I keep seeing the same panic loop on LinkedIn. People doom-scrolling headlines about AI replacing jobs, then closing the tab and doing absolutely nothing different the next morning. The anxiety is real, but it’s also a terrible career strategy.
Then I stumbled on this incredible breakdown from an AI professional who flipped the whole conversation. Instead of fueling the fear, the original poster laid out 12 concrete actions any knowledge worker can take starting today. No fluff, no doom, just a checklist you can run through this week.
I was blown away by how practical this list is. So I’m breaking it down for you, step by step, with my own take on why each one matters.
Why this list hits different
Most AI advice falls into two buckets: vague philosophy (“embrace change”) or hyper-specific tool tutorials that age out in a month. The author skipped both and went straight for behaviors. Things you control. Things that compound.
The framing here is what makes it click: stop worrying about what AI will do to you, and start designing what you’ll do with AI.
The 12 actions, unpacked
1. The certifications on your profile.
- Head to anthropic.skilljar.com.
- Take Claude 101, AI Fluency, and Intro to Cowork.
- Add each one to LinkedIn under Add section, Licenses and certifications, Issuer: Anthropic.
- Visit claude101.com to actually master Claude.
Why this matters: signals are cheap to earn and they pre-qualify you in recruiter searches. Free certifications are low-hanging fruit most people ignore.
2. The tools you open daily.
- Pin Claude on your Chrome and open it every morning.
- Pick one daily task and do it in Claude for 30 days.
Daily reps beat weekend marathons. The expert is right: muscle memory is built by opening the same tab every day, not by binging tutorials.
3. How you prompt.
- Start every prompt with scope, format, length, and tone.
- Use action verbs only. Cut every “don’t do this.”
- End prompts with “Go beyond the basics.”
That last line is gold. I tested it on a few drafting tasks and the outputs genuinely level up.
4. What you automate.
- Inbox: connect the Gmail connector to draft replies.
- Notes: connect Granola to summarize meetings.
- Research: ask Claude to read 5 sources and give you a brief.
5. What you refuse to automate.
- The first 10 minutes of any problem. Think on paper first.
- Your final edit. Read every output out loud.
- Never let AI write messages that change relationships.
This is the action I respect most. The creator isn’t selling “automate everything.” They’re drawing a line around the human parts, and it’s the right line.
6. Your output per hour.
- Pick one repeating task and time yourself today.
- Redo it with Claude next month and time it again.
- Pin both numbers in your notes and repeat every 30 days.
Measure the lift or you’re guessing. Simple as that.
7. The skills you build in Claude.
- Anything you prompt twice deserves to be a skill.
- Customize, Skills, plus icon, upload, name it.
- Use a slash command like /47. Build 3 this week.
8. Your “about me” file.
- Create a .md file with your role, writing style, goals, and non-negotiables.
- Add your projects and what AI should never do.
- Paste it into every new chat and reuse forever.
This one alone will 10x your output quality. Stop re-explaining yourself to every fresh chat window.
9. What you ship publicly.
- One post or one doc per week. Pick the format you hate.
- Use AI to draft. Rewrite the first and last line yourself.
- Ship on a fixed day (the author picks Tuesday).
10. Who sees your work.
- 5 new LinkedIn connections per week.
- 1 DM per week to someone whose work you actually use.
- Tag one person in a post. Loops bring more reach.
11. What you read offline.
- 30 minutes a day with no phone in the room.
- Books, long essays, and your own old notes.
- Highlight by hand. The friction is the point.
I love that this contributor put offline reading on an AI productivity list. The best prompters I know are also the best readers. Inputs matter.
12. The questions you ask.
- Before prompting, write down the questions you’d ask a human expert.
- Add 2 follow-ups you’d ask after their answers.
- Better questions beat better prompts.
The pattern underneath all 12
Look at the list again and you’ll spot the theme. Every single action is something you can do this week without permission from your boss, your company, or the AI gods. No waiting, no excuses.
The mind behind this post built a framework that respects two truths at once: AI is changing everything, AND your daily choices still matter more than the model release schedule.
How I’d start if I were you
- Pick 3 actions from the list. Not 12. Three.
- Block 30 minutes tomorrow morning for the first one.
- Do it badly. Ship it anyway.
- Next week, add one more.
- Re-read the full list in 30 days and check what stuck.
Momentum compounds. One certification this week, one skill next week, one published post the week after. By month three you’re a different professional.
Go check the full LinkedIn post for the original phrasing and a few details I trimmed. Then close the tab and actually do one of them today.