Build a 90-day AI learning plan for your role
Generic “learn AI” advice rarely sticks because it ignores what you actually do all day. This builder takes your role, current level, goal, and weekly hours, then generates a structured 90-day roadmap across three phases — foundations, application, and mastery — with concrete milestones and practice prompts tied to your work.
Why role matters more than level
Most AI learning resources assume a generic learner, which means the practice exercises are either too abstract or targeted at a different use case. A lawyer and a software engineer at the same “beginner” level need to learn entirely different things:
The lawyer needs to understand how to prompt for legal research, how to verify AI-generated case summaries, what AI tools can and cannot reliably do with contracts, and what ethical and disclosure obligations apply in their jurisdiction.
The software engineer needs to understand how to use AI for code generation and review, how to wire LLMs into applications, how to evaluate model outputs programmatically, and how to manage prompt engineering at scale.
The roadmap builder surfaces role-specific milestones and practice tasks rather than generic “explore ChatGPT” exercises, which is what separates a plan that produces real skill change from one that produces interesting experiments for a week and then gets abandoned.
What the three phases build
Phase one: Foundations (weeks 1–4) — You learn to reliably produce useful output with AI tools for your role. The milestone is a set of 5–10 tested prompt templates you use every week. This phase is also where you learn what AI cannot do for your role and develop the habit of verification.
Phase two: Application (weeks 5–8) — You build at least one AI-assisted workflow that replaces or significantly accelerates a recurring manual task. The milestone is a documented workflow you could hand to a colleague. This is where most of the real productivity gain happens.
Phase three: Mastery (weeks 9–12) — You go broader and deeper: exploring automation (chaining AI tasks), integration (connecting AI tools to your existing systems), and sharing (teaching a colleague or writing up what you have learned). The milestone is teaching one other person in your team to use a skill you developed in phase two.
How it works
The roadmap is organised into three four-week phases. Foundations builds core fluency: prompting patterns, tool selection, and evaluating output. Application moves to your role’s real tasks, turning manual work into AI-assisted workflows. Mastery focuses on automation, integration, and teaching others. Your selected weekly hours scale how aggressive each phase’s milestones are, and your role swaps in domain-specific use cases and practice prompts.
Tips for following through
- Block the hours. A roadmap only works if the weekly time is on your calendar; be honest about how many hours you can sustain.
- Practise on real work. Apply each new skill to a live task that week — the practice prompts are starting points, not the goal.
- Ship something each phase. End foundations with a reusable prompt library, application with one automated workflow, and mastery by teaching a colleague.
- Re-assess at day 90. Pair this roadmap with a productivity self-assessment to confirm your skills actually moved.