Before You Apply: The Unglamorous Week That Decides the Rest
WEEK 102 :: POST 3 :: CLAUDE
Directions Given To The A.I. This Week+
Instructions Given to each A.I. — Please provide 3 prompt variations that share this objective:
Each A.I. also received two static attachments: the blog post template (structure) and the authoring instructions (voice and standards). The text below is the week-specific assignment as sent — reflowed for the web; wording unchanged.
I'd like you to write this week's Ketelsen.ai post. Two files are attached: the blog post template (the structure to follow) and the authoring instructions (context, voice, and standards). Please read both before you begin, then produce the complete post in a single response.
This week's theme: "Getting Your Assets Right Before You Apply."
This is Week 3 of an eight-week series on running a job search with AI. The reader now has a target-role spec sheet from Week 2. This week they build the materials — résumé, LinkedIn presence, and positioning story — against that spec, not in a vacuum. The car-series parallel is "getting your money right before you shop": unglamorous preparation that determines how every later conversation goes.
This is also the series' myth-busting week. The reader has heard that an ATS robot rejects résumés for using the wrong font, that white-text keyword stuffing works, that six seconds is all any human ever spends reading. Some of this folklore contains a grain of truth; much of it is confidently wrong. The honest version — screening software filters and searches, recruiters scan before they read, tailoring beats tricks — is more useful than the folklore, and the prompts should be built on the honest version. What the prompts must not do is have the AI assert how any specific screening product behaves today as settled fact.
The deliverable the reader should walk away holding: a master résumé built against their target spec, a mined inventory of quantified achievements, and a positioning narrative that LinkedIn, cover letters, and interview answers all draw from.
THE SERIES CONTRACT — identical every week; it binds every prompt you design. This series' tagline is its editorial contract: "Use AI like an analyst, not a ghostwriter." It is written for a reader in a market unsettled by AI itself — some readers are searching precisely because AI eliminated their last role. Write with that reader at the table: no AI-efficiency cheerleading, no automation jokes, no promises that AI will "do it for you" anywhere a human hiring decision is involved. And hold one line in every prompt: the AI is the reader's private analyst, coach, and sparring partner — it structures, researches, rehearses, and questions. It does not ghostwrite. Anything a hiring human will read or hear — résumé lines, cover letters, outreach messages, interview answers, negotiation emails, resignation letters — must end in the reader's own words and be true. Prompts should drive toward drafts the reader rewrites and owns, and should say so explicitly. Employers increasingly restrict how AI may be used in their own hiring decisions for legal and compliance reasons, and recruiters increasingly recognize — and discard — material that reads machine-written. A prompt that makes a reader look AI-generated hurts them twice. Posts that ignore this contract should expect to lose the week. Two practical notes. First: a standing “About this series” notice covering these same points is added to every published post automatically at publication — acknowledge the frame in your own voice where your week's prompt calls for it, but do not write a formal disclaimer block of your own, and do not open every post with the same acknowledgment paragraph: outside the weeks whose prompts explicitly carry the series frame, this contract lives in your tone and your prompt design. Second, the framing is POSITIVE: used this way — analyst backstage, reader on the page — AI is an advantage no hiring human will ever hold against your reader. Write like that is true, because it is.
The three prompts should help a reader:
- Rewrite the résumé against a real job description. Take one actual posting for the target role and restructure the reader's résumé against it — matching the language the posting actually uses, surfacing the relevant experience, cutting what does not serve — with the reader supplying both documents.
- Mine achievements they forgot they had. The AI as a structured interviewer that digs quantified accomplishments out of the reader's work history — the project that saved a quarter of the budget, the process that cut a week to a day — because the raw material for every asset is specifics, and most people cannot list their own.
- Build the full asset system. The master résumé that tailored variants are cut from, a LinkedIn profile aligned to the same positioning, and a keyword strategy grounded in the language of real postings for the target role — one coherent story told at three surfaces, ready for Week 5's application engine.
At the advanced tier, the strongest version of this week is a master résumé plus tailoring engine — a system where the reader maintains one complete document and generates posting-specific variants on demand, rather than owning seventeen slightly different résumés. That structure is worth reaching for, and it sets up Week 5 directly.
A constraint for this week. Screening systems differ, change, and do not publish their rules. No prompt may ask the AI to state how a specific ATS product ranks or rejects candidates today, or to guarantee that a formatting choice will pass or fail screening. The prompts should build assets on durable principles — clear structure, the posting's own language, quantified specifics — and where the reader wants to verify a claim about screening, point them to named sources (the employer's own application guidance, recruiters in the target industry) rather than asserting it. And one line matters doubly here: everything in the résumé stays true. The AI structures and sharpens what the reader actually did; it does not invent metrics, titles, or dates, and a prompt that lets it should expect to be marked down.
Design the prompts so the AI does what it is genuinely good at: structured interviewing, restructuring documents against a target, translating accomplishments into the market's language. The reader supplies their history and their real numbers; the AI supplies structure, language, and the questions that surface what the reader forgot. And build the last step into the prompts themselves: the final pass belongs to the reader. A résumé line the reader cannot say out loud in an interview, in their own voice, is not theirs yet — and generic AI phrasing on a résumé is exactly what recruiters have learned to spot and discount. Every prompt that produces candidate-facing text should end by handing the draft back for the reader's own rewrite, and should say why that step is not optional.
Series dependency chain, for the Metadata block: Week 3 consumes Week 2's target-role spec sheet (the assets are built against it). Week 3 produces the master résumé, the achievement inventory, and the positioning narrative — consumed by Week 5 (the application engine cuts tailored variants from the master), Week 6 (the story bank is built from the achievement inventory), and Week 7 (the positioning narrative carries into negotiation).
Because readers may arrive at this post without having read the earlier weeks, the prompts should work for someone with any résumé and a role in mind, while making clear the assets come out sharper when they are built against a real Week 2 spec.
Three difficulty tiers as always — Beginner, Intermediate, Advanced — each a genuinely different approach to the same problem, not the same prompt at three lengths.
On examples: this is a career topic that touches every industry. The template lists tech startup / retail / freelance as suggested industry examples — those are marked MAY, and adapting them is expected here. A project manager translating operations work into product language, a laid-off analyst quantifying five years of "just doing my job," and a designer building a portfolio narrative are the right kinds of contexts. Choosing them over the suggested business examples is correct behaviour and will not be scored against you.
A note on supplied figures. Anything marked `[SUPPLIED — use as given]` above came from Ketelsen.ai's own research brief. Use it freely — you are not fabricating by repeating it, and you will not be marked down for leaving it uncited. Do not attach an invented source to it. (No supplied figures this week — including the folklore numbers: if a "six-second scan" or a "75 percent of résumés rejected" figure is reaching for your keyboard, write the sentence without it.)
## BEFORE YOU SUBMIT — STRUCTURAL CHECK
(This block is identical every week. It exists because these specific items are the ones posts drop, and a dropped structural item costs compliance points for something that takes one minute to add.)
Your post is parsed by a script before any human reads it. Confirm all seven:
1. ☐ Response begins with `PLATFORM: <your name>` and `WEEK: 3` 2. ☐ `## Lead` present once, at the very top, before Variation 1 3. ☐ `## In one line` present in all three variations 4. ☐ `## What this prompt gives you` present in all three variations 5. ☐ `## The Prompt` present in all three variations, with the prompt in double quotes beneath it 6. ☐ `## Introductory Hook` and `## Current Use` present in all three variations (three of each — not one) 7. ☐ Every template heading written as `##`, none bolded instead; prompt breakdown is running text split on ` : `, with no `###` headings inside it
A complete post has 57 `##` headings. If your count is well short, a section is missing or was bolded instead of hashed.
One extra check this week: confirm no prompt asks the AI to state how a specific screening product behaves as fact, or to guarantee a résumé will pass screening — and confirm nothing in any prompt invites the AI to invent achievements, metrics, titles, or dates that are not the reader's own. Also confirm every prompt that produces candidate-facing text ends with the reader rewriting and owning the final words (the series contract's final-pass rule).
About this series. AI is reshaping work — for some readers, it's part of why you're searching at all. We don't pretend otherwise. And the hiring world is wary of AI-written material: many employers restrict how AI may be used in their own hiring decisions, and recruiters increasingly recognize — and discard — machine-written applications. So this series teaches a different approach: use AI as your private research analyst, interview coach, and thinking partner, while every word an employer sees or hears from you stays genuinely, verifiably yours. AI behind the scenes. You on the page.
Week 2 gave you a target. This week you build the materials that go after it — and most job seekers build them backwards, polishing a résumé in a vacuum and hoping it lands somewhere useful. Three prompts, three depths: an interview that digs quantified accomplishments out of your own memory, a rewrite that reads your résumé against one real posting and shows you what a hiring reader would see, and a system that turns a single master document into a tailored variant for every application you send. None of the three writes your résumé for you, and that restraint is the entire point.
The Achievement Excavation Interview
Turn a hazy work history into a list of specifics.
Ask someone what they accomplished in their last job and you get adjectives. Reliable. Detail-oriented. Strong communicator. Ask that same person what actually changed because they were in the building — what was slow before and fast after, what cost more before and less after, what nobody could do until they worked it out — and the room goes quiet. That silence is not modesty. It is that nobody keeps a running ledger of their own impact while they are busy having it, and by the time the ledger is needed, the numbers have gone soft and four years of projects have blurred into one long Tuesday. Everything you build over the next six weeks is cut from that raw material. If the ledger is thin, every résumé bullet, cover letter, and interview answer downstream is thin too.
Why this matters now
Hiring teams are reading more applications than they used to, and reading them faster. Specifics are what survive a fast read — a number, a before-and-after, a named result that a stranger can picture. Generic capability statements do not survive, and they happen to be the exact shape of text that reads as machine-written, which is the one impression you cannot afford to make this year. Fifteen minutes of being interviewed by an AI that refuses to accept a vague answer will produce more usable material than an afternoon of staring at your own résumé. And every line of it will be true, because you supplied every line.
You are going to interview me to help me recover accomplishments from my own work history that I have forgotten or undersold. I am preparing job search materials and I need specifics, not adjectives.
My situation: I am targeting the role of [TARGET ROLE]. My most recent job was [JOB TITLE] at [TYPE OF ORGANIZATION], where I [ONE SENTENCE ABOUT WHAT YOU DID]. Before that, [ONE LINE ON THE PREVIOUS ROLE, OR WRITE 'nothing relevant'].
Interview me one question at a time. Wait for my answer before you ask the next one. Ask about twelve questions in total.
Aim your questions at the things people forget: problems I fixed that nobody assigned me, processes I changed, work that was slow or broken before me and better afterward, money or time saved, people I trained or unblocked, work of mine that got reused after I moved on, and the thing my manager always came to me for.
When an answer of mine is vague, do not accept it. Ask a follow-up that pushes for a number, a timeframe, a before-and-after, or the actual name of the thing. If I say I do not know the number, help me reconstruct an estimate by asking what I do remember — how often it happened, how long it used to take, how many people it involved — and then tell me plainly that the estimate is mine to accept or reject.
Do not invent any detail I have not given you. Do not fill a gap with a plausible-sounding number, tool, title, or date. If I cannot supply something, record it as unknown.
At the end, give me a plain list of everything I told you, one line per item, in this shape: what the situation was, what I did, what changed as a result. Use my own words wherever you can. Then flag the three items that are strongest for a [TARGET ROLE] application and explain why those three.
Finally, list which lines still need a number from me before they are usable, and stop there. I will write the final wording myself.
How the AI reads this prompt
Practical examples from different industries
An operations project manager targeting a product role.
She pastes in that she managed vendor onboarding for a mid-sized distributor and wants a product manager job. The interview asks what she changed rather than what she oversaw, and eleven minutes in she mentions that the onboarding packet used to be a forty-page PDF nobody read, that she rebuilt it as a checklist with owners, and that new vendors went live faster afterward. She does not remember the exact figure, so the model asks how long onboarding used to take, how long it took after, and how many vendors ran through the process in a year — and she reconstructs a defensible range herself. That single item is worth more to a product application than everything currently on her résumé, because it describes discovering a user problem and shipping a fix.
An analyst laid off after five years of "just doing my job."
He arrives convinced he has nothing to show, because his work was maintaining reports that other people used. The interview's questions about reuse and about what his manager always came to him for surface three items he had never counted as accomplishments: a reconciliation process he automated on his own initiative, a quarterly report he redesigned after the finance team kept asking him the same question about it, and the fact that two colleagues were trained by him rather than by anyone official. None of this was ever in a job description, which is precisely why it was invisible to him — and why the interview format finds it and a blank page does not.
A high school teacher moving into corporate training.
She has no business metrics and initially treats that as disqualifying. The interview does not ask for revenue; it asks what was slow or broken before her and better afterward. Out comes a department-wide grading rubric she wrote that colleagues still use, a summer curriculum project she scoped and delivered on a fixed timeline with three other teachers, and an intervention program whose participation she can actually quantify from her own records. Translated into the vocabulary of instructional design, those are curriculum development, project delivery, and program measurement — and every one of them is true, verifiable, and hers.
Creative use case ideas
- Your own performance review. Run the interview two weeks before your self-assessment is due and you will walk into the conversation with specifics instead of the three things that happened most recently. Most people write their review from short-term memory and undersell nine months of work.
- Helping someone re-enter the workforce. Sit with a partner or parent who has been out of paid work and run the interview on their behalf, reading questions aloud and typing their answers. Unpaid work — coordinating care, running a volunteer program, managing a household renovation — surfaces well under this questioning and is almost always undersold by the person who did it.
- A grant or fellowship application. The same excavation works for organizational accomplishments. Point it at your nonprofit's last two years instead of your career and it produces the specifics a funder actually wants in place of mission language.
- A community coach or club organizer building a season summary. Run it on a volunteer season — the youth team, the choir, the neighborhood association — and you get a factual record of what changed under your watch. Useful for the next organizer, and quietly useful for you if that experience ever needs to appear on a page.
- Rebuilding a project post-mortem from memory. When a project ends and nobody wrote anything down, the interview format reconstructs a timeline and a set of outcomes far faster than a meeting where everyone stares at each other.
Adaptability tips
Change the seeded categories and you change what the interview finds. If you are targeting a people-leadership role, replace the list with questions about hiring, developing, retaining, and difficult conversations. For a technical role, aim at systems you designed, incidents you handled, and decisions that outlived you. The category list is the steering wheel.
Adjust the question count to your history. Twelve is right for one substantial job; ask for twenty if you are covering a full career, and run separate sessions per employer if the history is long — one inventory per job is easier to maintain than one giant list.
If the interview stalls or the questions feel generic, paste in your actual job description mid-session and say "use this to ask sharper questions about the same period." Feeding real material to a stalled interview is almost always better than restarting.
For readers who find open questions hard, add "offer me three example answers for each question so I can see the shape of a good one, then let me answer in my own words." The examples prime the recall without putting words in your mouth — as long as you keep the second half of that sentence.
Pro tips
Run the interview by voice if your tool supports it. People remember more when speaking than when typing, and the throwaway aside is where the numbers hide.
Save the raw transcript, not just the final list. Week 6 builds a story bank, and the story bank needs the texture — who pushed back, what nearly went wrong — that the tidy summary strips out.
When the model flags your three strongest items, argue with it. Ask "what would make you rank the fourth item higher?" The answer usually tells you what evidence your inventory is missing, which is more useful than the ranking itself.
Prerequisites
You need a job title you are targeting and rough recall of your last one to three roles. No documents required — this prompt is designed to run before you open your résumé, and running it first is better, because an open résumé anchors you to what is already written on it. If you completed Week 2, have your target-role spec sheet nearby so you can paste the role definition rather than improvising it; the inventory comes out sharper when the model knows what it is digging toward. Set aside twenty uninterrupted minutes and answer in full sentences rather than fragments.
Required tools
Any general-purpose conversational AI — Claude, ChatGPT, or Gemini — on a free tier. The prompt requires no file uploads, no plugins, and no paid features. Voice input is a nice-to-have rather than a requirement. A plain text file or notes app for saving the resulting inventory is worth having open before you start.
Frequently asked questions
What if I genuinely cannot remember any numbers?
That is normal and it is not disqualifying. The prompt handles it by reconstructing estimates from what you do remember — frequency, duration, headcount — and then handing the judgment back to you. Old email, calendar entries, and performance reviews are also better sources than memory; a fifteen-minute search of your sent folder from that period often produces the exact figure. What you must not do is let the AI supply a number to fill the hole, which is why the prompt forbids it explicitly.
Isn't it dishonest to have an AI help me remember my own accomplishments?
No, and the distinction is worth being clear about. The AI supplies questions and structure; you supply every fact. That is the same thing a good career counselor or a sharp friend does over coffee, and nobody considers that dishonest. The line this series holds is that anything a hiring human reads or hears must be true and must be in your words — recovering true things you had forgotten is on the right side of that line by a wide margin.
The model started writing résumé bullets instead of asking questions. What went wrong?
It skipped ahead, which models do when they can see the eventual goal. Say "stop, you are running the interview, not writing the résumé — ask me question four." You should not need to restart. If it keeps happening, remove the mention of job search materials from the opening paragraph and reintroduce it once the interview is finished; models sometimes optimize toward the stated end product rather than the stated process.
How is this different from just using a résumé builder?
A résumé builder starts from a template and asks you to fill in fields, which means you can only tell it things you already knew you had. This starts from your memory and works outward, and the output is not a document — it is an inventory that three later prompts in this series consume. The inventory outlives any single résumé. You will rewrite the résumé a dozen times; you build this once and maintain it.
Should I do this before or after updating my résumé?
Before, ideally by a day or two. An open résumé anchors your recall to what is already on the page, and the whole value of this exercise is finding what is not. If your résumé is already open, close it.
Recommended follow-up prompts
Run Variation 2 next, feeding it the inventory alongside your current résumé — the rewrite is noticeably better when the model has raw material it can promote onto the page rather than only the compressed bullets already there.
A "translate this achievement for a different audience" prompt: paste one inventory item and ask for the version a technical peer would find credible, the version a non-technical hiring manager would understand, and the version that fits in twelve words. Useful preparation for the three surfaces Variation 3 builds.
A "what evidence am I missing" prompt: paste the full inventory and your target role, and ask which capabilities the role requires that the inventory does not yet evidence at all. That list becomes your search plan for old files, not a list of things to invent.
Tags and categories
Tags:
achievement inventory, résumé preparation, structured interviewing, job search, quantifying accomplishments, career change, AI interviewing, beginner prompts
Categories:
Career & Job Search, Beginner Prompts
Citations
NOT APPLICABLE — this variation makes no factual claims requiring external sources. The interview structure is a prompt design choice, not a research finding.
The Posting-Matched Rewrite
See your résumé the way a hiring reader sees it.
There is a particular kind of frustration that comes from sending a well-written résumé to forty postings and hearing nothing. The natural conclusion is that something invisible is rejecting you — a robot, a font, a missing keyword — and an entire folklore industry exists to sell you the fix. Some of that folklore contains a grain of truth. Screening software does filter and search. Recruiters do scan before they read. But the useful version of that insight is far less exotic than the mythology: a résumé written to describe your history in general does not answer the specific question a specific employer asked, and both the software and the human are looking for the answer to that question. The fix is not a trick. It is doing the boring thing forty times, which is exactly the kind of work an AI is good at making survivable.
Why this matters now
You are going to apply to more than one job, and each posting asks for something slightly different in slightly different language. Tailoring is the highest-return activity in a job search and the one people abandon first, because doing it by hand takes an hour per application and feels like it might be pointless. This prompt takes that hour down to something like fifteen minutes, and — more importantly — it shows you the reasoning, so that after four applications you start seeing the mismatches yourself before you ask. It also produces a gap register, which is the part most people skip and the part that tells you whether you are applying to the right jobs at all.
Act as a hiring-side reviewer for the role in the posting below — someone who reads this kind of application every week and decides who moves forward. You are not writing my résumé for me. You are showing me what this posting actually asks for and where my existing material does and does not answer it.
I am giving you two documents.
DOCUMENT 1 — THE POSTING:
[paste the full text of one real job posting for the role you want]
DOCUMENT 2 — MY CURRENT RÉSUMÉ:
[paste your résumé as plain text]
Work in four steps and label each one.
STEP 1 — Read the posting only, ignoring my résumé. List the ten things this employer appears to actually care about, ranked by the weight the posting gives them, and for each one quote the exact words the posting uses. Where the posting uses a term for something I might call by a different name, say so.
STEP 2 — Now read my résumé against that list. For each of the ten, tell me whether my résumé already evidences it, evidences it but buries it, or does not evidence it at all. Quote the line of mine you are judging so I can see what you saw.
STEP 3 — Propose a restructured version for this posting: what order the sections and bullets should be in, which bullets should be rewritten to describe work I already did using the posting's own vocabulary, and which should be cut or compressed because they do not serve this application. For every rewrite, show my original line and your proposed line side by side and name which words you changed and why.
STEP 4 — Give me a gap register: requirements I genuinely do not meet, with no attempt to paper over them. For each, say whether it is something I could evidence from experience that is not on the page yet, or something I actually lack.
Hard constraints. Every claim in your proposed rewrite must trace to something already in my résumé or in material I supply. Do not add a metric, title, date, employer, tool, or responsibility I did not give you. If a bullet would be stronger with a number, mark it [NEEDS NUMBER FROM ME] rather than supplying one. Do not tell me whether this résumé will pass any particular screening system, and do not describe how any named screening product ranks candidates — you have no way to know that, and I am not asking.
End by handing the work back: list the bullets I now need to rewrite in my own voice, and remind me why a line I cannot say out loud in an interview is a line I should not send.
How the AI reads this prompt
Practical examples from different industries
A graphic designer applying to an in-house brand role.
Her résumé reads like an agency portfolio — client names, campaign titles, awards. Step 1 shows that the posting barely mentions campaigns and repeatedly asks about brand systems, design consistency across teams, and working with non-designers. Step 2 finds that all three exist in her history and none are visible: the style guide she built for a client sits inside a bullet about a rebrand, three words from the end. Step 3 promotes it and rewrites the surrounding bullets in the posting's language, and Step 4 tells her honestly that she has never managed a designer, which the posting lists as preferred. She applies anyway, prepared for the question rather than blindsided by it.
A warehouse supervisor applying to a logistics coordinator role.
He assumes the gap is education, since the posting mentions a degree as preferred. The analysis says otherwise: eight of the ten things the posting weights are operational, and his résumé evidences six of them in language nobody outside his warehouse uses. "Ran the shift" becomes a description of daily throughput planning and labor allocation, using the posting's own terms, describing exactly what he already did. The gap register flags the two real gaps — a specific inventory system, and vendor negotiation — and separates them: he has never touched the system, but he has negotiated with carriers regularly and simply never wrote it down.
A nonprofit development officer moving to a corporate partnerships role.
Her résumé is written in nonprofit vocabulary — stewardship, cultivation, major gifts — and Step 1 makes clear that the posting is asking about pipeline management, renewal rates, and multi-year contracts. Those are the same activities under different names, and Step 2 shows the model naming that correspondence explicitly. What makes this example worth studying is Step 3's restraint: the rewrite does not claim she managed a sales pipeline, it describes the donor pipeline she actually managed in language a corporate reader recognizes. The distinction is subtle, it is the whole ballgame, and it only stays intact because the prompt forbids the model to reach past what she supplied.
Creative use case ideas
- A conference talk proposal against the call for papers. Same structure, different documents: paste the CFP and your draft abstract and run the four steps. Selection committees publish what they are looking for and most submissions ignore it entirely.
- A grant application against the funder's RFP. Your organization's boilerplate was written for a different funder. Step 2's buried-versus-absent distinction is especially useful here, because the thing the funder cares about is usually in your materials somewhere, on page four.
- A college or scholarship application against the stated criteria. Teenagers write personal statements about what matters to them and never check them against what the committee said it evaluates. The gap register also gives an honest read on whether an application is a reach.
- A rental or co-op board application in a competitive market. Less obvious, genuinely useful: the listing states what the landlord cares about, and your application usually answers a different question.
- Auditing a job posting you are about to publish. Run it from the other direction — paste your own posting and a résumé of the person you would love to hire, and see whether your posting actually asks for what you want. Most postings do not.
Adaptability tips
Run steps 1 and 2 alone as a fast triage pass. Before you invest in tailoring, the ranked list plus the evidenced-buried-absent verdict tells you in three minutes whether this posting is worth an hour of your evening. Applied across a week of listings, this is the single biggest time saver in the prompt.
Add the achievement inventory from Variation 1 as a third document. It changes the character of Step 3 substantially — instead of only reshuffling what is already on the page, the model can promote something from the inventory that the posting specifically asks for.
For senior roles, add a step between 3 and 4: "identify what this posting is not saying — what the role's real challenge appears to be, based on what the posting emphasizes and repeats." Senior postings often describe the problem they are hiring to solve if you read them for subtext, and that read shapes the cover letter more than the résumé.
If you are applying internationally or across sectors, add "flag any convention in my résumé that is standard where I am but unusual for this employer's market, and ask me before changing it." Résumé conventions vary considerably by country and sector, and you want that surfaced as a question rather than silently corrected.
Pro tips
Paste the posting text rather than a link. Postings behind application portals often do not fetch cleanly, and a partially loaded posting produces a confident analysis of half a job description with no indication that anything is missing.
Save Step 1's output for each posting you analyze. After five or six, lay them side by side — the requirements that appear every time are your real target profile, and they feed Variation 3's language map directly.
When the model's Step 3 rewrite sounds better than your original, be suspicious for a moment before being pleased. Check that the improvement came from clearer structure or better verbs rather than from a claim that quietly grew. Sharper is good; bigger is a problem.
Ask a follow-up: "which of your proposed rewrites would you expect a skeptical interviewer to probe first?" The answers tell you where to prepare, and occasionally reveal that a bullet is overstated in a way you did not notice.
Prerequisites
You need one real job posting for a role you would actually take — pasted in full, including the sections people skip — and your current résumé as plain text. Copying from a PDF often garbles bullets and columns, so check the pasted version reads sensibly before you run it. Ideally you have the achievement inventory from Variation 1 and the target-role spec from Week 2, though neither is required. Budget twenty minutes, and expect to run the prompt again with a second posting before the pattern becomes obvious.
Required tools
Any general-purpose conversational AI that accepts long pasted text. Free tiers work, though a posting plus a two-page résumé is a substantial paste and shorter context windows may struggle with both plus four steps of analysis. If output gets truncated, run steps 1–2 and steps 3–4 as separate messages in the same conversation. No file upload capability is required, and pasting as text is more reliable than uploading a PDF regardless of tier.
Frequently asked questions
Isn't matching the posting's language just keyword stuffing with extra steps?
No, and the difference is whether the underlying claim is true. Keyword stuffing means inserting terms for work you did not do so that a search finds you; using the posting's vocabulary means describing work you actually did in the words the market uses for it. The first is a lie that collapses in the first interview. The second is basic communication, and it is what you would do naturally if you knew that industry's dialect — the prompt just supplies the dialect.
Should I worry about fonts, columns, tables, or PDF versus Word?
Simple, clearly structured documents are easier for both software and humans to read, so the sensible default is a clean single-column layout with real text rather than text inside images. Beyond that, be skeptical of anyone who tells you exactly what a screening system will do with a specific formatting choice — those systems differ, change, and are configured per employer. When it matters, the employer's own application instructions are the authority, and following them precisely is worth more than any formatting theory.
The model rewrote my whole résumé instead of doing the four steps. Now what?
Say "you skipped to the end — go back and do Step 1 only, and stop." Models often collapse multi-step instructions into a finished product because the finished product looks like the goal. If it keeps happening, run each step as its own message rather than asking for all four at once; the sequencing matters more than the convenience.
How many postings should I do this for?
Every application you seriously care about, which is fewer than most people apply to. Twenty tailored applications beat two hundred untailored ones, and the gap register from the first five will probably change which jobs you apply to at all. By the time you reach Variation 3, you are building a system that makes this cheap enough to do every time.
What if the gap register says I am missing most of the requirements?
Then it did its job, and you have real information rather than a month of silence. Some gaps close by remembering — experience that exists but never made the page. Others are genuine, and the useful response is to decide whether to target adjacent roles, acquire the missing thing deliberately, or apply anyway with a clear-eyed cover letter that addresses it. All three are better than not knowing.
Recommended follow-up prompts
A cover letter prompt fed by Step 4's gap register: paste the gaps and ask for three honest ways to address the most significant one in a paragraph, then write the paragraph yourself. Cover letters that acknowledge a gap directly read as confident; cover letters that pretend it is not there read as hopeful.
An interview-preparation prompt built from Step 2: paste the buried-versus-absent verdicts and ask which of your claims a skeptical interviewer would probe first, and what evidence you should have ready for each. This is the seed of Week 6's story bank.
A "second opinion" run of the same prompt in a different AI tool with the identical posting and résumé. Where two models disagree about what the posting weights most, the disagreement is usually pointing at genuine ambiguity in the posting — which is worth knowing before you write to it.
Tags and categories
Tags:
résumé tailoring, job postings, keyword matching, gap analysis, application strategy, document rewriting, intermediate prompts, job search
Categories:
Career & Job Search, Intermediate Prompts
Citations
USAJOBS Help Center, "What should I include in my federal resume?" — U.S. Office of Personnel Management. Cited as an example of an employer publishing its own application requirements explicitly, which is the kind of source that settles a formatting question when a general rule of thumb cannot.
The Master Résumé and Tailoring Engine
Build one document that generates every version you need.
Somewhere on your machine there is a folder containing resume_final.docx, resume_final_v2.docx, resume_TechCo.docx, and resume_final_USE_THIS_ONE.docx, and you no longer know which of them contains the good version of the third bullet. This is what happens when tailoring is done as a series of one-off edits: each application forks the document, improvements get stranded in whichever fork received them, and after two months your best material exists in a file you will never open again. The alternative is a structure borrowed from anyone who maintains something complex — keep one complete source of truth, and generate the variants from it. Build it once, in an afternoon, and every application afterward is a cut rather than a rewrite. It is also the only version of this that stays true, because there is one place where the facts live.
Why this matters now
This is the week to build it, because Week 5's application engine assumes it exists and Weeks 6 and 7 draw on the same positioning. Doing it now converts what would be fifteen separate hours of tailoring across a search into a single build plus fifteen quick cuts. There is a second reason to do it now rather than later: your positioning gets sharper the more postings you read, and it gets fuzzier the more times you edit a document without a source of truth. A master résumé holds the sharpening. And when a role you genuinely want appears with a Friday deadline, the difference between having this and not having it is the difference between a considered application and whatever you can produce at eleven at night.
Act as my positioning strategist for a job search. We are building one system today and using it for months: a master résumé I maintain, a positioning narrative everything else draws from, and a tailoring procedure I can re-run for any posting without rebuilding anything.
Inputs I am supplying:
A. MY TARGET ROLE SPEC —
[paste your target role definition: the titles, the kind of organization, the level, and what you have decided the job should involve. If you have a spec sheet from earlier work, paste it. If not, write three sentences.]
B. THREE TO FIVE REAL POSTINGS for that role —
[paste each in full, separated by a line of dashes]
C. MY FULL WORK HISTORY —
[paste everything: current résumé, achievement inventory, projects, volunteer work, anything you might ever want on a page. Over-supply deliberately. This is raw material, not output.]
Produce five artifacts, each under its own heading.
ARTIFACT 1 — LANGUAGE MAP. Across the postings, identify the vocabulary that recurs: what this market calls the work, the responsibilities, the tools, the outcomes. Separate three tiers — terms in nearly every posting, terms in some, and terms appearing once that are probably one company's internal dialect. For each recurring term, tell me whether my history contains the thing the term describes, even if I have always called it something else. Build this map from the postings themselves; do not tell me what any screening software does with these words.
ARTIFACT 2 — POSITIONING NARRATIVE. One paragraph, third person, stating who I am professionally, the through-line across my history, and the specific value I bring to this target role. Then the one-sentence version, then the three-word version. Build it only from what my history supports. Where the narrative would be stronger with something I have not evidenced, say so explicitly instead of writing the stronger version.
ARTIFACT 3 — MASTER RÉSUMÉ. A complete document, deliberately too long to send anywhere: every role, every achievement, every bullet worth keeping, organized so variants can be cut from it. Group bullets by the capability they evidence and tag each with the terms from the language map it speaks to. Preserve my numbers exactly as I gave them. Where a bullet needs a number I did not supply, mark it [NEEDS NUMBER].
ARTIFACT 4 — TAILORING PROCEDURE. Write the reusable instruction I will paste back to you, together with a new posting, to generate a variant from the master. Specify how to select which bullets survive, how to order them, how much rewriting of my language is permitted, what must never change, and what to do when a posting asks for something the master cannot evidence. Write it as a prompt I can copy, not as advice about tailoring.
ARTIFACT 5 — LINKEDIN ALIGNMENT. Headline options, an About section outline, and how the experience entries should differ from the résumé, given that LinkedIn is a searchable public profile read by people who found me rather than a document I sent. State plainly where the positioning must be identical to the résumé and where it should be broader.
Then evaluate your own output against these criteria and report where it falls short: (1) every factual claim traces to material I supplied; (2) no invented metric, title, date, employer, or tool; (3) the narrative is one story rather than five; (4) a variant cut using Artifact 4 would still be entirely true; (5) nothing anywhere asserts how a specific screening product behaves.
Everything you produce is scaffolding. Before any of it reaches a hiring human, I rewrite it in my own words. Mark the passages you think I am most likely to send unchanged — those are the ones I most need to rewrite.
How the AI reads this prompt
Practical examples from different industries
A security analyst targeting detection engineering roles.
He supplies four postings and eleven years of incident work. The language map shows that "detection engineering" postings consistently ask for detection logic, tuning, and false-positive reduction — all of which he has done constantly and described on his résumé as "responded to alerts." Artifact 2 builds the through-line: someone who has spent a decade on the receiving end of bad detections and knows exactly why they fail. Artifact 3 groups his bullets by capability rather than by employer, which is what finally makes the tuning work visible; scattered across three jobs chronologically, it had read as routine operations. His tailoring procedure specifies that incident volume figures never change and vendor names get swapped per posting.
A freelance copywriter consolidating six years of client work.
Her problem is the opposite of a gap — she has fifty projects and no shape. Over-supplying the input means pasting a client list, sample briefs, and testimonials, and the language map immediately separates what agency postings ask for from what in-house postings ask for, which tells her she has been applying to two different jobs with one document. The positioning narrative forces a choice she had been avoiding for a year. Artifact 4's procedure ends up specifying that the client roster is filtered by sector for every variant, which is the single edit that makes her applications look intentional instead of scattered.
A teacher of fourteen years moving into instructional design.
Her history contains no vocabulary from the target market at all, which makes Artifact 1's translation instruction the whole exercise: curriculum design, learner assessment, stakeholder alignment with department heads, iterative revision based on outcome data. Every one of those is something she genuinely did, and none of them appear in her current résumé in those words. The self-audit matters here more than anywhere, because translation across sectors is where a model is most likely to reach past the evidence — and criterion (1) catches the two places where it did, both of which she then corrects with what actually happened.
Creative use case ideas
- An artist's or musician's press kit. The same architecture — a master document holding every show, review, collaboration, and piece, with variants cut for gallery submissions, grant applications, and venue bookings. Artists rebuild this from scratch every time and lose material with each rebuild.
- A small business's capability statement library. One master document of past work, certifications, and outcomes, with variants generated per RFP. Small firms lose contracts to larger competitors partly because the larger competitor has this and they do not.
- A volunteer's or board member's service record. Nonprofits ask for a bio, and most people write a new one badly each time. A master service record with a positioning narrative produces a consistent, accurate bio in a minute.
- An academic's materials system. Teaching statement, research statement, and CV all draw from one positioning narrative, and the persistent problem of three documents telling three slightly different stories disappears.
- A family's or community group's grant history. Less obvious: a master record of what a group has done, tagged by the outcomes funders ask about, makes the next application a selection problem rather than an act of collective memory.
Adaptability tips
Rebuild Artifact 1 every four to six weeks during an active search, using the postings you have read since. Market vocabulary shifts, and a language map from two months ago quietly stops matching what you are reading. The master and the narrative are stable; the map is the part that ages.
Split the prompt if your inputs are large. Run Artifacts 1 and 2 in one session, confirm the positioning is right before anything is built on it, then run 3 through 5 in a second session with the narrative pasted back in. Positioning errors propagate into everything downstream, so it is worth a checkpoint.
Point the same structure at a different target. If you are searching across two role types — the safe one and the ambitious one — build two language maps and two narratives against one master résumé. The master does not change; the cuts do. This is exactly the situation the architecture exists for.
For an internal move or promotion case, swap the postings for your organization's job architecture or leveling guide and the narrative for a case document. The mechanics are identical: evidence, mapped to a target's own language, cut for one audience.
Pro tips
Keep the master résumé in plain text or Markdown rather than a formatted document. You will paste it into an AI tool dozens of times, and formatted files paste badly. Format only at the variant stage, when a specific document is going to a specific employer.
After generating Artifact 4, test it immediately on a posting you already analyzed with Variation 2. If the procedure produces a materially worse result than your hand-tailored version, the procedure is underspecified — ask what it missed, and add that to the instruction. A tailoring procedure you have not tested is a plan, not a tool.
Keep a changelog at the top of your master. One line per edit, dated. Three months into a search you will want to know when a bullet changed and why, particularly if the version you sent to a company that just called you back is not the version currently in the file.
When the self-audit says the output passes all five criteria, ask "which criterion were you least confident about, and why?" The confident all-clear is less informative than the hesitation, and the hesitation is usually pointing at a real soft spot.
Prerequisites
You need three to five real postings for your target role, saved as text, and as much of your work history as you can assemble — old résumés, project files, performance reviews, the achievement inventory from Variation 1. Set aside an unhurried session; this is the one prompt in this post that is genuinely a project rather than a task. Week 2's target-role spec makes Artifact 2 substantially better, because a positioning narrative built against a defined target is sharper than one built against a job title. You also need somewhere durable to keep the output, since the entire premise is that this document persists and gets maintained.
Required tools
A conversational AI with a large context window, since this prompt supplies several postings plus a full work history in one message. Paid tiers of Claude, ChatGPT, or Gemini handle it comfortably; free tiers may require splitting the prompt across two sessions as described in the Adaptability Tips. A plain-text editor for maintaining the master document, and a place to store it where you will actually find it in three months. No specialized résumé software is needed or recommended.
Frequently asked questions
How long does this actually take?
Assembling the inputs takes longer than running the prompt — an hour or two of digging through old files, then twenty minutes of generation, then a real editing pass where you rewrite everything in your own voice. Call it half a day, done once. Compared against fifteen minutes per application forever, it pays for itself somewhere around application number eight, and the quality difference shows up much earlier than that.
Should the master résumé ever be sent to anyone?
No. It is deliberately too long, includes tagged capabilities and bracketed placeholders, and contains material irrelevant to any specific employer. Sending it would be like sending a photographer's contact sheet instead of a print. The variants cut from it are what you send, and each of those should be a document you have read aloud before it leaves.
What if the positioning narrative does not sound like me?
Then rewrite it, which was always the plan. The model's job is to find the through-line in your history and state it clearly; your job is to say it in a voice you can sustain in an interview. The three-word version is a useful test — if you cannot say those three words to a stranger without wincing, the narrative is aimed slightly wrong and it is worth another pass before anything else is built on it.
Does putting keywords in the master résumé help me get found?
On a public profile like LinkedIn, being findable by the words recruiters actually search for is a real and reasonable goal, which is why Artifact 5 exists. On a résumé you submit, the honest framing is different: use the market's language because it describes your work accurately to the person reading it. What nobody outside a given company can tell you is exactly how that company's screening configuration treats any particular word today. Build on accurate language and clear structure, and take specific claims about specific systems as folklore until an employer or a recruiter in that industry tells you otherwise.
Can I use this if I am employed and just keeping options open?
Yes, and it is arguably the better time to build it, because your accomplishments are fresh and you are not under deadline pressure. Update the master quarterly with what you have done since. The people who conduct short, effective job searches are almost always the people who maintained something like this before they needed it.
Recommended follow-up prompts
The tailoring procedure itself — Artifact 4 — is your next prompt, and it is the one you will use most. Run it against a live posting the same day you generate it, while the model still has the master in context and you can catch an underspecified instruction cheaply.
A "narrative stress test": paste the positioning narrative and ask what a skeptical interviewer would ask to test whether it is true, then check that you can answer every question with something from the master. Gaps between the story and the evidence are far better found now than in a room.
A quarterly maintenance prompt: paste the master résumé and a list of what you have done since the last update, and ask which existing bullets are now superseded and which new items belong in which capability group. Maintenance is what makes the system worth having, and a prompt makes it a fifteen-minute task.
Tags and categories
Tags:
master résumé, résumé system, positioning narrative, LinkedIn optimization, tailoring engine, reusable prompts, career strategy, advanced prompts, job search
Categories:
Career & Job Search, Advanced Prompts
Citations
New York City Local Law 144 of 2021, governing automated employment decision tools — cited as an example of jurisdiction-specific regulation of hiring software, which is part of why blanket claims about "how the ATS works" are unreliable across employers.
Illinois Artificial Intelligence Video Interview Act (820 ILCS 42) — cited for the same reason: employer obligations around AI in hiring vary by jurisdiction and continue to change.
LinkedIn Help Center, profile and search visibility guidance — cited as the platform's own documentation, which is the appropriate authority for how a public profile surfaces in search, as opposed to third-party speculation.
Which of the three should you use?
The three prompts solve three different problems and are best understood by what each one takes as its input. Variation 1 takes nothing but your memory and produces facts. Variation 2 takes facts and one posting and produces a diagnosis. Variation 3 takes facts and several postings and produces a system. That progression is also the natural order of use — the interview before the rewrite, the rewrite before the engine — and a reader with an hour this week should spend it on Variation 1, because the other two are only as good as the material they are given.
They overlap least where it matters most. All three refuse to invent, all three end by handing the words back to you, and all three treat market vocabulary as a translation problem rather than a keyword problem. Beyond that they behave quite differently: Variation 1 is conversational and slow, Variation 2 is analytical and produces a document you argue with, and Variation 3 is architectural and produces something you maintain. Running Variation 2 twice in a week is normal. Running Variation 3 twice in a week means something went wrong the first time.
Choose by where you are stuck. If your résumé feels thin and you cannot say why, the problem is raw material and Variation 1 is the answer. If your résumé feels fine but nothing lands, the problem is that it answers a general question rather than a specific one, and Variation 2 will show you exactly where. If tailoring is working but costing you an hour per application and your files have started multiplying, the problem is architecture, and Variation 3 is the fix. A reader who runs all three in sequence ends the week holding a master résumé, an achievement inventory, and a positioning narrative — which is precisely what Weeks 5, 6, and 7 expect you to bring.
TAGS:
