How to Create an AI-Personalized Script (Adapting to Japanese Business Practices): Real Examples of Coffee Chat Conversations and Follow-Up Emails

1) First, fix the “counterpart image” and the “purpose of the setting”.

Script creation improves accuracy more by first fixing the other party and the purpose rather than deciding the order of lines in advance. In Japanese business contexts, the success or failure often hinges on the tone of the small talk, how respect is expressed, and the wording used to close.

• Target person: Job title, expertise, age range (whether they are a mid-career professional), and past involvement (shared points)
• Purpose of the meeting: Information gathering, introductions, maintaining relationships, next appointment
• Prerequisite: Adjust the approach (LINE/Slack/email), and consider the days of the week when the other person is more likely to have time.

Once you decide these three, AI is less likely to get wrong who it should address, what it should cover, and how far it should go.

2) Create a “Conversation Flow” template that fits Japanese business practices

The script for a coffee chat isn’t just about the content of the conversation—the structure and flow of the discussion are also important. In Japanese settings, what’s valued is that the opening greeting, the level of detail in your self-introduction, the cushion you use to acknowledge and respond to the other person’s words, and the closing (including a clear next action) feel natural.

1. Introduction: Thank-you + Confirm the time required (brief)
2. Purpose shared: One word for the topic we want to discuss today
3. Question: The order that makes it easiest for the other person to talk (experience → approach → learning)
4. Reply: Summary + empathy + a comment with a brief pause
5. Close: A thank-you plus a sentence that leads to next steps (follow up in an email/message)

The key is to create a rhythm where the conversation doesn’t stall and the other person feels at ease to keep talking, rather than just saying something amusing.

3) Specify the prompt to be given to the AI, all the way up to the “output format”

To make the most of AI’s writing ability, the trick is to have it produce the script as “conversation parts,” rather than letting it create it from “writing fragments.” By asking it to output in the same structure as the example below, template-based operations become easy.

4) Separate the conversation scripts and follow-up emails by “role”

Conversation plays the role of creating a state where the other person is ready to speak. Email serves the roles of “a memory aid” and “making the next step clear.” If we divide the AI’s output into these roles, the tone stays consistent.

5) Example: Email after a coffee chat (Japanese → English)

Below is a practical way to write out the conversation flow “in email form.” Replace the numbers and proper nouns with those that fit your conditions.

6) LINE/Slack is designed as a “short version”

The key to message apps isn’t just “making sentences shorter”—it’s also about “making them easier to reply to” by organizing them. Start with a brief line that includes a thank-you and the context, ask just one question, and end with a single, one-line closing sentence.

• LINE/Slack: 3〜5 lines, 1 question
• To account for a possible delay in my reply, I’ll avoid stating a strict deadline
• Translate from Japanese to English so that it has the same meaning and use consistent proper terms.

7) The finishing is done using the “Quality check items”

AI output is convenient, but in the end, people refine it. When you standardize a checklist, quality becomes consistent.

• Is the level of formality in the wording appropriate for the other person?
• Is your question specific enough and easy for the other person to answer?
• Is the closing message unambiguous, with the next action clearly defined?
• Check for any contradictions between your conversations and emails.