Every generation thinks its moment of change is unique.
It's not. What is unique is who pays attention early, who ignores it, and who actually does something about it.
Right now, it's AI. Before that, it was cloud, mobile, the internet, the personal computer. Go back further and the pattern is even clearer.
When railroads and electricity spread, the big winners weren't only the companies that owned them. It was the merchant who expanded because shipping got faster. The factory that redesigned how work got done. When cars arrived, they didn't just create car companies. They created suburbs, logistics networks and whole new kinds of jobs.
In the personal computer era, you didn't need to be Microsoft. You just needed to learn Excel before your peers, or use email before others did. At one point, "good at PowerPoint and Excel" was a competitive edge.
Every time, the people who moved ahead asked one question: "What becomes possible because this exists?"
If you're early in your career, this is good news. You don't need seniority to get ahead on a new tool. You need a habit. This module gives you one, plus a way to use AI without handing over your judgment.
The short version
- Tech shifts follow a pattern. The winners aren't always the inventors. They're the people who learn early and apply it to real problems.
- Build a weekly practice: one hour, one real task, done with AI, checked and logged.
- AI amplifies learning. It doesn't replace it. Without understanding, you just produce wrong work faster.
- Give AI a job description. Decide what it does on its own, what it drafts for you, and what it never touches.
- Your judgment is the product. As the doing gets cheaper, thinking clearly gets more valuable.
The pattern, and where you fit in it
Most people react to new technology in one of three ways. They ignore it until their company forces it. They dabble, try a few prompts, get a mediocre answer and decide it's overhyped. Or they build it into how they work, a little every week.
The gap between the second and third group is where careers change. Access isn't the advantage. Everyone has access. The advantage is how quickly you learn the tools, how deeply you understand the problem, and how well you apply one to the other.
Key idea: The real skill isn't "learning AI." It's building the muscle of seeing change early, leaning in, learning fast and applying it to something real. There will be another shift after this one.
That muscle is what you're really training. AI is just this decade's version of the reps.
A weekly AI practice habit
You don't need a course or a weekend retreat to start. You need a recurring hour on your calendar and a real task to point it at.
Pick a real task. Not a toy prompt. A meeting summary, a messy spreadsheet, a first draft of a process doc, a customer email you struggled with. Real work teaches you where the tool is strong and where it breaks.
Try it with AI. Give it context the way you'd brief a new coworker: who it's for, what good looks like, what to avoid. "Make a summary" gets you slop. "Summarize these notes for my manager in five bullets, decisions first, then owners and dates" gets you something useful.
Check it. Read it like your name is on it, because it is. What's wrong? What's missing? What did it confidently make up?
Log it. Keep a running doc: the task, the prompt, what worked, time saved. Good prompts become templates you reuse.
Share one win. Show a teammate. This is how you become the person people ask, which is a visibility win on its own.
That's more hands-on practice than most of your peers will get, and you can start Monday.
Two habits help the hour stick. Block it on your calendar like a meeting. And once a quarter, try something bigger: one workflow you hate, a slow process or an unsolved problem, and build something that works. Not perfectly. Just something real.
Use AI without outsourcing your judgment
There's already a pattern emerging. An investor told me he's tired of AI slop landing in his inbox. He can tell when something is AI-generated, because it doesn't make sense.
He's right. You can fake it for a bit. String together prompts, generate something that looks polished. But without context, judgment and real understanding, it falls apart quickly. And the people who matter can tell.
Give your AI a rulebook
I run my business with a team of AI assistants. My executive assistant, Jill, is Claude. She doesn't work because the AI is clever. She works because she has a rulebook, one file with three headings:
- What she does on her own, like confirming meetings and triaging newsletters.
- What she drafts and I decide, like first replies to anyone new, or anything involving pricing or commitments.
- What she never does, like making payments or committing to contract terms.
It's a job description written clearly enough that software can follow it. You can do the same thing for how you use AI at work, even if you're two years in and using a chat window. The rulebook keeps you fast where speed is safe and careful where it isn't.
Check your company's AI policy first. Many companies have approved tools and clear rules about customer or employee data. Following them is part of using AI well.
What it looks like in practice
A composite example, not a real person.
Nadia is a recruiting coordinator, a year and a half in. Her company rolled out an AI assistant six months ago. Most of her team tried it once and went back to the old way.
She blocks Friday 3–4 p.m. Week one, she uses it to turn her messy interview-loop notes into a scheduling summary. It invents an interviewer who doesn't exist. She logs that: always paste the full list. Week three, she builds a reusable prompt for candidate follow-up emails that she edits before sending. Week six, she uses it to compare three job descriptions and spot inconsistent requirements.
She writes a three-part rulebook. On its own: formatting, summarizing her own notes. Drafts, she decides: every email to a candidate. Never: anything with salary details or candidate feedback pasted into an unapproved tool.
By month three, she's saving a few hours a week, and her manager asks her to show the team how she does it. Same tool everyone else had. Different habit.
Common traps
- Waiting for training. Your company may never run it, or run it late. Practise on your own time with real tasks.
- Dabbling, then quitting. One bad answer isn't a verdict. Better context usually gets better output.
- Sending the first draft. That's how AI slop happens. Edit until it sounds like you and every fact checks out.
- Ignoring data rules. Pasting customer or employee data into the wrong tool is a fast way to lose trust. Check the policy.
- Learning the tool, not the work. AI amplifies what you understand. Keep learning your field, or you'll just be wrong faster.
Do this this week
Start the habit.
- Block one hour on your calendar, same time every week, labelled "AI practice."
- Pick one real task from this week and run it through the loop: try, check, log.
- Write your three-heading rulebook using the tool below. Five lines under each.
- Share one thing that worked with a teammate.
Your tool
MY AI PRACTICE + RULEBOOK
WEEKLY PRACTICE LOG
Date | Task (real work) | Prompt I used | What was wrong | Time saved | Reuse? Y/N
_____|__________________|_______________|________________|____________|__________
GOOD BRIEF TEMPLATE
"You're helping me with [task]. It's for [audience].
Good looks like [format, length, tone].
Here's the context: [paste].
Avoid [things you don't want]. Ask me questions if anything is unclear."
MY AI RULEBOOK
Does it on its own (low risk, I skim):
1.
2.
3.
4.
5.
Drafts it, I decide (I edit and own it):
1.
2.
3.
4.
5.
Never does:
1. Handle data my company hasn't approved for this tool
2. Send anything in my name without my review
3.
4.
5.
QUARTERLY BUILD
The workflow I hate / process that's slow / problem I haven't solved:
What I'll build with AI:
Done when:
THE QUESTION TO KEEP ASKING
"What becomes possible in my job because this exists?"
Talk about it
- Which part of your job would someone from ten years ago find unrecognizable? What might look that way ten years from now?
- Where have you seen AI slop at work? What made it obvious?
- If you wrote a rulebook for AI in your role, what's the one thing that belongs under "never does"?
— Evan