Learn a New Skill: 90-Day Roadmap

Get a project-driven roadmap for a new skill, with a first output in week one, checkpoints that prove progress, and the plateau plan.

The Prompt — replace [BRACKETS] with your details

Act as a coach who has taught this skill to adults with limited time. Build me a 90-day roadmap.

The skill: [WHAT I WANT TO LEARN]

About me:
- Why I want it: [THE REAL REASON — job, project, curiosity, income]
- What "good enough" looks like for me: [THE CONCRETE END STATE — not "be good at it"]
- Current level: [HONESTLY, INCLUDING ADJACENT SKILLS I ALREADY HAVE]
- Time available: [HOURS PER WEEK, AND WHEN]
- How I learn best: [READING / VIDEO / DOING / TEACHING OTHERS]
- Budget for materials: [AMOUNT OR "free only"]
- What has stopped me before: [BE HONEST]

Deliver:
1. The 20% of this skill that produces most of the useful capability, and what to deliberately ignore for now
2. Week 1: something I build, make, or do by day 7 — however small. No week of theory first
3. Three phases across 90 days, each with a project that forces the skills of that phase, and a definition of done
4. A weekly rhythm: what a typical week looks like given my available hours
5. Checkpoints at day 30, 60, and 90: a specific thing I should be able to do, so I can tell progress from time spent
6. Resources: a small number of specific, high-quality ones with what each is for. Do not give me a list of twenty. Flag any you are not certain currently exists rather than inventing a title
7. The plateau: where people stall in this skill, roughly when it hits, and the specific way through
8. How to get feedback, since practising without it entrenches mistakes
9. Given what has stopped me before, the two structural changes most likely to prevent it happening again

Be realistic about what my stated hours can achieve in 90 days. If my target is not reachable, say so and reset it.

How to use this prompt

Verify every recommended resource exists before buying or committing time — models invent plausible book and course titles.
Do the week-one project even if you feel unprepared. That is the point of it.
Use the day-30 checkpoint honestly; adjusting the plan then is normal, abandoning it is not.

Why this prompt works

Self-directed learning fails at the same two points: a long theory phase with no output, and an unrecognised plateau. Anchoring each phase to a project makes progress observable, and naming the plateau in advance means hitting it reads as expected rather than as evidence you cannot do this.

Variations to try

Add "I have 3 hours a week, not 10" and see how honestly it rescopes the target.
Ask for the day-30 checkpoint as a specific test you can actually attempt.
Ask what to learn next once the 90 days are done, and what would be a distraction.

Common mistakes to avoid

Accepting invented book titles, course names, or URLs. Check each one.
Collecting resources instead of practising. Two good ones used beats twenty saved.
Setting a vague target, so you can never tell whether you have arrived.

Works well with

ChatGPT

Claude

Gemini

Need a custom version of this prompt?

The free prompt generator builds a prompt tailored to your exact goal, framework, and target AI model — or paste this template into the optimizer to refine it.

→ Open the Prompt Generator → Prompt Optimizer

Related prompts

Explain Any Concept Simply

Get any difficult concept explained in layers — from a one-line version to a deeper explan…

Personalized Study Plan Creator

Build a week-by-week study plan for any exam or skill, fitted to your schedule, with space…

Flashcard Generator

Turn any notes, article, or chapter into well-formed flashcards ready for Anki or any spac…

Socratic Tutor (Learn by Questions)

Instead of getting answers, get guided to them. This prompt makes the AI teach through que…