Career & Job Search prompt of the day

Achievement mining: let the model interview you until vague duties become measurable results

Your resume lists duties, not results, because you never wrote the numbers down. Have the model interview you, one question at a time, until each role yields results you can actually stand behind.

You are a career coach running an achievement-mining interview. My resume reads like a list of duties; your job is to pull real, verifiable results out of my memory by asking good questions, not by suggesting results for me.

Process:
1. I paste one role at a time (title, dates, the bullets I have now).
2. You ask ONE question at a time, waiting for my answer. Ask about scale (how many, how often, how much), before-and-after (what was it like before I touched it), what would have happened without me, who noticed, and what I am proudest of that is not on the page.
3. When I give a number, ask how I know it — a report, a dashboard, an email, my own estimate. Label each number by its source.
4. After eight to ten questions on a role, stop and produce the output below for that role, then ask whether to move to the next one.

Output per role:
A) RESULT BULLETS — five bullets in the pattern action + scope + result, built only from my answers. Each number carries its source label: (documented) or (my estimate).
B) THE EVIDENCE LIST — what I should go and find to back up each estimate before it goes on a resume: an old report, a performance review, a thank-you email.
C) THE STORIES — which two results would make the best interview stories, and the one detail each still needs.

Inputs: [ROLE TITLE + DATES] · [CURRENT BULLETS FOR THAT ROLE] · [TARGET ROLE I AM APPLYING FOR]

Rules: Do not invent results, numbers or outcomes, and never suggest a figure for me to agree to — ask, then write down what I say. Keep estimates labelled as estimates; I will verify them before they go on a resume. If I cannot remember a number, write "[FIND THE NUMBER]" rather than rounding to something plausible. Never paste confidential information from a current or past employer into a consumer AI tool, and keep names and personal identifiers out of the output beyond what I provided.

Why it works — People write duties instead of results because nobody asked them the right questions at the time, and a blank page does not ask. A one-question-at-a-time interview recovers the scale and before-and-after detail that memory holds, while labelling every number as documented or estimated, and listing the evidence to find, keeps the resulting bullets defensible when a hiring manager asks how you measured it.