If you've been working through this course from Module 1, you've produced a real artefact, learnt three prompt structures, understood the failure modes, and built a working mental model of three different tools.
Here's the problem most beginners hit next: the outputs still feel generic.
You've asked good questions. You've used the right structures. But the answers keep sliding back to "it depends on your situation" or advice that could apply to any of the million people who type similar questions every day.
That's not the AI's fault. It's a context problem. The AI doesn't know who you are. And it's not going to figure that out across a session or two. You have to give it the context explicitly, deliberately, and in a form you can reuse.
This module is about doing exactly that.
The personal context block
The single most useful thing you can build today is a personal context block: a short paragraph (150 to 200 words) that you paste at the beginning of any conversation where personal career context matters.
Think of it as a briefing document for a new colleague who knows a lot about the world but nothing about you. Every time you start a relevant conversation, you brief them. No re-explaining. No generic answers.
Here's a template, with an anonymised example filled in:
Template:
Here's my context before we start:
I'm [brief description of current role/background, 1 to 2 sentences].
Before that I [relevant previous experience, 1 sentence].
I'm currently trying to [what you're working toward, 1 sentence].
My constraints: [time, financial, geographic, or other relevant limitations, 1 to 2 sentences].
What I want from you when I ask career-related questions: [specific guidance
on what good answers look like for you, e.g. "don't give me generic options,
give me specific recommendations given what I've told you" or "always tell me
when you're uncertain rather than being confident"].
Feel free to refer back to this context throughout our conversation.Anonymised example:
Here's my context before we start:
I'm a registered nurse with 8 years of clinical experience in acute care,
currently working part-time while caring for two young children. I have a
bachelor's degree in nursing and no formal business or management training.
I'm trying to move into healthcare administration or health system improvement
roles over the next 18 months, preferably without having to do a full master's
degree first.
My constraints: 10 to 12 hours per week available for study, no employer
sponsorship, and I'm based in regional Victoria so remote work is important.
When I ask you career-related questions: don't give me generic frameworks.
Give me specific recommendations that acknowledge my constraints. If you're
not sure about something Australian-specific, say so and tell me to verify.Paste this at the start of a session, then ask your actual question. The difference in output quality is significant.
Build your own version. Keep it in a notes app or document you can copy from quickly. Refine it as your situation changes.
Prompt libraries you actually reuse
You've already learnt three prompt structures. Now it's worth building a small personal library, five to ten prompts you know work, saved somewhere you can grab them.
These aren't generic tips from a blog post. They're prompts you've tested on your own situation and found useful.
Here are the five that tend to produce the most useful outputs for career-shifters. These are starting points; adapt them for your context.
Decision stress-test:
[Context block]
I'm considering [specific decision]. My reasons for doing it: [reasons].
My reasons against: [reasons]. The thing I'm least sure about: [specific uncertainty].
Challenge my reasoning. Don't just validate my thinking. Tell me what I might
be missing or what assumption I might be making that's worth examining.Learning checkpoint:
I've been studying [topic] for [time period]. Here's what I think I understand
about it:
[Your explanation in plain English]
Tell me: where is my understanding correct, where is it incomplete, and what
is the most important concept I haven't mentioned that someone with this knowledge
should understand?Explain this to me like I just arrived:
I'm new to [field/topic]. I have a background in [your background].
Explain [concept you're trying to understand] to me, starting from the
assumptions someone with my background would have, and building up from there.
Use an analogy where that helps. Don't over-explain the basics I likely know.Job ad gap analysis:
[Context block]
Here is a job ad I'm considering applying for:
[paste job ad]
Given my background, where are the strongest matches, where are the obvious
gaps, and what would I need to address or reframe in my application to
give myself a genuine chance?Skill priority ranker:
[Context block]
I have [X hours per week] for skill development. I'm targeting [role/field].
Given what I've told you about my background, rank these potential learning
investments in order of likely ROI for my specific situation: [list your options].
Be direct. Don't tell me "all of these are valuable." Tell me what you'd
prioritise if you were me, and why.Save these. Use them. Update them when you find they need refining.
Memory and saved instructions in ChatGPT
ChatGPT has a "memory" feature (in settings, can be turned on or off) that allows it to remember things you've told it across sessions. If you tell ChatGPT your job target, your background, your constraints, and what you're working on, and memory is enabled, it will carry that across future conversations. You won't need to re-paste your context block every time.
This is useful but worth understanding properly. A few notes:
First, you can view what ChatGPT has stored about you (Settings > Personalisation > Memory). Do this after a few sessions. It's illuminating, and occasionally incorrect. If ChatGPT has misremembered something, correct it explicitly: "That's wrong. Update your memory: [correct information]."
Second, memory is a convenience, not a substitute for a context block. Memory accumulates slowly and may not include the most relevant details until you've had many conversations. For a new or important conversation, paste the context block anyway.
Third, ChatGPT's memory stores information about you but not documents. It won't remember the contents of a long document you pasted in a previous session.
Claude Projects: ongoing context for a specific topic
Claude has a feature called Projects, available on Claude.ai. A Project is a saved conversation space where you can upload documents, set instructions, and return to the same context across multiple sessions.
For career-shifters using Claude for ongoing learning or job searching, this is genuinely useful.
You might set up a Project for "Career Pivot 2026" with:
- Your personal context block as a saved system instruction
- Your current CV uploaded as a reference document
- Notes on the field you're entering (key concepts, key players, key frameworks)
Every conversation in that Project starts from that context. You're not starting from zero every time.
How to set it up: in Claude.ai, click "New Project" in the left sidebar. Add a title. In the Project instructions, paste your context block. Upload any relevant documents. Start a conversation.
The decision lens pattern
This is a prompt structure worth highlighting specifically because it's so useful for career decisions.
The idea: give the AI your criteria, your options, and your context, and ask it to evaluate each option against the criteria. Not to make the decision for you. To make the comparison rigorous.
[Context block]
I'm deciding between these options:
Option A: [description]
Option B: [description]
Option C: [description, if relevant]
Here are the criteria I care about, roughly in order of importance:
1. [Criterion 1, e.g. financial sustainability in the next 18 months]
2. [Criterion 2, e.g. fits with family commitments]
3. [Criterion 3, e.g. likelihood of landing my target role within 2 years]
4. [Any others]
For each option, evaluate how well it performs against each criterion,
given what you know about my situation. Be direct about weaknesses.
If you need more information from me to give a useful assessment on a
specific criterion, ask.This doesn't replace your judgement. What it does is force clarity on what you actually value, and make the comparison visible rather than keeping it vague in your head. Often the act of filling in the criteria is itself clarifying.
What doesn't work
Treating memory as infallible. AI memory, whether in ChatGPT or Claude Projects, is useful but not perfect. Always re-read what's been stored about you and correct anything wrong.
Building a 500-word context block. Longer is not better. If your context block requires scrolling, it's too long. Keep it to 150 to 200 words. The discipline of keeping it short forces you to include only what actually changes the answer.
Using AI as the decision-maker. The decision lens pattern produces analysis. It doesn't produce decisions. You make the call. The AI helps you see the comparison clearly.