icp-persona-builder

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

ICP Persona Builder

ICP 用户角色构建工具

Research a company's buyer segments and build detailed synthetic personas that model their ideal customers. These personas become a reusable client asset — once built, any skill can load them to evaluate content, messaging, websites, or campaigns through buyer eyes.

研究公司的买家细分群体，构建能够模拟其理想客户的详细合成用户角色。这些角色将成为可复用的客户资产——一旦创建完成，任何技能都可以加载它们，从买家视角评估内容、营销话术、网站或推广活动。

Quick Start

快速开始

Build ICP personas for [company]. Their site is [url].

With known ICPs:

Build personas for [company]. Their ICPs are: [ICP 1], [ICP 2], [ICP 3].

Build ICP personas for [company]. Their site is [url].

若已知ICP：

Build personas for [company]. Their ICPs are: [ICP 1], [ICP 2], [ICP 3].

Inputs

输入项

Input	Required	Source
Company name	Yes	User provides
Company URL	Recommended	Helps with research
Known ICPs	No	User provides, or discovered via research
Client context file	No	`clients/<client>/context.md` if available

输入项	是否必填	来源
公司名称	是	用户提供
公司官网URL	推荐提供	辅助调研
已知ICP	否	用户提供，或通过调研发现
客户上下文文件	否	若存在则位于 `clients/<client>/context.md`

Process

执行流程

Phase 1: Company Research

阶段1：公司调研

Understand what the company does and who they serve:

WebFetch their website — homepage, product/solutions pages, pricing, "who it's for" pages
WebSearch for:
- "[company] customers" / "[company] case studies"
- "[company] reviews" (G2, Capterra, TrustRadius)
- "[company] vs" (comparison searches reveal buyer segments)
- "[company] jobs" (who they're hiring to sell to / support)
Extract signals:
- What problem do they solve?
- What's their pricing/packaging? (Signals ACV and buyer type)
- What industries/verticals do they serve?
- What company sizes do they target?
- What roles/titles appear in case studies and testimonials?
- What's their go-to-market motion? (Self-serve, sales-led, hybrid)

了解公司业务及服务对象：

抓取公司官网内容——首页、产品/解决方案页面、定价页、“适用人群”页面
进行网络搜索，关键词包括：
- "[company] customers" / "[company] case studies"
- "[company] reviews" (G2, Capterra, TrustRadius)
- "[company] vs" (对比搜索可挖掘买家细分群体)
- "[company] jobs" (从招聘岗位判断其目标服务对象)
提取关键信号：
- 他们解决什么问题？
- 他们的定价/套餐模式是什么？（反映客户平均价值（ACV）和买家类型）
- 他们服务哪些行业/垂直领域？
- 他们针对哪些规模的公司？
- 客户案例和推荐语中出现哪些岗位/头衔？
- 他们的市场推广模式是什么？（自助式、销售主导式、混合式）

Phase 2: Identify ICP Segments

阶段2：识别ICP细分群体

From the research, identify 4-6 distinct buyer segments. Each segment should represent a meaningfully different type of buyer — different role, different company profile, or different buying motivation.

For each segment, define:

Attribute	Description
Segment name	Short label (e.g., "Enterprise IT Leader", "Startup Founder", "Agency Operator")
Role/titles	Typical job titles in this segment
Company profile	Size, stage, industry, tech stack
Core pain point	The #1 problem driving them to look for a solution
Buying trigger	What event makes them start searching NOW
Decision criteria	What matters most when evaluating (ranked)
Sophistication	How well they understand the problem space and solution landscape
Alternatives	What else they'd consider (competitors, DIY, status quo)
Segment size estimate	Rough sense of how big this segment is for the company (primary, secondary, emerging)

Segment diversity rules:

At least one technical buyer (evaluates capabilities, architecture, integrations)
At least one business buyer (evaluates ROI, outcomes, competitive advantage)
At least one skeptical profile (has been burned before, hard to convince)
At least one junior/researcher (doing initial research for a decision-maker)
Try to cover different company sizes if the company serves multiple tiers

基于调研结果，识别4-6个不同的买家细分群体。每个群体应代表具有显著差异的买家类型——比如不同岗位、不同公司特征或不同购买动机。

为每个群体定义以下属性：

属性	说明
细分群体名称	简短标签（例如：“企业IT负责人”、“初创公司创始人”、“代理机构运营者”）
岗位/头衔	该群体的典型岗位头衔
公司特征	规模、发展阶段、行业、技术栈
核心痛点	促使他们寻找解决方案的首要问题
购买触发因素	促使他们立即开始搜索解决方案的事件
决策标准	评估时最看重的因素（按优先级排序）
专业程度	他们对问题领域和解决方案格局的了解程度
替代选项	他们可能考虑的其他选择（竞品、自制方案、维持现状）
群体规模预估	该群体对公司的重要性规模（核心、次要、新兴）

细分群体多样性规则：

至少包含一个技术型买家（评估产品功能、架构、集成能力）
至少包含一个业务型买家（评估投资回报率、业务成果、竞争优势）
至少包含一个疑虑型用户（曾有过负面经历，难以说服）
至少包含一个初级/调研型用户（为决策者做初步调研）
如果公司服务多个规模层级的客户，尽量覆盖不同公司规模

Phase 3: Build Synthetic Personas

阶段3：构建合成用户角色

For each segment, create a detailed synthetic persona. The persona should feel like a real, specific person — not a marketing abstraction.

Persona structure:

json

{
  "id": "persona-slug",
  "name": "Jordan Chen",
  "segment": "Enterprise IT Leader",
  "title": "VP of Engineering",
  "company": {
    "type": "Mid-market SaaS company",
    "size": "200-500 employees",
    "stage": "Series B, scaling fast",
    "industry": "Financial services technology"
  },
  "demographics": {
    "experience_years": 12,
    "reports_to": "CTO",
    "team_size": 35,
    "budget_authority": "$50K-200K without board approval"
  },
  "situation": "Jordan's team is growing faster than their tooling can support. They've been using a patchwork of internal scripts and are losing engineering hours to maintenance. The CTO has asked Jordan to evaluate modern solutions before next quarter's planning cycle.",
  "pain_points": [
    "Team productivity is dropping as they scale",
    "Current tools don't integrate well",
    "Onboarding new engineers takes too long"
  ],
  "buying_trigger": "CTO mandate to evaluate solutions before Q3 planning",
  "decision_criteria_ranked": [
    "Enterprise security and compliance (SOC2, SSO)",
    "Integration with existing stack (GitHub, Jira, Datadog)",
    "Scalability — will this work at 2x team size?",
    "Total cost of ownership, not just sticker price",
    "Implementation timeline — needs to be live in 6 weeks"
  ],
  "skepticism_profile": {
    "trust_level": "Low — has been burned by vendor promises before",
    "research_style": "Deep dive. Reads docs, checks GitHub issues, asks peers in Slack communities",
    "key_objections": [
      "Will this actually scale or will we outgrow it in a year?",
      "What's the real implementation cost beyond the license?",
      "How good is the support when things break at 2am?"
    ]
  },
  "technical_sophistication": "High — understands the technical landscape well, can evaluate architecture decisions, wants to see under the hood",
  "language": {
    "describes_problem_as": "We need to consolidate our toolchain and reduce operational overhead",
    "searches_for": ["engineering productivity platform", "developer tools consolidation", "[competitor] alternative enterprise"],
    "red_flag_words": ["revolutionary", "AI-powered", "seamless" — overpromising triggers skepticism],
    "trust_signals": ["SOC2 badge", "customer logos in their industry", "transparent pricing", "public changelog"]
  },
  "evaluation_behavior": {
    "first_visit": "Scans headline, checks if it's for their company size, looks for enterprise/security page",
    "deep_evaluation": "Reads docs, checks integrations list, looks for case studies from similar companies",
    "social_proof_needs": "Wants to see companies their size in their industry, not just FAANG logos",
    "deal_breakers": ["No SSO/SAML", "No self-hosted option", "Pricing only available via sales call"]
  }
}

为每个细分群体创建详细的合成用户角色。角色应贴近真实、具体的个人，而非抽象的营销概念。

用户角色结构：

json

{
  "id": "persona-slug",
  "name": "Jordan Chen",
  "segment": "Enterprise IT Leader",
  "title": "VP of Engineering",
  "company": {
    "type": "Mid-market SaaS company",
    "size": "200-500 employees",
    "stage": "Series B, scaling fast",
    "industry": "Financial services technology"
  },
  "demographics": {
    "experience_years": 12,
    "reports_to": "CTO",
    "team_size": 35,
    "budget_authority": "$50K-200K without board approval"
  },
  "situation": "Jordan's team is growing faster than their tooling can support. They've been using a patchwork of internal scripts and are losing engineering hours to maintenance. The CTO has asked Jordan to evaluate modern solutions before next quarter's planning cycle.",
  "pain_points": [
    "Team productivity is dropping as they scale",
    "Current tools don't integrate well",
    "Onboarding new engineers takes too long"
  ],
  "buying_trigger": "CTO mandate to evaluate solutions before Q3 planning",
  "decision_criteria_ranked": [
    "Enterprise security and compliance (SOC2, SSO)",
    "Integration with existing stack (GitHub, Jira, Datadog)",
    "Scalability — will this work at 2x team size?",
    "Total cost of ownership, not just sticker price",
    "Implementation timeline — needs to be live in 6 weeks"
  ],
  "skepticism_profile": {
    "trust_level": "Low — has been burned by vendor promises before",
    "research_style": "Deep dive. Reads docs, checks GitHub issues, asks peers in Slack communities",
    "key_objections": [
      "Will this actually scale or will we outgrow it in a year?",
      "What's the real implementation cost beyond the license?",
      "How good is the support when things break at 2am?"
    ]
  },
  "technical_sophistication": "High — understands the technical landscape well, can evaluate architecture decisions, wants to see under the hood",
  "language": {
    "describes_problem_as": "We need to consolidate our toolchain and reduce operational overhead",
    "searches_for": ["engineering productivity platform", "developer tools consolidation", "[competitor] alternative enterprise"],
    "red_flag_words": ["revolutionary", "AI-powered", "seamless" — overpromising triggers skepticism],
    "trust_signals": ["SOC2 badge", "customer logos in their industry", "transparent pricing", "public changelog"]
  },
  "evaluation_behavior": {
    "first_visit": "Scans headline, checks if it's for their company size, looks for enterprise/security page",
    "deep_evaluation": "Reads docs, checks integrations list, looks for case studies from similar companies",
    "social_proof_needs": "Wants to see companies their size in their industry, not just FAANG logos",
    "deal_breakers": ["No SSO/SAML", "No self-hosted option", "Pricing only available via sales call"]
  }
}

Phase 4: Save Persona Assets

阶段4：保存用户角色资产

Save to the client directory as reusable assets:

clients/<client>/personas/personas.json
— Machine-readable, all personas in an array:

json

{
  "company": "Acme Corp",
  "url": "https://acme.com",
  "created": "2026-02-26",
  "segment_count": 5,
  "personas": [ ... ]
}

clients/<client>/personas/personas.md
— Human-readable Markdown with all personas written out in prose form, easy to review and share.

clients/<client>/personas/segments.md
— Summary table of all segments with key attributes, useful as a quick reference.

将角色保存到客户目录中，作为可复用资产：

clients/<client>/personas/personas.json
— 机器可读格式，所有角色以数组形式存储：

json

{
  "company": "Acme Corp",
  "url": "https://acme.com",
  "created": "2026-02-26",
  "segment_count": 5,
  "personas": [ ... ]
}

clients/<client>/personas/personas.md
— 人类可读的Markdown格式，所有角色以散文形式呈现，便于审阅和分享。

clients/<client>/personas/segments.md
— 包含所有细分群体关键属性的汇总表格，可作为快速参考工具。

Output Summary

输出内容摘要

After building, present:

Segment overview table — All segments with key attributes at a glance
Persona summaries — 2-3 sentence summary of each persona
Coverage check — Confirm diversity rules are met (technical, business, skeptical, researcher)
Next steps — Suggest running
```
icp-website-review
```
or other skills that can use the personas

角色构建完成后，输出以下内容：

细分群体概览表——一目了然展示所有细分群体的关键属性
用户角色摘要——每个角色用2-3句话概括
覆盖度检查——确认符合多样性规则（技术型、业务型、疑虑型、调研型用户均已覆盖）
后续建议——推荐运行
```
icp-website-review
```
或其他可调用这些角色的技能

Tips

小贴士

Research depth matters. Spend real time in Phase 1. The better you understand the company's actual customers, the more realistic the personas. Don't just read the homepage — dig into reviews, case studies, job postings.
Make personas specific. "Marketing Manager" is too generic. "Sarah, Senior Demand Gen Manager at a 50-person B2B SaaS startup who just lost her SDR team to budget cuts" tells you exactly how she'll evaluate a tool.
Include the language dimension. How the persona describes their problem is often completely different from how the vendor describes their solution. This gap is where messaging fails.
Skepticism is the most important trait. Every persona needs a clear skepticism profile. What would make them NOT buy? What's their default assumption about vendors?
This skill has no code script. It's agent-executed using WebSearch and WebFetch. The structured process above guides the research and persona creation.

调研深度至关重要。 在阶段1投入足够时间。对公司实际客户的了解越深入，创建的角色就越贴近真实。不要仅浏览首页——深入研究客户评价、案例和招聘信息。
让角色具体化。 “营销经理”过于笼统。而“Sarah，50人规模B2B SaaS初创公司的高级需求生成经理，其SDR团队因预算削减被解散”这样的描述能明确体现她评估工具的视角。
涵盖语言维度。 用户角色描述问题的方式往往与供应商描述解决方案的方式截然不同。这种差异正是营销话术失效的原因。
疑虑特征是最重要的属性。 每个角色都需要明确的疑虑特征。什么会让他们放弃购买？他们对供应商的默认看法是什么？
该技能无代码脚本。 由Agent通过WebSearch和WebFetch执行。上述结构化流程将指导调研和角色创建工作。

Dependencies

依赖项

Web search capability (for company and ICP research)
Web fetch capability (for reading website pages)
No API keys or paid tools required

网络搜索能力（用于公司和ICP调研）
网页抓取能力（用于读取官网页面内容）
无需API密钥或付费工具