Beginners Guide to Claude (Save Time & Make Money)

Overview

The video is a beginner-focused guide to using Anthropic’s Claude to save time and improve ROI across multiple businesses (e-commerce, digital products, managing a software company, and more). It covers practical setup steps and a progression from basic usage to automation.

Key updates mentioned

• Opus 4.7 was released during recording, making upcoming workflows “more powerful.”

Core product features & how to use them

1) System Prompt (custom instructions)

• Set a custom system prompt to control Claude’s behavior in every chat.
• The creator recommends keeping it lean with only a few high-impact rules, such as:
  • Clarify before action: ask 3–4 clarifying questions before complex work.
  • Improve and suggest: request opportunities for automation, improvement, and repeatability while working.
  • Output preference: default outputs to Markdown unless another format is required.
• Optional note: prompt preferences can be downloaded from below the video.

2) Memory / capabilities

• In the Capabilities tab, Memory can be enabled so Claude remembers more details over time.
• The creator prefers excluding memory from general preferences, because memory is managed separately.

3) Interactive “UI” flow using system prompt

When starting a task (example: creating a roadmap for an online digital art business), Claude:

• Asks targeted questions to gather context
• Produces a higher-quality roadmap based on the user’s answers

Emphasis: a well-crafted system prompt improves both result quality and efficiency.

4) Project folders (organization + persistent context)

• A Projects feature groups chats, files, custom instructions, and memory together.
• Projects help Claude respond better because it can reference prior chats within that project.
• Creating a project triggers the system prompt to ask initial questions and generate files automatically.

Modes inside Claude (and project limitations)

Claude has three modes where files/instructions/memory can differ:

• Default chat mode
  • Projects work but are more limited.
• Co-work
  • Projects behave like a real location on your computer.
  • Requires the Claude app installed locally.
  • Can reference/control local folders (example: organizing ~207 files on the desktop into categories in real time).
• Code
  • Mentioned as robust and capable even for non-coders.

5) Artifacts (create “apps” and files, not just text)

• Artifacts are Claude outputs that are not plain text, including:
  • files (e.g., markdown)
  • PDF/app/game-like interactive tools, etc.
• In the creator’s “learn” project, Claude generates multiple markdown artifacts.
• Artifacts can become interactive, app-like tools inside Claude.

Examples:

• Wall art cropper tool: a single-page web app to crop designs into print-ready dimensions (avoids writing Photoshop scripts manually).
• Website component artifact: turn a reference image into a reusable pop-up component; outputs code that can be saved and reused.

Purpose: quick prototyping without coding a full solution from scratch.

6) Connectors (integrations that let Claude do actions in other apps)

Under Customize → Connectors, users can add pre-built integrations (often “one or two click”).

• Connectors act like “arms and legs” so Claude can act in external tools.

Examples:

• Create content/code in a Claude project → send to Notion (permission prompt: “always allow”).
• Gmail + Notion workflow:
  • analyze the last 15 emails
  • summarize them into a Notion document
• Canva connector:
  • authenticate Canva
  • generate design outputs inside the Claude workflow

7) Co-work schedules (automation)

In Co-work schedules, tasks can run on a recurring basis.

Examples:

• Weekly:
  • organize desktop/downloads files by type
  • log moves to a Notion change log
• Every 4 hours:
  • pull unread emails
  • draft responses according to policies
  • require periodic human approval only

Additional examples mentioned:

• scanning/unsubscribing marketing emails
• pulling analytics weekly
• analyzing spreadsheets for ideas
• compiling expense/P&L data

8) Skills (repeatable step-by-step instruction packs)

• Skills are instruction files that standardize complex workflows.
• They’re framed as solving the “AI consistency” problem using checklists/SOPs.

Examples:

• Policy auto-email skill:
  • load policy source of truth
  • pull unread emails (via Gmail connector)
  • classify
  • draft responses
  • save/summarize
• Product research skill for e-commerce: perform research actions that would take hours manually.
• Computer-control skill: automate building and publishing a product listing.
• JSON prompt builder skill: generate consistently formatted JSON prompts for downstream usage (example references an image tool like “Nano Banana”).

9) Claude Code for business workflows (sub-agents + parallel work)

• Claude Code is positioned as the most valuable/robust model for business use.
• Key emphasized capability: deploy sub-agents to run tasks in parallel.
• Claude Code can:
  • navigate many files
  • search the internet for tasks
  • use connectors

The creator uses it with tools like Anti-gravity and VS Code-style environments as text editors/file managers.

Example structure: a workspace with folders for digital product research, employees, finances/taxes, scripts, video editing, YouTube, and skills (e.g., thumbnail creation).

Note: Claude Code is described as similar to other Claude models but more “technical” in interface.

Main speakers / sources

• Primary speaker/source: the video creator (not identified by name in the subtitles) demonstrating Claude and Anthropic features.
• Company source referenced: Anthropic (release of Opus 4.7)