\r \r

Agentic Workflow Designer\r

\r

From messy manual processes to autonomous AI pipelines — design, document, and deploy.\r \r

What This Skill Does\r

\r Agentic AI (AI that can autonomously execute multi-step tasks) is the #1 enterprise tech trend in 2026 with a projected $8.5B market and 40% CAGR. Yet most teams struggle to:\r

• Map which workflows are actually suitable for agentic automation\r
• Design reliable pipelines that don't break silently\r
• Choose between n8n, Make, Zapier, or custom agent frameworks\r
• Justify the ROI to business stakeholders\r \r This skill bridges the gap between AI hype and practical workflow automation:\r \r
• Workflow Discovery — Identify and prioritize automation opportunities in any business process\r
• Agentic Pipeline Design — Create detailed workflow blueprints with triggers, agents, tools, and fallbacks\r
• Platform Selection — Compare n8n / Make / Zapier / custom LangGraph for your use case\r
• Generate Workflow Specs — Produce JSON/YAML specs importable into n8n or Make\r
• ROI Calculator — Estimate time/cost savings from automation\r
• Human-in-the-Loop (HITL) Design — Design appropriate checkpoints for sensitive decisions\r \r

Trigger Words\r

\r Agentic workflow, automate my process, workflow automation, n8n, Make automation, Zapier flow, design a workflow, workflow design, process automation, automate with AI, AI pipeline, autonomous workflow, HITL pattern, 工作流设计, 自动化工作流, 流程自动化, 智能体工作流, 帮我设计流程, 自动化这个流程, n8n工作流, 企业自动化, RPA替代, agentic AI pipeline\r \r

Target Users\r

\r

• Operations managers digitizing manual business processes\r
• Developers building production AI automation systems\r
• Product managers scoping automation features\r
• Consultants delivering workflow automation projects\r
• Entrepreneurs building AI-native products\r \r

Workflow\r

\r

新增内容（2026版）\r

Step 2 新增技术评估（2026）：\r

• LangGraph v1.0生产就绪：状态机工作流/长期记忆/错误恢复三大核心能力，企业级部署支持Kubernetes自动扩缩容，GitHub Stars突破85K\r
• CrewAI v1.10多智能体协作：支持6种角色类型+并行任务编排，内置20+企业级连接器（Slack/Notion/Airtable/GitHub），2026年Q1新增中文文档\r
• Claude Agent SDK / OpenAI Agents SDK横向对比：工具调用准确率(94% vs 91%)/上下文利用率(78% vs 82%)/成本效率(¥0.8/千Token vs ¥1.2/千Token)三大维度全面评测\r
• MCP(Model Context Protocol)生态爆发：50+官方服务器覆盖GitHub/Slack/Notion/Postgres等，企业内部MCP注册表成为新基础设施\r
• LLM长上下文之战：Gemini 2M Token / Claude 200K / GPT-4o 128K技术选型指南，针对金融长文档(招股书/年报)场景给出最优性价比方案\r \r ---\r \r

新增内容（2026版）\r

Step 2 新增技术评估（2026）：\r

• LangGraph v1.0生产就绪：状态机工作流/长期记忆/错误恢复三大核心能力，企业级部署支持Kubernetes自动扩缩容，GitHub Stars突破85K\r
• CrewAI v1.10多智能体协作：支持6种角色类型+并行任务编排，内置20+企业级连接器（Slack/Notion/Airtable/GitHub），2026年Q1新增中文文档\r
• Claude Agent SDK / OpenAI Agents SDK横向对比：工具调用准确率(94% vs 91%)/上下文利用率(78% vs 82%)/成本效率(¥0.8/千Token vs ¥1.2/千Token)三大维度全面评测\r
• MCP(Model Context Protocol)生态爆发：50+官方服务器覆盖GitHub/Slack/Notion/Postgres等，企业内部MCP注册表成为新基础设施\r
• LLM长上下文之战：Gemini 2M Token / Claude 200K / GPT-4o 128K技术选型指南，针对金融长文档(招股书/年报)场景给出最优性价比方案\r \r ---\r \r

Step 1 — Process Discovery\r

Ask the user to describe their current workflow:\r

• What triggers it? (email, schedule, webhook, human action?)\r
• What are the key steps? (list them in plain language)\r
• Who (or what system) does each step today?\r
• Where do errors/delays typically occur?\r
• What's the desired output/outcome?\r \r

Step 2 — Automation Suitability Assessment\r

\r Score the workflow across 5 dimensions:\r \r | Dimension | Score | Notes |\r |-----------|-------|-------|\r | Repetitiveness | /10 | How often does this run identically? |\r | Rule-based | /10 | Are decisions clear-cut or judgment-based? |\r | Data availability | /10 | Is input data structured and accessible? |\r | Error tolerance | /10 | Can errors be caught and recovered automatically? |\r | Stakes | /10 (inverted) | Low-stakes = easier to automate |\r | Automation Score | /50 | >35 = High priority, 20–35 = Medium, \x3C20 = Keep manual |\r \r

Step 3 — Agentic Pipeline Design\r

Generate a detailed pipeline blueprint:\r \r

🎯 Workflow: [Name]\r
⚡ Trigger: [webhook / cron / event / manual]\r
🤖 Agents:\r
  ├── Agent 1 [Role]: [Tool 1, Tool 2] → Output: [description]\r
  ├── Agent 2 [Role]: [Tool 3] → Output: [description]\r
  └── Agent 3 [Role]: [Tool 4, Tool 5] → Output: [description]\r
🔄 Flow: Sequential / Parallel / Conditional\r
🧠 Memory: [ephemeral / Redis / vector DB]\r
🚨 Error Handling: [retry / fallback agent / human escalation]\r
👤 HITL Checkpoints: [list high-stakes decision points]\r
📊 Output: [final deliverable description]\r
```\r
\r
**Example — Lead Qualification Pipeline:**\r
```\r
🎯 Workflow: B2B Lead Qualification & Outreach\r
⚡ Trigger: New form submission webhook\r
🤖 Agents:\r
  ├── Enrichment Agent [Clearbit + LinkedIn scraper] → Company profile JSON\r
  ├── Scoring Agent [GPT-4o] → Lead score (0–100) + reasoning\r
  ├── Decision Gate [Human] → Approve for outreach? (HITL)\r
  └── Outreach Agent [Email API + CRM API] → Personalized email + CRM update\r
🔄 Flow: Sequential with HITL gate\r
🧠 Memory: PostgreSQL (lead history)\r
🚨 Error: Retry enrichment 3x → flag for manual review\r
👤 HITL: Score > 80 auto-approves; 50–80 requires human review; \x3C50 auto-rejects\r
📊 Output: CRM updated + email queued\r
```\r
\r
### Step 4 — Platform Recommendation\r
\r
| Platform | Best For | Agent Support | Self-host | Price |\r
|----------|----------|--------------|-----------|-------|\r
| n8n | Technical teams, complex logic | ✅ via AI nodes | ✅ Yes | Free/OSS |\r
| Make (Integromat) | Non-technical, API integrations | Partial | ❌ No | ~$9+/mo |\r
| Zapier | Simple triggers, non-technical | Partial | ❌ No | ~$20+/mo |\r
| LangGraph (custom) | Complex state machines, production | ✅ Native | ✅ Yes | Dev hours |\r
| CrewAI | Role-based agent teams | ✅ Native | ✅ Yes | Dev hours |\r
\r
### Step 5 — n8n Workflow JSON Spec (Sample Output)\r
```json\r
{\r
  "name": "Lead Qualification Pipeline",\r
  "nodes": [\r
    {\r
      "name": "Webhook Trigger",\r
      "type": "n8n-nodes-base.webhook",\r
      "parameters": { "path": "lead-inbound" }\r
    },\r
    {\r
      "name": "Enrich Lead",\r
      "type": "@n8n/n8n-nodes-langchain.agent",\r
      "parameters": {\r
        "promptType": "define",\r
        "text": "Enrich this lead data using Clearbit: {{ $json.email }}"\r
      }\r
    },\r
    {\r
      "name": "Score Lead",\r
      "type": "@n8n/n8n-nodes-langchain.openAi",\r
      "parameters": {\r
        "resource": "text",\r
        "operation": "message",\r
        "modelId": "gpt-4o",\r
        "messages": { "values": [{ "content": "Score this lead 0-100..." }] }\r
      }\r
    }\r
  ]\r
}\r
```\r
\r
### Step 6 — ROI Calculator\r
\r
| Metric | Before Automation | After Automation | Savings |\r
|--------|------------------|-----------------|---------|\r
| Time per run | [X hours] | [Y minutes] | [Z%] |\r
| Runs per week | [N] | [N] | — |\r
| Total time saved/week | — | — | [hours] |\r
| Cost saved/month | — | — | [$$$] |\r
| Automation setup cost | — | — | [one-time] |\r
| **Payback period** | — | — | [weeks] |\r
\r
## Example Interactions\r
\r
**User:** "I spend 3 hours every Monday pulling sales data from 5 spreadsheets, writing a summary email, and updating our CRM. Can this be automated?"\r
\r
**Skill response:** Scores the workflow (42/50 — High priority), designs a 4-agent pipeline (data collector → analyzer → email writer → CRM updater), recommends n8n as the platform (self-hostable, native AI nodes), generates a complete n8n JSON spec, and estimates 11.5 hours/month saved = ~$580 value at $50/hr.\r
\r
---\r
\r
**User:** "I want to build a customer support triage system that reads emails, classifies them, and routes to the right team."\r
\r
**Skill response:** Designs a HITL-enabled pipeline with email reading, classification, confidence threshold (>85% auto-route, \x3C85% human review), CRM ticket creation, and Slack notification. Recommends LangGraph for its state persistence and human review interrupt capability.\r
\r
## Notes & Constraints\r
\r
- Always design **HITL checkpoints** for: financial decisions, customer communications, data deletions, external API calls with side effects\r
- For **regulated industries** (finance, healthcare, insurance): flag compliance requirements\r
- Workflows involving PII must include data retention and access control considerations\r
- Recommend starting with a **pilot workflow** (lowest risk, highest frequency) before scaling\r
- Provide rollback strategies: every agentic workflow should have a manual fallback\r