AI Starter Kit
Your first 7 days with AI
A plan for busy professionals who've been meaning to start.
Seven days of short, practical steps that build on each other. Fifteen to thirty minutes a day. By the end you'll have a working habit, a prompt sheet you'll use, and a real sense of what AI can and can't do.
By Nicolle Weeks · Updated April 2026
Before we start
This is a free 7-day plan for non-technical professionals who want to start using AI at work. Fifteen to thirty minutes a day. No technical background required.
If you've been lurking on AI content for months (saving posts, watching demos, feeling vaguely behind), this is for you.
You don't need to become an “AI person.” You don't need to understand how it's built. You need to know how to use it well enough that it helps you, and how to question it when it shouldn't be trusted.
One rule
Do the task each day. Reading about AI is not the same as using it. This plan only works if you open a tool and try something.
The 7-day plan
The key content for each day is shown below. Expand any day for the full explanation and task.
Today's task (5 minutes)
Write down the one thing you've been avoiding doing because you didn't know where to start with AI. That's your north star for the week. By Day 6, you'll do it.
Read more
The panic is manufactured. Most of what you've seen about AI is designed to make you feel like you're already losing: by people who want to sell you something, or by media that benefits from your anxiety.
Here's what the data shows: most professionals, including the ones who sound confident about AI in meetings, are figuring this out as they go. The gap between where you are and where you need to be is smaller than you think.
What AI does well
- ✓Drafting, summarising, rewriting text
- ✓Explaining complex topics plainly
- ✓Brainstorming and generating options
- ✓Reframing problems from new angles
- ✓Helping you think and write faster
Where AI falls short
- ✗Current facts (training data has a cutoff)
- ✗Specific citations: always verify
- ✗Nuanced judgement calls
- ✗Knowing when to push back on you
- ✗Anything requiring real-world verification
Read more
The AI tools you're probably using (ChatGPT, Claude, Gemini, Copilot) are language models. They were trained on enormous amounts of text, and they learned to predict what words come next.
This is fundamentally different from a search engine. A search engine retrieves existing pages. A language model generates a response from scratch, word by word. It is not looking anything up. It is predicting.
Today's task (15 minutes)
Ask an AI tool to explain something you've always been curious about (from your field, or completely unrelated). Notice how it responds. Does it sound confident? Does the explanation hold up? This is your first real feel for how these tools behave.
Want to go deeper? Read the full curriculum →
ChatGPT
- Most widely used; easiest to find help with
- Web browsing built into paid tier
- Strong for research and analysis
- Good image generation (DALL·E)
Claude
What I use
- Notably better at long, nuanced writing
- More likely to push back thoughtfully
- Excellent for editing and matching your tone
- Large context window; paste whole documents
Read more
The two tools worth your time right now are ChatGPT (from OpenAI) and Claude (from Anthropic). Both have free tiers. Both are genuinely good. You don't need to master either. Pick one and use it.
Today's task (20 minutes)
Open one tool. Find something real from your workday (an email to write, a document to summarise, a meeting to prepare for). Use the tool to help. Don't overthink the prompt. Describe what you need in plain language. Edit the output. Notice what works and what doesn't.
Role
Who should AI be? “You are a sharp executive editor…”
Context
What does it need to know? “I’m preparing for a board meeting…”
Task
What specifically do you want? “Write a 150-word summary…”
Constraints
What to avoid or include? “No jargon. Lead with the key finding.”
Read more
The quality of your AI output is almost entirely determined by the quality of your input. Vague prompts produce vague output. Specific prompts produce useful output.
✗ Weak prompt
“Write a summary of this document.”
✓ Strong prompt
“You're a sharp executive summariser. Write a 150-word summary for a non-technical audience. Lead with the most important finding. No jargon.”
Today's task (20 minutes)
Take something you wrote this week and ask AI to rewrite it using a specific role and constraint. Then try a second version with a different tone or length. Compare all three. Notice how much the specificity of your prompt changes the result.
Read this next
| Mode | When to use it |
|---|---|
| Acceleration | Speed up the parts of your job that drain your energy: drafting, summarising, reformatting. |
| Exploration | Brainstorming, finding angles you hadn't considered, stress-testing your assumptions. |
| Reflection | Processing your own thinking. Use AI to clarify, not confirm what you already believe. |
| Delegation | Repetitive, low-stakes work. Automate what genuinely doesn't need you. |
Read more
The most powerful shift is moving from “I tried AI for a random thing” to “I used AI for something that matters at work.” Today's the day.
Today's task (20–30 minutes)
Use this prompt, adapted to your role:
“I'm a [your role] and my main responsibilities include [2–3 things]. I've used AI a little for [what you've tried]. Acting as an AI productivity coach, give me a SWOT analysis of my AI usage and suggest 5 specific ways I could use AI to save time or improve my work. Be concrete and practical. No generic advice.”
Is this verifiable?
Can you check the claim against an independent source?
Is this time-sensitive?
AI training data has a cutoff; it may not know about recent events.
Does this need a human in the loop?
Real names, legal claims, health decisions, sensitive judgement calls.
Does something feel off?
Trust your instincts. AI can produce confident nonsense.
Read more
AI sounds confident even when it's wrong. This is a consequence of how these systems work, not a glitch. They generate plausible text, not verified facts. This is called hallucination, and it happens to everyone.
Key point
Treat every AI response as a first draft that needs your judgement, not a final answer.
Today's task (15 minutes)
Ask AI something you know well (a topic from your expertise). Read its response carefully. Find one thing that's imprecise, overstated, or wrong. This exercise builds the critical eye you need to use AI well for the long term.
A Monday morning habit (5 minutes)
- →What's the most draining task I have this week? Can AI help with any part of it?
- →What am I writing from scratch that I could get a first draft for?
- →Is there anything I need to understand better before a meeting or decision?
Read more
You've done the hardest part: getting started. The question is how to keep going without making AI another thing that falls off your list after two weeks.
The professionals who get the most out of AI aren't the ones who use it the most. They're the ones who use it intentionally, for the right tasks, at the right moments.
This weekend's task
Go back to what you wrote on Day 0: the one thing you'd been avoiding. Use everything you've learned this week and do it. It won't be perfect. That's fine. The point is to stop waiting.
Keep this forever
10 prompts for every knowledge worker
These work regardless of your role, your industry, or which tool you use. Fill in the brackets with your specifics.
Drowning in email
"Summarise this email thread. Tell me what needs a response, what I can ignore, and what I should flag."
Preparing for a meeting
"I have a meeting with [person/team] about [topic]. What are the most important questions I should ask and points I should make?"
Procrastinating on a piece of writing
"Write a rough first draft of [thing]. It doesn't have to be perfect. I'll edit it. Key info: [paste your notes]."
Getting pushback on your idea
"Give me the five strongest objections to this proposal and a one-sentence response to each: [paste proposal]."
Something is too long
"Rewrite this to be 30% shorter without losing the core point or changing my voice: [paste text]."
You don't understand something
"Explain [concept] as if I'm smart but completely new to this. Give me an analogy from everyday life."
Stuck on a decision
"Help me map the trade-offs between [Option A] and [Option B] for someone in my situation: [brief context]."
Turning bullets into writing
"Turn these bullet points into a clear, professional paragraph in a conversational tone: [paste bullets]."
Need a second opinion
"Play devil's advocate on my thinking here. What am I missing or getting wrong? [paste your thinking]."
Getting started on anything
"I need to [task]. I don't know where to start. Give me the first three steps."
One rule: Always read the output before you use it. You are the editor. AI gives you the first draft.
Keep reading
Curated from the Human+AI archive. Pick the one that matches where you are right now.
If you want the basics
A Human+AI guide to common AI terms
The vocabulary you need to sound current in any AI conversation.
If you're wondering which tool to use
I switched from ChatGPT to Claude. Here's how I did it.
A practical walkthrough, including how to bring your chat history with you.
If you're worried about your job
AI job disruption: 5 steps to future-proof your career
3 lies and 5 things you can do.
If you want to use it at work, now
Build your AI-first workflow and pull ahead at work
The advantage goes to people who are getting good at using it.
If you're worried about overreliance
Will AI make us stupid?
An honest look at what overreliance does to our thinking.
If you have kids
A practical AI literacy guide for parents
Research-based. Written for the real questions parents are asking.
Keep learning with Human+AI
Keep learning with Human+AIOne email a week. Plain-English takes on AI for non-technical professionals. No hype, no doom, no jargon. The next step after this plan.