CHAPTERS
- 0:04 – 0:34
Claude as a collaborator and tour of the sidebar navigation
The video opens by positioning Claude as an intelligent partner that helps you break down complex projects and move from a blank page to an actionable plan. It then orients you to the Claude.ai sidebar and the main areas you’ll use to manage work.
- •Claude complements your expertise by helping structure and accelerate complex work
- •Sidebar basics: start new chats and return to previous conversations
- •Projects: organize conversations with persistent context and custom instructions
- •Artifacts: turn ideas into shareable apps, tools, or content
- 0:34 – 1:05
What makes a prompt effective: natural, concise, coworker-style communication
This section explains that Claude’s output depends heavily on your prompt and any added context. The recommended approach is to talk to Claude like a capable coworker—clear, concise, and conversational.
- •All interactions start with a prompt plus surrounding context
- •Prompt quality directly affects response quality
- •Use natural language and be concise rather than overly formal
- •Aim for collaborative instruction (as you would with a teammate)
- 1:05 – 1:35
Prompt framework: set the stage, define the task, specify rules
The video provides a three-part structure for crafting strong prompts. It emphasizes clarifying your role and goals, stating the exact action you want, and adding style/format constraints and examples.
- •Set the stage: role, objectives, and relevant background Claude should know
- •Define the task: write, analyze, build, etc., with clear deliverables
- •Specify rules: tone, style, format, and constraints
- •Attach examples when possible to anchor outputs
- 1:35 – 2:06
Worked example: investor pitch deck prompt with research and citations
A sample prompt is reviewed to show how the framework looks in practice. The example includes context (an indie streaming app pitch), a clear task, and strict output requirements like web research with citations and professional formatting.
- •Context + objective: investor pitch deck for a new indie streaming app
- •Task definition: provide relevant details and what to produce
- •Rules: require current web research with citations
- •Rules: specify professional report structure and format
- 2:06 – 2:36
Adding your context via uploads to tailor Claude’s responses
This chapter highlights the “power move” of providing Claude with your specific documents and background materials within a chat. Uploading relevant files helps Claude align its responses with your real constraints, data, and goals.
- •Upload company docs, project files, or research to deepen relevance
- •Claude uses the added chat context when generating responses
- •Treat uploads as a shortcut for shared understanding
- •Supports multiple file types (e.g., PDF, CSV, Docs)
- 2:36 – 3:06
Search & tools: expanding Claude’s context with web and connected sources
The video explains how enabling tools gives Claude access to additional information sources as needed. Options like web search and Google Drive connections can provide current or real-time data to improve results.
- •Use the search/tools menu to expand available context
- •Enable web search for up-to-date market or external information
- •Connect data sources (e.g., Google Drive) for real-time/internal info access
- •Claude selects the appropriate tool approach while solving requests
- 3:06 – 3:36
Choosing the right model: how to use the model selector
This segment shows where to find the model selector in the UI and how to switch models. It notes that changing models starts a new chat and that different models are tuned for different needs.
- •Model name appears under the text input in a chat
- •Click the model name to switch to a different Claude model
- •Switching models results in a new chat
- •Models have distinct capabilities aimed at different use cases
- 3:36 – 4:07
Model guidance: Opus for complexity, Sonnet for everyday work
Here the video offers practical guidance on matching the model to the task. Opus is positioned for the most complex, multi-step work, while Sonnet is the recommended default balancing capability and cost effectiveness.
- •Claude Opus: largest-scale model with hybrid reasoning for complex tasks
- •Example use case: multi-step financial analysis
- •Claude Sonnet: best for everyday use and recommended default
- •Choose the tool (model) based on what you’re trying to accomplish
- 4:07 – 4:37
Extended thinking mode and Research: deeper reasoning and cited investigations
This chapter introduces two advanced capabilities: extended thinking mode for harder reasoning (with higher latency) and Research for comprehensive, multi-source investigations. Research automatically decomposes complex questions and delivers cited reports over a longer time window.
- •Extended thinking mode helps with complex analysis but increases latency
- •For quick tasks, use Sonnet with thinking mode off
- •Research mode: systematic, multi-angle investigation across many sources
- •Delivers comprehensive cited reports in ~5–45 minutes
- 4:37 – 5:18
Getting better results over time through ongoing collaboration
The closing emphasizes that Claude is most powerful with continued back-and-forth, not one-off prompts. Frequent interaction helps Claude adapt to your communication preferences and become a stronger day-to-day collaborator.
- •Best results come from iterative conversation, not single prompts
- •Ongoing communication helps Claude learn your style and preferences
- •Claude amplifies your capabilities across work when paired with your expertise
- •Reinforces the partnership: AI intelligence + your context and judgment
