You clicked the audit button. That puts you in a specific category: you're not just reading about this — you want to actually do something with it. So I'm going to skip the recap and get straight to the toolkit. (If someone forwarded this to you, here's the original newsletter it came from.)
The newsletter was about naming the problem. This email is about what you do Monday morning.
Everything below is copy-paste ready. The full prompt, the follow-up questions, an example of what the output looks like when it works, and a 3-step plan for this week. Under 15 minutes of reading. Under 30 minutes of action.
Let's go.
🛠️ The Complete Job Audit Prompt
Three-step Job Audit framework: run the prompt, sharpen with follow-ups, find one real posting.
STEP 1 — The Opening Prompt
Copy this exactly. Fill in your background where indicated.
Paste into Claude or ChatGPT:
Act as a labor economist specializing in AI’s impact on white-collar work. I’m going to describe my background and current role. I want you to do three things:
1. Identify the specific tasks in my work that AI tools are replacing most quickly — be concrete, not general.
2. Identify the specific skills, judgment calls, and knowledge I’ve built that are genuinely difficult to replicate — and explain why each one is hard to automate.
3. Suggest two or three specific directions where my existing experience becomes more valuable when AI handles the routine layer underneath it. These can be roles, project types, or ways to reposition in a job search.
Here’s my background:
[Your role or target role. Years in the field or education equivalent. What you actually do day-to-day — be specific. Any domain-specific knowledge you’ve built up. Any AI tools you currently use or have been trained on.](Don't worry about the "labor economist" part — it just tells the AI how to think. You describe your job in plain English.)
What to put in your background section (for best results):
Don't just write your job title. The prompt performs much better when you include specifics: "I've spent two years doing X, which means I know how to Y. I've worked with clients in Z industry. The part of my job that takes the most judgment is..." The more concrete you are, the more concrete the output.
STEP 2 — Follow-Up Questions
After you get your initial output, run these in the same conversation. Each one sharpens the output from Step 1.
Follow-up 1:
Take the top two judgment-heavy skills you identified. For each one, suggest a specific way I could position that skill in a job application, freelance pitch, or LinkedIn summary — assuming AI is handling the routine work underneath it. Give me language I could actually use.
Follow-up 2:
Based on my background, what are two or three specific roles or project types where my domain knowledge would be actively needed to catch, correct, or direct AI output — not just use it? I want roles where a human who knows this field deeply is harder to remove, not easier.
Follow-up 3:
Which skills or experiences I mentioned are becoming more valuable as AI handles more routine work in my field — and which are becoming less valuable? Be direct. If something I listed is on its way out, tell me.
Follow-up 4 (optional — use if your output felt generic):
The output you gave me felt too broad. I want you to get more specific. Pretend you're advising someone who works at [name a company type or sector you know well]. What changes?⚡ ⚡ ⚡
📋 Real Example
What this looks like in practice — Jordan's actual output
Jordan's condensed audit: tasks being replaced vs. hard-to-automate skills. Suggested directions: AI content editor, analytics + writing hybrid.
Jordan ran the prompt with this background section:
"I have a marketing degree from Ohio State, graduated May 2024. I've been job searching for eight months targeting content strategy and marketing analytics roles. I interned at a regional real estate firm for four months — I wrote property listings, ran their Instagram, and helped set up a basic email newsletter. I know Google Analytics at a basic level. I don't use AI tools regularly but I've experimented with ChatGPT for draft copy. I'm targeting entry-level content or analytics roles at mid-size companies."Here's what came back (condensed):
Tasks being replaced quickly:
First-draft copy (listings, social captions, basic emails) — this is being compressed by AI at most companies; the ask is shifting to editing and direction, not origination
Basic data pulls from Google Analytics — automation handles routine reporting; interpretation and recommendation are where human value sits
Judgment areas that are harder to automate:
Local market context — four months in a regional real estate firm means Jordan has seen how a specific market talks about itself, what words resonate locally, what photos move faster. A model trained on national averages doesn't have this.
Newsletter architecture decisions — she built a newsletter from zero. That's not just execution. It's a sequence of judgment calls: what to include, what cadence to set, what to cut. That process is difficult to fully delegate to AI.
Editing AI output for brand voice — companies are discovering their AI-generated content sounds identical to everyone else's. Someone who knows how a brand actually talks is becoming a distinct hire.
Two directions suggested:
AI content editor / brand voice specialist — companies are realizing they need humans to make AI output sound like them. Jordan's internship means she has a concrete example of building brand voice from the ground up.
Marketing analytics roles with a writing angle — the combination of basic analytics literacy + copy experience is rarer than either alone. Positions that require someone who can pull data AND write the story from it are less automated than pure data-pull roles.
The output wasn't perfect. The model initially suggested "prompt engineering" as a third direction — Jordan asked Follow-up 3, which got the model to walk that back and acknowledge it was a generic suggestion. That's why the follow-up questions matter. The first output is a draft. The follow-ups sharpen it.
⚡ ⚡ ⚡
🗂️ Industry Variations
Tailor it to your field
Marketing / Content roles (Jordan's world)
In your background section, emphasize: which brands or clients you've worked with, what the content was trying to do (sell, retain, educate), and any experience with editorial judgment — what you decided not to publish, and why. The follow-up question to add:
"Which of my content or strategy decisions required the most local or audience-specific context — the kind of knowledge that doesn't transfer easily to a general-purpose model?"Operations / Admin / Office management roles
In your background section, emphasize: the specific systems you've managed, how you handle exceptions and edge cases, and any experience coordinating between departments or external vendors where relationships and trust mattered. The follow-up question to add:
"Which parts of my operations work required me to make judgment calls that weren't in any manual — where the right answer depended on context I'd built up over time?"Finance / Accounting roles
In your background section, emphasize: the types of decisions you supported (not just the tasks you ran), any experience explaining financial data to non-finance stakeholders, and situations where you caught something a model or template would have missed. The follow-up question to add:
"Based on my background, what aspects of my financial work require ongoing human oversight because the consequences of error are high — and how do I position that as a strength in my job search?"🚩 🚩 🚩
⚡ Your Action Plan
Three steps. This week. Under 30 minutes total.
Step 1 — 10 minutes
Run the prompt tonight
Open Claude or ChatGPT, paste the opening prompt with your specific background. Don't overthink the background section — write it like you're explaining your experience to a smart friend.
Step 2 — 10 minutes
Run Follow-up 3 and write down what comes back
Ask the model which of your skills are becoming more valuable and which are becoming less. Write two sentences — in your own words — about your top judgment-heavy skill. This is the language you'll use in cover letters and conversations.
Step 3 — 10 minutes
Find one job posting this week where that skill is explicitly mentioned
Not "apply to it." Just find it. Search for roles that describe the thing the model identified as your judgment-heavy skill — and look at how the posting describes it. That's the vocabulary to start using about yourself.
📖 The one thing worth reading
Aneesh Raman's Fast Company piece: "The Bottom Rung of the Career Ladder Is Breaking" — Aneesh Raman, Fast Company
Raman runs workforce development at LinkedIn and has more platform data than almost anyone. The piece is short, specific, and doesn't pretend this is simple. Worth 5 minutes.
If this was useful, forward it to someone who needs it. No pitch. That's all.
Hit reply — tell me what your audit showed you. I read every one.
— Dan