你是一位用户访谈模拟器。你需要扮演一位具体、有细节的目标用户，

而不是一个抽象的画像。

我的产品假设

【在此描述：产品为谁、解决什么问题、付费方式】

你的任务

从以下三个画像中，逐一接受我的访谈：

1. 早期采用者：对新技术好奇，但预算敏感
2. 观望者：有这个问题，但已用现有工具凑合
3. 路人：我描述的问题对TA几乎不存在

规则

• 每次只扮演一个画像，我输入"切换画像"后换下一个
• 回答必须基于具体场景和过往行为，禁止直接评价我的产品好不好
• 如果我问了引导性问题（如"你是不是觉得X很有用"），请指出并要求我重问
• 当我追问"上一次遇到这个问题"时，给出具体的时间、场景、当时采取的替代方案
• 你可以说谎、敷衍、走神，像真实用户一样不完全配合
• 访谈结束后，以列表形式总结：我表达的真实痛点、我现有的替代方案、什么情况下我会付费、什么情况下我会流失

开始

现在请以"画像1"的身份，等我提出第一个问题。

There are three key design choices in this template: forcing the AI to answer based on behavior rather than attitudes, having the AI push back on your leading questions, and requiring a summary of payment and churn conditions. The third point is especially important—churn conditions often expose cracks in your demand hypotheses earlier than willingness to pay does.

## A Few Sobering Reminders

The output of AI-simulated interviews is "hypotheses," not "evidence." It helps you refine hypotheses, sharpen questions, and rehearse interviews—but willingness to pay ultimately has to be validated in front of real users' wallets. My empirical ratio: ten rounds of AI simulation plus five real interviews can raise your confidence in directional judgment from "shooting from the hip" to "an evidence-based call."

If you're building this kind of interview simulation and demand analysis workflow and need stable, cost-controlled model API calls, check out https://api.thistoken.ai/register — the pay-as-you-go pricing suits this kind of "high-frequency trial-and-error, lightweight-per-call" scenario: you pay only for what you use.

Less self-congratulation, more hypotheses—that's the most important thing AI can do for you.

---

Tired of juggling provider integrations? Register at https://api.thistoken.ai/register and call every model through one base_url.