AI ChatGPT Crash Course

How It Works

Tell the course your main goals (writing, coding, research, customer work, personal productivity) and your current comfort level. It builds a focused curriculum just for you: the newest models and which one to use when, the 5 prompting patterns that actually move the needle, concrete prompts for your exact domain, common traps to dodge, and a simple daily practice schedule: so in a few weeks you're not just using ChatGPT, you're the one explaining it to everyone else.

What to Provide

ChatGPT Crash Course

Replace [[tokens]] with your details. This is the finished deliverable.

1. Account and Model Overview (2026)

Here's what's new so you're never the one quoting outdated info: every ChatGPT model today is part of the GPT-5 family, and the model picker is now a simple speed/reasoning dial instead of a wall of confusing model names.

Plans and pricing:

• Free: GPT-5 Instant only, capped at 10 messages per 5 hours, no GPT-5.5 access.
• Go (~$8/mo): Higher limits than Free, still Instant-tier reasoning, budget option in supported regions.
• Plus (~$20/mo): Full GPT-5.5 access, Advanced Voice, deeper memory, Projects, custom GPTs, higher limits.
• Pro ($100–$200/mo): 5–20x Plus quotas, extended Deep Research sessions, GPT-5.5 Pro, large-context access on the $200 tier.
• Business (~$25–30/user/mo): Admin controls, shared workspaces, higher privacy, SCIM, audit logs.
• Enterprise: Custom-negotiated, typically ~$60/user/mo with seat minimums.

When to use which reasoning level (simple rule):

• Everyday questions, drafting, summarization: Instant (fast, cheap, good for 80% of work).
• Hard reasoning, math, complex planning, debugging tough code: Medium / High / Extra High (slower, more thorough: pick higher only when the task needs it).
• Highest-stakes or longest-context work (Pro plans): Pro Standard / Pro Extended.

Recommendation for you: [[Start with Plus if you use it >5 hours/week. Use Instant for 80% of work, step up to Medium/High for the hard 20%.]]

2. Core Features You Should Enable Immediately

Custom instructions (most important lever):

Go to Settings → Custom instructions. Fill:

• What you do for work / study: [[Your role and industry]]
• How you want responses: [[Concise, use bullet points, always give examples, ask clarifying questions when needed]]

Projects (Plus/Team):

Create one project per major area ([[Marketing, Coding side projects, Personal finance, Research]]). Upload relevant files once; chat inside the project stays grounded in those files.

Memory:

Review and edit what it remembers about you. Add key facts: [[Your company size, tools, writing style preferences, common constraints]].

Advanced Voice (mobile):

Great for thinking out loud, language practice, or walking meetings. Use for brainstorming when typing feels slow.

3. Prompting Fundamentals That Actually Move the Needle

Pattern 1: Role + Goal + Constraints + Format

"You are a senior [[role]]. Your goal is to [[outcome]]. Constraints: [[time, budget, tone, must include X, never do Y]]. Output in [[markdown table / numbered steps / email format / JSON]]."

Pattern 2: Give examples (few-shot)

Show 1–2 good examples before asking for new work. The model matches style and structure.

Pattern 3: Chain of thought + verify

"Ask me any clarifying questions first. Then think step by step. At the end, list assumptions you made and suggest how to validate them."

Pattern 4: Iterate explicitly

After first output: "Improve the previous version. Make it 30% shorter. Add one concrete example for [[specific audience]]."

Pattern 5: Use delimiters and structure

Use ``` or XML-style tags to separate instructions, context, and the thing to transform.

4. Killer Use Cases with Ready Prompts (Adapt These)

Use case A: [[Email and message drafting]]

Prompt:

"You are my [[role]]. Write a [[tone]] email to [[recipient type]].

Situation: [[paste context]]

Goal: [[what you want them to do]]

Constraints: Keep under [[150 words]], start with the ask, one clear CTA with deadline.

Subject line first, then body."

Use case B: [[Research and synthesis]]

Prompt:

"Research [[topic]] for [[audience and purpose]]. Use only reputable recent sources. Structure: Key facts (bullet), Main debates (2 sides with sources), Practical implications for [[your situation]], Open questions. Cite sources inline with dates."

Use case C: [[Code help: explain or write]]

Prompt:

"Language: [[Python / TS etc.]]

Task: [[what to build or fix]]

Constraints: Use [[version or library rules]], include comments, handle [[edge cases]].

First give a short plan, then the code in a single block."

Use case D: [[Planning and decision making]]

Prompt:

"I need to [[decide or plan]]. Options I'm considering: [[list]].

Criteria that matter to me: [[cost, time, risk, learning, alignment with X]].

For each option give: Pros, Cons, Risk level, One concrete next step, Overall score 1–10 with justification.

End with your recommended path and why."

Use case E: [[Content and outlines]]

Prompt:

"Create a [[blog post / script / LinkedIn carousel / newsletter]] outline for [[audience]].

Angle: [[specific tension or promise]]

Length target: [[X words / slides]]

Must include: [[3 data points or stories, CTA at end]]

Tone: [[conversational expert]]

Deliver: Title options (3), outline with H2s, key points per section, suggested visuals or hooks."

5. Safety, Accuracy, and Professional Use Rules

Never do these:

• Paste confidential or customer PII without anonymizing or using an enterprise plan with data controls.
• Treat model output as fact without checking sources for YMYL (health, finance, legal).
• Submit AI-generated school or client work without disclosure if required.
• Rely on it for real-time info (stock prices, breaking news) without verification.

Accuracy habits:

1. Ask for sources.
2. Cross-check 1–2 key claims.
3. When it cites, open the link or search the claim.
4. For numbers: ask it to show the calculation or source table.

Professional disclosure (when appropriate):

"This draft was created with AI assistance and reviewed by me."

6. Common Traps and How to Escape Them

• Vague prompts → generic answers. Fix: add role, audience, constraints, format.
• One-shot and accept → mediocre. Fix: always do 1–2 iterations with specific feedback ("make it punchier", "add data").
• Forgetting context: long chats drift. Fix: start new chat or use Projects.
• Over-trusting on specialized topics. Fix: treat as smart junior colleague; verify.
• Prompting in natural language only for code. Fix: give the constraints and desired interface first.

7. 30-Day Practice Plan

Week 1: Foundations (15–25 min/day)

• Day 1–2: Set up custom instructions + one Project for your main use case.
• Day 3–5: Practice Pattern 1 + 2 on 3 real tasks from this week.
• Day 6–7: Review 2 outputs you liked; save the prompts as templates.

Week 2: Your Top Use Cases

• Pick 2 use cases from section 4. Run 4–5 examples each.
• Create 3 reusable prompt templates in your notes or custom GPT.

Week 3: Advanced Patterns

• Bump the reasoning level to Medium/High on one hard problem.
• Try Advanced Voice for thinking or language.
• Practice "explain this to me like I'm [[level]]" on something you use daily.

Week 4: Workflow Integration

• Put AI inside your actual tools (Notion AI, Google Docs, Cursor, Zapier).
• Build one small custom GPT or project that contains your style + examples.
• Track time saved on 5 tasks; decide where to double down.

Ongoing: Once a week, review one "best" and one "worst" output from the week. Update your custom instructions.

Worked Examples (Tailored)

Beginner: Marketing Coordinator

Use cases: Social posts, email subject lines, competitor notes.

First 3 prompts: Hero + CTA social, subject line + preview, 5-bullet competitor summary.

Result after 2 weeks: 40 min/day saved on first drafts; manager now asks for AI-assisted versions.

Intermediate: Software Engineer

Use cases: Code explanation, test ideas, refactor plans, commit messages.

Patterns: Always "plan then code", ask for edge cases, "rewrite this for clarity".

Result: Uses High reasoning for architecture decisions; Instant for daily refactors.

Power User: Solo Consultant

Use cases: Proposal outlines, client workshop design, research packs, pricing logic.

Advanced: Heavy use of Projects with past proposals + client notes; custom GPT for "my voice".

Result: Proposal time cut from 4 hours to 45 min first draft.

Format Checklist

Quick Reference Card (Save This)

• Role + goal + constraints + format = strong base prompt.
• Always iterate once or twice.
• Use Projects for ongoing work.
• Verify anything that matters.
• Save your best prompts.

Start the 30-day plan today. In a month you will be in the top 10% of ChatGPT users in your field.