emails, chat exchanges, and file updates, automatically moving cards from “In Progress” to “Under Review” to “Completed”.

And it’s free. Zero yen.

For small and medium-sized enterprises, project management tools that cost several thousand yen a month can be a significant burden. Moreover, such tools are often too multifunctional, leading to underutilization. QodFlow adds AI’s automatic updates to the simplest form of kanban, making it usable on the ground.

What to Do This Week: Setup Steps

1. Create an Account (5 minutes): Create a free account on the QodFlow official website.
2. Create a Board (15 minutes): Start by creating just one board. Recommended topics include “Orders for This Week” or “Inquiry Responses”, where progress checks frequently occur. Four columns — “Not Started”, “In Progress”, “Under Review”, and “Completed” — are sufficient.
3. Input Tasks (20 minutes): Add currently ongoing projects as cards. Link them to responsible persons, deadlines, and related files.
4. Set Up AI Automatic Updates (15 minutes): Configure integration settings with email and chat tools. QodFlow’s AI will detect related exchanges and automatically update the status.
5. Share with the Team (10 minutes): Share the board’s URL with the team. A simple note saying, “Check here for progress” will suffice.

Time Required: 1 hour on the first day. Operational costs thereafter are nearly zero.

Why It Works on the Ground

Just eliminating the question “What’s the progress?” can save 30 minutes to an hour a day. For a team of five, this translates to 50 to 100 hours of “confirmation work” disappearing each month. At an hourly wage of 1,500 yen, that’s equivalent to 75,000 to 150,000 yen a month. All with a zero-yen tool.

—

3. Engram — Offline AI Memory Eliminates “Where Did That Information Go?”

What Changes?

Information management in small and medium-sized enterprises is often stored in the “heads of veteran employees”. The name of the contact person at a client, the history of past troubles, machine settings — they can tell you if you ask, but if that person is absent, no one else knows.

This is the essence of dependency on individuals. The moment this person leaves, the company’s knowledge assets can vanish.

Engram is an offline AI memory tool. It can be installed on your company’s server or local PC. Since there’s no need to store information in the cloud, confidential information and customer data can be securely stored.

Transform the “knowledge in veteran heads” into a searchable database.

And it’s free. Just download and install it from GitHub.

What to Do This Week: Setup Steps

1. Installation (20 minutes): Download Engram from GitHub and install it on your company’s PC or server. If you have a Docker environment, it can be set up with a few commands. If not, a Windows installer is also available.
2. Input Initial Data (1 hour): Start by entering “frequently asked questions”. Client lists, key points from business manuals, records of past troubles. There’s no need to create perfect documentation; bullet points are fine. The AI will organize it into a searchable format.
3. Search Test (10 minutes): Ask questions like, “Who is the contact person for Company XX?” or “What were the steps for handling the last machine trouble?” Confirm that the AI retrieves the relevant information.
4. Establish Daily Operation Rules (15 minutes): Just set a rule that says, “When new information arises, leave a quick note in Engram”. Once a week for 10 minutes is sufficient. This will accumulate information over time.

Time Required: 2 hours on the first day. Thereafter, 10 minutes weekly.

The Importance of Being Offline

Cloud-based knowledge tools are convenient, but there’s always the question of whether it’s acceptable to store customer information on external servers. Especially for local small and medium-sized enterprises, trust with clients is the foundation of business. They want to minimize the risk of information leaks.

Engram operates offline, so this issue does not exist. It works even without an internet connection. It can be used in factories or offices in mountainous areas. Not being dependent on internet connectivity is a significant value for local small and medium-sized enterprises.

—

Cost Comparison: Three Operations That “Self-End” for 5,000 Yen a Month

With an investment of 5,000 yen a month, operational costs equivalent to 225,000 to 350,000 yen disappear. That’s 2.6 to 4.2 million yen annually. Can any small or medium-sized enterprise look at these numbers and say, “This doesn’t concern us”?

—

The Meaning of “What to Do This Week”

The total initial setup time for the three tools is about 4 hours. You can complete it in one afternoon.

The important thing is not to “prepare perfectly before starting”, but to start moving first and adjust while using it. Clara’s response accuracy can be 60% at first. With feedback over a week, it can reach 80%. QodFlow can start with just one board. If it works well, you can add more. Engram’s data can begin with just ten bullet points.

The strength of small and medium-sized enterprises is their speed in decision-making. While large companies launch “AI implementation projects”, select vendors, and go through approvals, taking six months for a PoC, small and medium-sized enterprises can implement these three tools and get operations running in that time.

“Not next month, but this week.”

This sense of speed is the only and greatest weapon that small and medium-sized enterprises have to compete with large companies. Just 5,000 yen and four hours one afternoon. Even if you fail, you have almost nothing to lose. If you succeed, operations worth hundreds of thousands of yen a year will be in a “self-ending” state.

Instead of looking for reasons not to do it, I’d like you to clear just one hour in your afternoon this week. Start by creating an account for Clara.