오픈클로? 자피어? 다 필요 없는 이유(클로드 하나로 끝내는 1인 기업 자동화)

Overview

The video argues that many “automation” tools (e.g., Google Opal, Zapier, and simple Python scripts) fall short because real work involves messy steps like searching, organizing, evaluating, formatting, and sharing information.

Instead, it positions Claude (via “Claude in Chrome” / a Claude browser coworking workflow) as the core automation method because it can read web pages in context and continue the workflow with judgment, not just execute commands.

The presenter demonstrates three practical automation workflows by combining:

• Claude in Chrome (real-time in-browser assistance, reading the currently opened page)
• Claude cowork / connectors to automate outputs into tools like Notion, Slack, and Google Drive

---

Key demos / tutorials (3 workflows)

1) Solo investor stock research → Notion “investigation cards” (≈10 minutes)

Problem: Gathering and organizing stock research candidates is exhausting; people abandon the process.

Workflow:

1. Use Claude in Chrome to gather scattered info by browsing needed pages/tabs.
2. Turn on “confirmation” before running (safety/visibility step).
3. Instead of summarizing official announcements, ask Claude to organize from an investigation-candidate perspective with a strict format:
   • Company title
   • Summary
   • What the company does (2 sentences)
   • Why you’re paying attention (framework)
   • Two risks
   • Sources / verify via the source URL if unclear
4. Claude selects ~5 candidates out of 10 and generates Notion cards with:
   • Title, source, URL, reference date
   • One-line summary
   • Why it matters, risks, and what to check next
5. Organize in Notion workflow stages: Investigation → Observation → Reunion.

Claimed benefit: turns stock info into a reusable “investment idea bank” rather than repeated manual work.

---

2) Legal contract review comments → issue table + negotiation memo (from comments)

Problem: Legal review time is consumed less by reading the contract and more by organizing comments/issues (often in Google Docs/Word).

Workflow:

1. Open the relevant agreement (Google Docs example).
2. Use Claude in Chrome to extract only the comments from the currently open document.
3. Convert the organized comments into a table of five key issues, including:
   • Counterparty demands
   • Our risks
   • Our position
   • Suggested conversation/phrasing (“counter-phrases”)
   • Confirmation questions to ask
4. Request generation of a contract negotiation memo document that includes:
   • Key conclusions
   • Five risks
   • Overall issue table
   • First paragraph draft of the counterparty reply email
   • Next-action checklist
5. Use Google Drive integration (button to upload/save).

Claimed benefit: AI handles the organization, making negotiations more stable and faster while humans keep judgment.

---

3) Marketer campaign request in Slack → mini tool + Notion card + Slack draft/send

Problem: Marketing campaign planning isn’t hard logically, but execution energy is drained by organizing incoming requests and producing deliverables (copy, UTM, checklists, etc.).

Workflow:

1. A campaign request arrives in a Slack “campaign request” channel (fields like name, goal, target audience, benefits, channels, budget, duration, landing page, URL).
2. Create an input app (HTML form approach described) that standardizes the request fields.
3. Tip: instead of manually entering data, Claude can pull/auto-fill from a posted message in the Slack request channel.
4. Press a button (e.g., “Create Preparations Automatically”) to generate deliverables at once:
   • 3-line summary
   • Ad copy
   • UTM examples
   • Phrase checklist
   • Content to save in Notion and (another “Seolle” tool mentioned)
5. Use Notion connector to create a new card in a campaign management sheet.
6. Generate a Slack team channel announcement as a draft first, then:
   • preview
   • send only if approved (confirmation step)

Claimed benefit: reduces exhaustion and rework as campaign volume grows; standardized pipeline with fewer missing details.

---

“3 automation principles” (tips from the presenter)

1. Don’t try to build perfect automation
   • Cover ~80% automatically; fix the rest manually faster than chasing edge cases.
2. Decide the output format first
   • Define a fixed structure (e.g., legal issue table columns). Consistent output becomes searchable/comparable over time.
3. Always include a confirmation step before transmission
   • Especially for sending Slack messages/emails/sharing documents to prevent accidental mistakes.

---

Main products / tools referenced

• Claude and “Claude in Chrome” (browser-based real-time page work)
• Claude cowork (workflow/assist mode)
• Notion (card creation + structured organization)
• Slack (campaign request intake + sharing/publishing)
• Google Docs (contract comments source)
• Google Drive (upload negotiation memo)
• Zapier and Python scripts (mentioned as less capable “auxiliary tools” compared to this approach)

---

Main speakers / sources

• Primary speaker: The video presenter (a single narrator claiming ~20 years of automation development experience)
• AI source used in demos: Claude (Claude in Chrome / Claude cowork workflow)