TL;DR
- You don't need tidy notes or a great memory — just raw material in one file.
- Mine your existing AI chat history (ChatGPT, Copilot, Gemini, Claude) for work you've already described in real time.
- No chat history in your tool? Use your commit history, PRs, and code review comments instead.
- Ask a former manager, skip-level, or HR contact for a short written reflection — and be explicit about leaving out confidential info and others' personal details.
- Paste everything into one file, mess and all — text, PDF, Word, Markdown, or JSON all work.
- Bulk import splits it into dated, individual stories, keeps your original wording, and lets you review before anything saves.
Most people sit down to write a career story file and freeze — not because they haven't done good work, but because none of it is written down anywhere in one place. It's scattered across old chat threads, half-remembered conversations, a performance review from two roles ago, a Slack message someone sent you that you never saved.
The good news: you don't have to reconstruct any of it from memory. Bulk import doesn't care how messy the source is — paste it in as one long ramble, upload a PDF, a Word doc, a Markdown file, or a JSON export, and it'll split it into individual, dated stories for you. Your only job is getting the raw material into one file. Here are the two fastest ways to do that.
1. Mine the AI assistant you already use
If you've spent the last year (or few) working alongside ChatGPT, Microsoft Copilot, Google Gemini, Claude, Claude Code, Cursor, or GitHub Copilot, that history is a goldmine. You've probably described real problems to it in real time — what broke, what you decided, what you shipped — long before you ever thought to write any of it down for yourself.
Open the tool tied to how you actually work — Copilot if your company runs on Microsoft 365, Gemini if you're in Google Workspace, Claude or ChatGPT if that's your daily driver, Cursor or GitHub Copilot if most of your work lives in code — and ask it to go looking on your behalf.
Copy this prompt
"Looking back at our conversation history together, help me reconstruct a record of my recent work. Go through everything we've discussed — projects, debugging sessions, decisions, documents I've asked you to help write or review — and pull out every moment that shows real professional work, especially:
— Wins: things I shipped, fixed, launched, or got approved
— Achievements: measurable outcomes — improved numbers, saved time, unblocked a team, closed a deal
— Challenges: hard problems I worked through, including ones that were technically or politically difficult
— Conflicts: disagreements I navigated, tradeoffs I had to defend, or stakeholders I had to align
For each one, give me a rough date or timeframe if you can tell, what the situation was and why it mattered, what I actually did, and what happened as a result — numbers if there are any. Don't clean it up or make it sound impressive, just give me the raw, specific version in plain language. Skip anything vague or generic; I only want moments where you can point to something concrete I did or decided."
A couple of things worth knowing before you run it. Tools with persistent memory across conversations (ChatGPT, Copilot, Gemini, Claude) can usually answer this directly. Tools that don't keep long-term history the same way, like Cursor or GitHub Copilot in an editor, won't have a chat archive to search — for those, your commit history, pull request descriptions, and code review comments are the equivalent record. A quick "summarize my commits and PRs from the last six months into the same wins/challenges/achievements format" pointed at your own git log works just as well.
Copy whatever comes back, however long or rough, straight into a text file. That's your first draft — don't edit it yet.
2. Ask the people who watched you work
For older roles, or work that never touched an AI tool, the best record lives in someone else's memory — usually a former manager, skip-level, or HR partner who wrote your reviews or nominated you for something. A short, easy-to-answer email gets better results than a vague "can we catch up sometime."
Copy this email
Subject: Quick favor — helping me remember my time on [team/project]
Hi [Name],
I'm putting together a record of my career for [a performance review / a job search / a promotion case — whichever is true], and I'd love your perspective on my time on [team or project], from around [rough dates].
If you have a few minutes, could you share anything that stands out — a project that went well, a hard problem I helped solve, feedback from a past review, or a moment you remember me handling well? Even a couple of sentences would help a lot.
One ask: please leave out anything confidential to [Company] — financials, strategy, client names, or anything that wouldn't be appropriate to share outside the company — and please don't include anyone else's personal details. I'm only looking for your take on my work, not anyone else's information.
No pressure at all if this isn't the right time — I really appreciate you considering it either way.
Thanks,
[Your name]
Say this part out loud, don't just imply it: asking explicitly for confidential business details and other people's information to be left out makes it easy for them to say yes, and protects them as much as it protects you.
Whatever comes back — three bullet points or three paragraphs — drop it into the same file as everything else.
3. Get it into one file, mess and all
Don't worry about formatting, order, or whether it reads well. Paste the AI export and the email reply into the same document, keep any dates you have, and leave the rest as-is. Structure doesn't matter here — paste it directly, or save it as a PDF, Word doc, Markdown, or JSON file and upload it. Bulk import will split it into separate stories, keep your original wording alongside a STAR breakdown for each one, and let you review and prune before anything is saved.
The only real prep work is getting the raw material down in one place. Everything after that, the tool does for you.
Frequently asked questions
What file formats can I use for bulk import?
Pasted text, PDF, Word (.docx), Markdown (.md), or JSON — whatever format your source material is already in. There's no need to convert or reformat anything first.
What if I don't have any notes or old documents to start from?
Start with your AI chat history — most people have more of it than they realize — and fill in the gaps by asking a former manager or HR contact for their recollection of your work.
Is it OK to ask a former manager for input like this?
Yes, as long as you're explicit about what to leave out: anything confidential to their company, and any other employee's personal details. Most people are glad to help when the ask is this specific and short.
Will the AI invent details if my source material is messy or incomplete?
No — bulk import keeps your original wording (shown as "Original story") alongside a structured STAR breakdown for each entry, and you review every story before anything is saved. Nothing is fabricated.
Related reading: How to build a brag doc with AI, for keeping this kind of record going year-round instead of rebuilding it from scratch each time.