WEEK 107 :: 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: "Sticking the Landing — Decision, Resignation, and the First 90 Days."
This is Week 8, the final week of an eight-week series on running a job search with AI. The negotiation is done and a final offer is on the table. Three endings can still go wrong from here: the decision (excitement making the call instead of the Week 2 criteria), the exit (a burned bridge or a counteroffer trap — the current employer suddenly finding money, which the reader should evaluate with open eyes given how often an accepted counteroffer just postpones the departure), and the landing (the sink-or-swim first quarter that decides the story of this job before most people have found the good coffee).
The reader started this series restless in Week 1. This week they finish it: signed, resigned gracefully, and onboarding with a plan — the same full-journey capstone the car series closed on.
The deliverable the reader should walk away holding: a scored final decision, a resignation plan with scripts, and a 30/60/90-day onboarding plan with stakeholder map and early wins identified.
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:
- Score the final decision. A weighted decision matrix that scores the final offer (or offers) against the Week 2 spec sheet — must-haves, deal-breakers, and priorities the reader set before any offer existed — so the call is made by criteria written by a calmer version of the same person, with a written record of why.
- Exit gracefully and survive the counteroffer. Resignation scripts for the conversation and the letter, a plan for the awkward two weeks, and a counteroffer framework that has the reader decide their response before the counter comes — including an honest look at why the money appeared only when they stood up to leave.
- Plan the first 90 days. A 30/60/90 onboarding plan built around stakeholder mapping (who decides, who influences, who knows where the bodies are buried), early wins matched to what the reader learned about the team in Weeks 4 and 6, and a cadence for turning a new job into an established position.
At the advanced tier, the strongest version of this week is the complete landing package — decision record, exit plan, and onboarding plan as one coherent handoff from "candidate" to "new hire," each part feeding the next. And because this is the series capstone, the strongest posts can close the loop the reader has lived: eight weeks of artifacts, each built on the last, is a working demonstration that structured, reusable prompting is how a long project holds together — said quietly, as an observation, not a pitch.
A constraint for the last week. Resignations touch contracts and law: notice obligations, non-competes, equity forfeiture, benefits timing — jurisdiction-specific, contract-specific, and not the AI's call. No prompt may ask the AI to interpret the reader's employment contract, state notice or non-compete rules as settled fact, or adjudicate what the reader is legally owed on exit. The scripts handle the human conversation; the contract questions get pointed at the reader's actual documents, HR, and where it matters, a lawyer. Counteroffer statistics follow the series rule: no invented figures — the pattern can be described honestly without a percentage stapled to it.
Design the prompts so the AI does what it is genuinely good at: scoring options against pre-set criteria, drafting graceful language for hard conversations, and structuring ninety days of ambiguity into a plan. The reader supplies the offer, the criteria, and what they learned about the company; the AI supplies the matrix, the scripts, and the calendar.
Series dependency chain, for the Metadata block: Week 8 consumes the whole series — Week 1's decision criteria and baseline, Week 2's spec sheet (the decision matrix scores against it), Week 4's dossier and Week 6's diagnostic reads (the landing plan's intelligence), and Week 7's negotiated outcome (the numbers being decided on). Week 8 produces the signed decision, the exit, and the onboarding plan — and, in the pattern this series inherited from the last one, its real successor is the next reader's Week 1: the restlessness this plan is designed to prevent.
Because readers may arrive at this post with an offer in hand and none of the earlier weeks read, the prompts should work from the offer and the reader's own priorities, while making clear the decision matrix is only as good as the criteria behind it — and criteria written before the offer existed are the trustworthy kind.
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 reader choosing between the offer and a surprise counteroffer, one resigning from a team they love for a role they need, and a new hire mapping stakeholders in their first month 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 — the counteroffer-departure statistic that circulates in this territory is exactly the kind of number to write around, not assert.)
## 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: 8` 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 interpret an employment contract, state notice or non-compete rules as fact, or assert a counteroffer statistic. Contract questions go to the reader's documents, HR, or a lawyer; the decision matrix runs on the reader's own Week 2 criteria. Also confirm resignation and counteroffer-response scripts end in the reader's own words — the series contract's final-pass rule applies to exits too: a resignation letter is read by the humans who will be the reader's references for years.
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.
An offer on the table is not the end of a job search — it is the last place one can still go badly wrong. This week closes the series with three prompts at three depths: a cold-light check that holds the offer against the criteria you wrote before it existed, an exit rehearsal that settles your counteroffer answer before the counteroffer arrives, and a full landing package that carries you from decision to resignation to a mapped first ninety days. Eight weeks come down to three conversations: the one with yourself, the one with your current manager, and the one you have not had yet with a team that does not know you.
The Cold-Light Offer Check
Hold the offer against what you wanted before it existed.
Offers arrive with a chemical attached. Someone chose you, the number is bigger than the old number, and a version of you that has been tired for months would very much like to say yes before anything can spoil it. That version is not a bad person. It is simply a poor analyst, and it is the one holding the phone. The check on it already exists: you wrote it in Week 2, before any of this was on the table, back when you had nothing to gain by lying to yourself about what you actually needed. This prompt does one small, unglamorous thing. It puts the offer and those criteria in the same room, in plain language, and shows you exactly where they fail to meet.
Why this matters now
Decision windows are short. Most offers come with a few days attached, often spanning a weekend, and that compression is doing work that has nothing to do with whether the job is right for you. An unsettled market makes it sharper still — when roles are scarce, any offer starts to feel like the only offer, and scarcity is a terrible advisor. Ten minutes of structured comparison before you reply does not slow the decision down. It stops you from having to make it twice.
I have a job offer I need to decide on, and I want to think it through clearly rather than emotionally. Act as a practical decision coach, not a cheerleader. Do not tell me what to choose.
Here is what I wanted from a new job, written down before this offer existed: [YOUR MUST-HAVES, YOUR DEAL-BREAKERS, AND THE THREE THINGS THAT MATTER MOST — pay, manager, commute, stability, learning, hours, whatever they honestly are]
Here is the offer: [PASTE WHAT YOU KNOW — title, pay, benefits summary, start date, team, location, reporting line]
Here is my current job: [ONE PARAGRAPH ON WHAT IS GOOD AND WHAT IS NOT ABOUT WHERE YOU ARE NOW]
Do four things, in plain language, in this order.
1. Go through my must-haves and deal-breakers one at a time and say whether the offer clearly meets it, clearly misses it, or is unclear from what I have told you. Do not guess at anything I did not give you — mark it unclear and say what you would need to know.
2. Write a short plain-English description of what accepting this would actually change about my week-to-week life, and what it would leave exactly as it is.
3. List the three most important questions I have not answered yet — the ones I would want answered before I sign anything.
4. Finish with a paragraph headed 'What would have to be true' that completes this sentence honestly: this is the right move if the following things turn out to be true.
One boundary: do not interpret my offer letter or employment contract, and do not tell me what notice I owe, what a non-compete means, or what I am entitled to on the way out. Anything in that territory goes on a short list at the end for me to take to my own documents, to HR, or to a lawyer. Ask me no questions until all four sections are written.
How the AI reads this prompt
Practical examples from different industries
Leaving a team you like.
A support lead at a mid-size software company has an offer with a meaningful raise and a step up in title. She pastes in her Week 2 criteria — stable hours, a manager who gives real feedback, no on-call — alongside the offer and a paragraph about her current team, whom she genuinely likes. The memo comes back marking pay and title as clear hits, feedback quality as unclear because nobody in three interviews described how reviews work there, and on-call as a miss she had somehow stopped noticing. The "what would have to be true" paragraph names the condition plainly: this works if the new manager turns out to be a coach, which is a thing she has evidence for from exactly one conversation. She does not decline the offer. She emails the hiring manager and asks how feedback happens, and the answer decides it.
One offer, no leverage.
A logistics coordinator laid off in a restructuring has a single offer, arriving in week eleven of unemployment, paying about twelve percent below his old salary. The temptation is to skip the analysis entirely, because what is there to compare against. Running the check anyway does something useful: it separates "this misses my pay criterion" from "I should refuse it," which are different findings, and it produces the three questions worth asking before he signs — about review timing, about how the role was funded, about who held it last. He takes the job. He also takes it knowing precisely which criteria he set aside and why, which is the difference between a compromise and a drift.
A career-changer trading down on paper.
A pharmacy technician moving into clinical data work has an offer that is a step backwards in seniority and roughly flat on pay. Against a generic "is this a good offer" standard it looks weak. Against her own written criteria — get into the field, work with people who will teach me, survive financially for two years — it scores well on the first two and flags the third as unclear, because she never worked out the actual monthly numbers. That gap is the real finding, and it sends her to a spreadsheet before it sends her to a decision.
Creative use case ideas
- The apartment or house decision. Write your criteria before the first viewing, then run each place against them. Property viewings are engineered to overwrite your priorities with a feeling about a kitchen, which is exactly the failure this prompt is built to catch.
- Choosing a school or a childcare arrangement. Parents typically hold their real criteria — proximity, temperament fit, what happens when a kid struggles — in their heads, where a good tour can quietly rearrange them. Writing them down first makes the tour evidence rather than persuasion.
- Whether to take a big freelance client. Rates make a contract look better than the working relationship will feel. Score it against what you wanted from your client list — payment behaviour, scope discipline, whether the work is a portfolio piece — and the answer often changes.
- The volunteer commitment. Board seats, coaching a team, running the committee: these are almost always accepted in the warmth of being asked. The same four-step memo works, and the "what would have to be true" paragraph usually reveals a time assumption that will not hold.
- Deciding to stay. Run the check with your current job in the offer slot. Most people have never once evaluated where they already work against criteria they wrote deliberately, and the result is frequently more interesting than the offer.
Adaptability tips
Swap the criteria block and the prompt does something else entirely — that is its whole flexibility. Any decision where you can state what you wanted before the option appeared fits the same four steps. If you have two offers rather than one, paste both and add "compare them against my criteria and against each other, in that order," so the criteria stay primary and the offers do not simply grade each other. If your criteria are vague, ask for a first pass that turns them into something checkable before you run the comparison: "before you do anything else, restate my criteria as testable statements and show me." If you want more challenge, add "then argue the opposite case in one paragraph" — the model will find the strongest argument against the reading it just gave you, which is a cheap and effective test of whether the memo was actually thinking or merely agreeing.
Pro tips
- Paste your criteria and the offer in separate messages, criteria first, and let the model acknowledge them before it sees the offer. Order shapes weighting more than most people expect.
- Save the finished memo somewhere dated. Six months into any job there is a week where you cannot remember why you took it, and a contemporaneous record beats reconstruction every time.
- If the memo reads as unreservedly positive, be suspicious of your inputs rather than pleased with your luck. It usually means the criteria you supplied were written after you had already started wanting the job.
- Read the "unclear" list as your recruiter question list. It converts the weakest part of the analysis into the most productive email you will send this week.
Prerequisites
Have the offer details in front of you — title, compensation, benefits summary, start date, reporting line, location and work pattern. Have the criteria you set at the start of your search; if you followed this series, that is your Week 2 spec sheet. If you did not write criteria in advance, spend ten minutes writing them now, and be honest that criteria written with an offer in hand are weaker evidence than criteria written without one. You do not need to have accepted or declined anything to run this.
Required tools
Any general-purpose AI assistant — Claude, ChatGPT, Gemini or similar. Free tiers handle this comfortably; the prompt is a single turn with a modest amount of pasted context. No plugins, no file uploads, no special features required. If your assistant offers a private or temporary chat mode, use it, since you are pasting compensation details and candid notes about your current employer.
Frequently asked questions
What if I never wrote down criteria before the search?
Write them now, but flag it to yourself. Criteria composed while an offer is sitting in your inbox tend to bend toward the offer — you will find yourself listing "growth opportunity" precisely because the job description used the phrase. A partial workaround is to write your criteria from the perspective of the job you are leaving: what made you start looking, and what would need to be different for you to stop. That framing pulls from a period before this offer existed, which is what gives criteria their authority.
The memo said something is unclear that I thought I knew. Now what?
That is the prompt working. There is a large gap between what you sensed during interviews and what anyone actually told you, and the model can only see the second category. Treat every "unclear" as a question for the recruiter or hiring manager. Asking them before you sign is normal, expected, and quietly informative — how someone answers a direct question about feedback or workload is itself data about the place.
Can I ask it to just tell me whether to take the job?
You can, and it will produce something confident, and you should not weight it. The model does not know how your Sunday nights feel, what your household needs, or what you can tolerate for eighteen months. What it does well is holding two things side by side without flinching. Keep it in the analyst seat and make the call yourself — the reasoning it surfaces is the value, not the verdict.
Is it safe to paste an offer letter into an AI tool?
Use judgement. Salary figures, your name, and your current employer's details are sensitive, and consumer AI tools vary in how they handle conversation data. The safest version is to paste the substance without the identifying wrapper: the numbers, the terms and the role, without the letterhead, the names, or anything marked confidential. Check your employer's and the new employer's policies, and check your tool's data settings before you decide.
Recommended follow-up prompts
- The exit rehearsal (Variation 2 of this post) — the natural next step once the decision is made, covering the resignation conversation and the counteroffer you should answer before it arrives.
- A "questions for the hiring manager" prompt — feed the unclear items from your memo back in and ask for the three most useful ways to raise them without sounding like you are having second thoughts.
- Week 2's spec sheet prompt — worth re-running from scratch if you are starting a new search, since the criteria you write at the beginning are the only ones this check can lean on later.
Tags and categories
Tags:
job offer, decision making, career criteria, offer evaluation, job search, decision memo, beginner prompts, career change Categories: Career & Job Search, Decision Support
Citations
Hammond, J. S., Keeney, R. L., & Raiffa, H. Smart Choices: A Practical Guide to Making Better Decisions. Harvard Business School Press — the standard general-audience treatment of setting objectives before evaluating alternatives, which is the principle this variation applies to a single job offer.
The Exit Rehearsal
Rehearse the resignation and the counteroffer before either one happens.
Resignations are short conversations with long shadows. The person across the desk will be a reference for years, will turn up at a conference in 2031, and will remember the tone more accurately than the words. Most people prepare for it the way they prepare for a dentist appointment — by not thinking about it and then getting through it — which is how a clean exit becomes a five-minute monologue about everything that has been wrong since March. And then there is the counteroffer: money that was apparently available all along, produced under the only condition that ever generates it, offered to a person who is emotionally exhausted and would love for the hard part to be over. That is not a good moment to be deciding anything. This prompt moves the deciding earlier.
Why this matters now
Two things happen in the same week, and they interact. You resign, and your employer responds — sometimes gracefully, sometimes with an offer designed to make you stay, sometimes both. Handled badly, the second one unwinds a decision that took eight weeks to make. Handled well, it becomes a piece of information you already knew how to interpret, because you wrote down what you would say to a match, a beat, and a shiny new title before anyone said any of them out loud. The rehearsal also protects the relationship, and in a market where people move often, that relationship is an asset you will draw on again.
Act as a resignation coach and sparring partner. I am leaving a job and I want the exit planned before I open my mouth, not improvised afterwards.
My situation:
- Role, and how long I have been here: [FILL IN]
- Who I resign to, and what our relationship is like: [FILL IN]
- How this organisation usually reacts when people resign: [FILL IN, or write unknown]
- What I want from these people in five years: [reference, rehire eligibility, friendship, nothing]
- Notice period as I understand it from my own paperwork: [FILL IN — take this as given, do not verify or interpret it]
- Chance they respond with more money or a new role: [low / medium / high]
- Why I am actually leaving: [BE HONEST HERE — this is for calibration, not for the letter]
Settings to follow:
- Tone: [warm and personal / brief and professional / formal]
- Letter length: [short / standard]
- Directness in conversation: [gentle / plain / very direct]
Produce five things.
1. A script for the resignation meeting: my opening two sentences, what I say when asked why, what I say if pressed for more detail than I want to give, what I say if the reaction is anger or guilt, and how I close. Keep the language simple enough that I can say it in my own voice without sounding rehearsed.
2. A resignation letter draft at the tone and length I set, containing no complaints, no grievances, and no explanation I did not ask for. Write it so I would be comfortable having it read aloud two years from now.
3. A counteroffer framework I complete now, before any counter exists. Cover: what change, if any, would genuinely alter my decision; what I say if they match the offer; what I say if they beat it; what I say if it is a title or a promise rather than money. Include a short, unsentimental list of the reasons compensation can appear at exactly this moment and not before, so I can weigh them with open eyes instead of in the room. Do not attach statistics or percentages to any of it.
4. A plan for the notice period: what I document, what I hand over and in what order, who I tell and when, and what I do not say to anyone.
5. A list of every question I should take to my own contract, to HR, or to a lawyer rather than to you — anything touching notice obligations, non-competes, equity or bonus timing, benefits, or what I am owed on exit.
Do not interpret my contract, tell me what notice I legally owe, or explain what any clause means. Route all of that to section 5.
When the five sections are done, stop and ask me one question: which line here sounds least like something I would actually say? Rewrite that section after I answer. Everything you draft is a starting point — I will rewrite it in my own words before any human reads or hears it.
How the AI reads this prompt
Practical examples from different industries
The counteroffer that was always available.
A senior analyst at an insurance firm has been asking for a market adjustment for eighteen months and hearing about budget cycles. She resigns on a Tuesday; by Thursday there is a number on the table that exceeds her new offer. Because she ran the framework first, she already has her answer written down — she had decided that money alone would not change her decision, because the reasons she was leaving were the review cycle and a manager who did not advocate for her, neither of which a raise addresses. The framework also handed her the graceful language for declining, which matters, because the person making the counteroffer is not the enemy and will be in her professional life for a long time.
Leaving a team you love for a role you need.
A veterinary practice manager is moving to a larger clinic for a schedule that lets him see his kids on weekdays. He does not have a grievance; he has a life change, which is a harder conversation in some ways because there is nothing to point at. He sets tone to warm and personal, directness to gentle, and answers the five-year question with "friendship." The script that comes back opens by naming the decision plainly before explaining anything — the sequencing detail that keeps a resignation from turning into a negotiation — and the notice plan front-loads the handover of the two things only he knows how to do, which is the concrete way to be generous on the way out.
The awkward two weeks in a small team.
A designer at a nine-person agency knows the news will be around the studio within an hour of the meeting. His plan matters more than his letter. The notice-period section produces an order of telling — manager first, then the two people whose projects he is on, then everyone — plus a short list of what not to say, including anything about where he is going and anything about compensation. He also gets a documentation checklist, which is the part everyone intends to do and finishes at 11pm on their last day.
Creative use case ideas
- Leaving a volunteer role or a board seat. Structurally identical, socially harder, because the people you are disappointing are friends and the counteroffer takes the form of "could you just stay on for the gala?"
- Ending a client relationship as a freelancer. Same script, different stakes: you need the reference, the referral pipeline and the possibility of coming back in two years, all of which survive or die in one conversation.
- Stepping down as a coach, a chair, or an organiser. Community groups run on individuals who cannot leave, and the exit plan — handover, documentation, who is told in what order — is often the most valuable thing the departing person produces.
- Rehearsing any conversation you are dreading. Telling a landlord you are leaving, telling a co-founder you are out, telling a family member you will not be hosting this year. The branching structure is what transfers: script the hard reactions, not the easy one.
- Preparing for a conversation you are on the receiving end of. Run it with the roles reversed to think through how you would want to handle someone resigning to you. Managers rarely rehearse this and it shows.
Adaptability tips
The parameter block is the adjustment surface — run the prompt twice at different tone settings and read both letters before choosing, since the gap between "warm and personal" and "brief and professional" is larger than it sounds and you may not know which one you want until you see them side by side. If you are not resigning but restructuring a relationship — going part-time, dropping a client to a smaller retainer, leaving one team for another internally — replace the resignation letter with the equivalent artifact and keep the rest intact. If your notice period is long, ask for the plan segmented by week rather than as a single block, because eight weeks of lame-duck status has a shape that two weeks does not. And if you know the counteroffer is coming, ask the model to run the counteroffer conversation as a live roleplay after the framework is built: it plays your manager, you answer out loud, and you find out which of your written answers you can actually say.
Pro tips
- Read the conversation script aloud before you use it. Anything you stumble over is a line that is not yours, and stumbling in the meeting is what makes a prepared exit sound prepared.
- Write the counteroffer framework before you have even accepted the new offer. Once you have resigned, every hour that passes makes it harder to think about neutrally.
- Ask for the letter at one paragraph shorter than you think you need. Resignation letters have exactly two jobs — stating that you are leaving and giving your last day — and every additional sentence is an opportunity to create a document you regret.
- Keep the "what I do not say" list somewhere you will see it during your notice period. The gap between resigning well and resigning badly is usually a conversation on day nine.
Prerequisites
You need a decision you have actually made — this prompt is for executing an exit, not for weighing one, and running it while undecided produces scripts you cannot deliver convincingly. Have your notice period from your own paperwork, your last day if you have agreed one, and a realistic read on how your workplace handles departures. Know what you want from these people afterwards. If a counteroffer is plausible, have the numbers from your new offer to hand. Nothing here requires you to have read the earlier weeks of this series, though the decision behind it will be steadier if you have.
Required tools
Any general-purpose AI assistant on any tier. The prompt is longer than the beginner version but still a single turn, and the optional roleplay extension works in ordinary chat with no extra features. Voice mode, if your tool has it, is genuinely useful for the rehearsal — saying the words out loud surfaces problems that reading silently does not. Given that you will be pasting candid material about your current employer, run it in a private or temporary chat and, obviously, not on a work device or a work account.
Frequently asked questions
Should I ever accept a counteroffer?
Sometimes, and the framework exists to make that a real question rather than a reflex in either direction. The useful test is whether the counteroffer addresses the reason you were leaving. If you were underpaid and the number fixes it, and nothing else was wrong, that is a coherent outcome. If you were leaving because of the work, the manager, or the trajectory, more money for the same situation is a solution to a different problem. Worth thinking about too: you have now shown your hand, and how your employer treats that over the following year is something you cannot know in advance.
How much detail do I owe them about why I am leaving?
Less than you think, and you get to decide the amount before the conversation rather than during it. A brief, true, non-grievance answer — the role, the scope, the schedule, the commute — is sufficient and complete. Exit conversations feel like a moment for candour, and occasionally they are, but feedback delivered while resigning is rarely received as feedback. If your organisation runs exit interviews, that is a separate conversation with different rules, and you can decide separately how frank to be there.
Can the AI tell me what my notice period is or whether my non-compete holds?
No, and the prompt is written to stop it from trying. Those answers depend on your contract, your jurisdiction, your role, and sometimes on case law that has moved recently. A model will produce something confident and general-sounding, and the confidence is precisely the hazard. Take your actual contract to HR or a lawyer for anything involving notice, non-competes, equity forfeiture, bonus timing, or benefits — the prompt hands you the list of questions to ask.
The letter it wrote sounds nothing like me. Is that a failure?
It is the expected first draft, and it is why the prompt ends by asking which lines sound least like you. Every draft here is raw material. Rewrite the letter in your own words, cut anything you would not say, and shorten it — the person receiving it has read your writing for years and will notice a voice that is not yours. That noticing costs you something real, because this document sits in a file next to the reference they will give you.
What if my employer walks people out immediately?
Then your handover happens before you resign, not after, which is exactly the sort of thing the "how does this organisation react" field is there to catch. Get your documentation in order, make sure nothing critical exists only in your head or your inbox, retrieve anything personal from your machine, and go in expecting your last day to be today. It is not a hostile assumption. It is common in some industries and it is much better to be ready for it than surprised by it.
Recommended follow-up prompts
- The landing package (Variation 3 of this post) — the full version, where the exit plan becomes one component of a single handoff from candidate to new hire.
- A reference-network prompt — a short prompt to list everyone at your current employer worth staying in touch with, what each of them would be able to say about your work, and what a reasonable check-in cadence looks like once you have left.
- A first-week logistics prompt — the unglamorous handover of benefits enrolment, final pay, expense claims and equipment return, built as a checklist with dates, and pointed at HR for anything it cannot answer.
Tags and categories
Tags:
resignation, counteroffer, notice period, exit planning, difficult conversations, professional references, job search, career transitions, intermediate prompts Categories: Career & Job Search, Workplace Communication
Citations
NOT APPLICABLE
The Landing Package
One handoff from candidate to new hire, built in one sitting.
Three artifacts, built separately, tend to contradict each other. The decision says you took the job for scope; the resignation script says you left for the money; the onboarding plan targets early wins in an area nobody who interviewed you thinks you were hired for. Nothing catches this, because each was made in its own afternoon, and the contradiction only surfaces in month four when someone asks what you came here to do and you give an answer that surprises you. The alternative is to build all three in one pass, with each feeding the next: the decision record supplies the reasons, the reasons shape the exit, and what you learned while interviewing becomes the intelligence behind ninety days of a plan. Then you break the whole thing on purpose, before reality gets the chance.
Why this matters now
The first quarter has an outsized effect on the rest of the job. Impressions form fast, they are formed by people with limited information, and they are stubborn — the read your new colleagues arrive at in week three tends to survive a great deal of contradicting evidence later. Meanwhile you are the only person in the building with a complete picture of why you came, and that picture starts eroding on day one under the weight of logins, onboarding videos and the meetings you are invited to for reasons nobody explains. Building the package while everything is still in your head is the only moment it can be done well.
Act as my transition analyst for the last mile of a job change. We are going to build one connected landing package in four phases. Do not skip ahead and do not begin a phase until I have confirmed the previous one. Everything you produce is analysis and draft material — anything a human at either company will read or hear, I rewrite in my own words before it leaves my hands.
Context I am supplying:
- Criteria I wrote before any offer existed: [PASTE — must-haves, deal-breakers, ranked priorities]
- The final offer, and any competing offer: [PASTE]
- What I learned about the team, the manager and the company while interviewing, including anything that worried me: [PASTE]
- My current role, and what leaving it costs me: [PASTE]
PHASE 1 — DECISION RECORD.
Build a weighted decision matrix. Use my ranked priorities as the criteria and assign each a weight out of 100 based on the order I gave; show me the weights and let me correct them before you score anything. Score each option against each criterion from 1 to 5, and beside every score name the specific evidence from my context that justifies it. No score without evidence — where you have none, mark it unknown and list what I would need to find out. Then give me the weighted totals, a one-paragraph verdict, and a sensitivity check: which single criterion, weighted differently, would flip the result. Close with a dated decision record under 200 words stating what I chose and why, written so that a version of me eighteen months from now, having a bad week, can read it and recover the reasoning.
PHASE 2 — EXIT PLAN.
Working from the decision record, produce: a resignation conversation script including the difficult branches, a letter draft, a counteroffer response decided in advance for a match, a beat and a title change, a handover outline for my notice period, and a list of every question touching my contract, notice, non-compete, equity or benefits — which you will not answer and which I will take to my documents, HR or a lawyer. Do not interpret contract terms or state notice or non-compete rules as fact. Attach no statistics to the counteroffer discussion.
PHASE 3 — LANDING PLAN.
Build a 30/60/90-day plan for the new role. Begin with a stakeholder map drawn only from what I told you: who decides, who influences, who holds the institutional memory, and who I have not met yet and should. Then for each thirty-day block give me learning goals, relationship goals, and one or two candidate early wins. Every early win must be visible, achievable without authority I do not yet have, and connected to something I actually learned about this team while interviewing. For each one, name the failure mode — the way it could make me look presumptuous rather than useful.
PHASE 4 — PRESSURE TEST.
Critique your own package. Name the three weakest assumptions in it. Identify any point where the decision record and the landing plan contradict each other. Name the single thing most likely to go wrong in my first sixty days given everything I have told you, and what early signal would tell me it is happening. Then compress the entire package into one page I can keep open on my first day.
Formatting: headed sections and short bullets throughout. End every phase with the line 'Confirm or correct before I continue.'
How the AI reads this prompt
Practical examples from different industries
A hospital operations manager with two offers and a counteroffer brewing.
She has three options and no clean way to compare them, since one pays more, one is closer to home, and her current employer has started making noises about a new role. Phase one produces a matrix where two of the three columns come back partly marked unknown, because nobody in her interviews described how the operations team is measured — which becomes the question she takes back to both hiring managers before deciding. The sensitivity check tells her the result flips if commute is weighted above development, and the moment she sees that written down she knows which one she actually means. The decision record she writes is the thing she reads in November, when the new job is hard for ordinary reasons.
A first-time engineering manager landing in an unfamiliar team.
He was hired to stabilise a group that lost two seniors in a quarter, and his interview notes are full of things people said carefully. The stakeholder map surfaces something he had noticed and not registered: the person who explained the deployment history was not the tech lead and was not in his reporting line, but has been there nine years and holds every piece of context about why the current architecture exists. Phase three puts a conversation with that person in the first two weeks. The early wins are deliberately small — the failure mode named beside each one is "arriving with a reorganisation," which is precisely what a new manager under pressure to show impact is most tempted to do.
A contractor going permanent inside a company they already know.
She has worked with this team for a year, which makes the landing plan look unnecessary and makes it more useful than average. The matrix forces her to separate what she likes about the people from what she knows about the role, which are different, and the sensitivity check reveals the decision rests almost entirely on a promise about scope that only one person made in one conversation. Phase three's stakeholder map is where the value lands: she discovers she knows one department well and three others not at all, and that her reputation as "the contractor who does the reporting work" is a starting position she now has to move deliberately.
Creative use case ideas
- A move to a new city. Same three phases: decide against criteria you set before you fell in love with a neighbourhood, exit your current place and relationships deliberately, and plan the first ninety days of building a life somewhere you know nobody.
- Taking over a community organisation. New chairs of school boards, clubs and neighbourhood associations inherit a stakeholder map they cannot see, complete with people who hold two decades of institutional memory and no title. The phase-three structure fits almost unchanged.
- Starting a degree or a serious course as an adult. The decision record is worth having when the second term gets hard, and the 30/60/90 shape works for building relationships with faculty and cohort just as well as with colleagues.
- Handing something over that you built. Run the exit phase alone when you are stepping away from a project, a codebase, or a role in a family business, and the handover outline becomes the artifact that decides whether your work survives you.
- Preparing to receive a new hire. Run the landing phase as the manager rather than the new hire, and you will have anticipated most of what they will need in their first month, including the institutional-memory introduction they will otherwise take five months to stumble into.
Adaptability tips
The phases detach cleanly. Run phase one alone when you are deciding, phase three alone when you have already started somewhere and never made a plan, phase four alone against any plan you built by other means. Compress the timeline when the job demands it — a startup role may be better served by 14/30/60, and saying so in the prompt is all it takes. If you are joining a team you will manage rather than join, add a fifth category to the stakeholder map for the people who will report to you and ask for their questions about you rather than yours about them, since that is the map that actually determines your first month. And if you have already accepted and resigned, skip phase two entirely, keep the confirmation gates, and spend the saved effort on a deeper landing plan.
Pro tips
- Run the whole thing before you resign, not after. Phase four regularly surfaces a question you should have asked before signing, and there is a narrow window where asking it is still easy.
- Paste your interview notes raw rather than summarised. Your summary has already discarded the ambivalent details, and the ambivalent details are what phases three and four run on.
- When a phase comes back wrong, correct the input rather than the output. The gate exists so that a bad weighting gets fixed at source instead of patched three sections later.
- Re-run phase four at day forty-five with what you have actually learned pasted in. The plan built from interview intelligence is a hypothesis; six weeks inside the building is the first real data you have.
- Keep the one-page compression somewhere visible for the whole first quarter. The failure mode of a good plan is not being wrong. It is being closed.
Prerequisites
You need real material to feed this — criteria set before the offer arrived, the offer itself, and genuine notes on what you learned while interviewing, including the parts that gave you pause. If you followed this series, that is Week 2's spec sheet, Week 7's negotiated outcome, and the dossier and diagnostic reads from Weeks 4 and 6. If you did not, reconstruct what you can and be honest with yourself about which parts are reconstruction, because the matrix inherits the quality of its inputs and presents it with far more confidence than it earned. Set aside about an hour, since the confirmation gates mean this is a working session rather than a paste-and-read.
Required tools
A general-purpose AI assistant with a reasonably long context window and a chat interface that keeps the conversation intact across turns, since the four phases build on one another. Paid tiers help here — not for capability so much as for context length and message limits, given that you are pasting substantial notes and running an hour-long session. Projects or custom-instruction features are useful if you want to keep the context available for the re-run at day forty-five. No plugins or integrations are required; everything happens in the chat.
Frequently asked questions
An hour seems like a lot for a job I have already accepted.
It is a lot, and it is smaller than the alternative. The first quarter sets a read on you that takes far longer than an hour to change, and the specific intelligence you are working from — what people said in interviews, what you noticed, who seemed to hold the real answers — has a short shelf life. Two weeks into the job it will have blurred into general impressions. If an hour is genuinely impossible, run phase three alone and keep the stakeholder map.
The matrix picked the option I did not want. Should I follow it?
No, and that outcome is one of the more useful things the exercise produces. A matrix is a structured way of examining your own criteria, not an oracle, and a strong pull against its verdict is information: either a criterion is missing, or a weight is wrong, or something you have not articulated matters more than everything you wrote down. Say what that is, adjust the model, and re-run it. The version that survives your honest objection is the one worth keeping.
Can I trust the scores if I mostly filled the context with impressions?
Trust them exactly as far as the evidence noted beside each one, which is why the prompt requires it. Cells backed by something specific — a stated policy, a number in the offer, something a named person said — are reasonably solid. Cells backed by a general feeling about the interview are your inference wearing a number, and the honest response is to mark them unknown and find out. A matrix full of unknowns is not a failed exercise; it is a very clear research list.
How is this different from asking for a 30/60/90 template?
A template gives you the shape without the content, and shapes are freely available. What makes phase three useful is that the early wins are constrained to what you specifically learned about this team, that the stakeholder map is built from your own interview notes rather than a generic org chart, and that every recommendation arrives with its failure mode attached. The plan is not better because it is more elaborate. It is better because it is about your actual situation.
What about the legal side of leaving — can phase two cover that?
It deliberately cannot, and it will hand you a list instead. Notice obligations, non-compete enforceability, equity and bonus timing, and benefits continuation depend on your contract and your jurisdiction, and they are also the questions where a confident wrong answer is genuinely expensive. Take that list to your own documents first, then to HR, and where money or restrictions are significant, to a lawyer. The scripts handle the human conversation; the paperwork is not the AI's call.
Recommended follow-up prompts
- A day-45 review prompt — paste your actual first six weeks against the plan you built, and ask which assumptions held, which stakeholders turned out to matter more than mapped, and what the next forty-five days should change.
- A first-90-days manager alignment prompt — turn your early wins into the three things you want your new manager to say about you at the end of the quarter, then work backwards to what they would need to have seen.
- Week 1's restlessness audit — worth re-running a year into the new role, not because you are leaving, but because the criteria that made this decision are the ones you will want on file if you ever are.
Tags and categories
Tags:
30/60/90 plan, onboarding, decision matrix, stakeholder mapping, weighted scoring, decision record, early wins, resignation, career transitions, advanced prompts, multi-phase prompting Categories: Career & Job Search, Onboarding & Leadership
Citations
Watkins, M. D. The First 90 Days: Proven Strategies for Getting Up to Speed Faster and Smarter. Harvard Business Review Press — the standard reference for structured transitions into a new role, and the origin of the 30/60/90 framing this variation adapts.
Which of the three should you use?
The three prompts split the last mile into its three failure points, and they escalate in ambition rather than in length. The Cold-Light Offer Check is a single-turn reality test: you supply criteria and an offer, and it tells you where they fail to meet and what you have not asked yet. It is the right choice when the decision is genuinely close, when you have twenty minutes rather than an evening, or when you have never run a structured comparison before and want to see what the technique feels like at low stakes. The Exit Rehearsal is the operational one — it assumes the decision is made and turns attention to the conversation that follows, which is where careful people most often improvise badly. Its distinguishing feature is the parameter block: it is built to be run more than once, at different settings, which makes it the most reusable of the three across a career.
The Landing Package is different in kind rather than degree. Where the other two produce one artifact each, it produces three that are checked against one another, and it adds a phase whose only job is to attack the work. That coherence is the point — the decision record supplies the reasons, the exit inherits them, the landing plan is built from the intelligence you gathered while interviewing, and phase four asks whether any of it contradicts itself. It also costs the most: an hour, real interview notes, and a willingness to be interrupted by confirmation gates.
Choose by where you actually are. Still deciding, start with the first. Decided and dreading Monday's conversation, go straight to the second. Signed, with the interview details still fresh and a first day on the calendar, run the third and take the pieces you need. They also stack in sequence over about a week, which is roughly how the last mile of a job change unfolds anyway. One thing worth noticing on the way out of this series: none of these prompts would work as well as they do without the artifacts the earlier weeks produced — the criteria from Week 2, the dossier from Week 4, the diagnostic reads from Week 6, the outcome of Week 7. Eight weeks of work, each week consuming the last, is what let this week's prompts be about judgement rather than about gathering. That is not a claim about AI. It is just what happens when the work you do is built to be picked up again.
TAGS:
