You finally sit down to do the thing everyone keeps telling you to do: give your AI some context about who you are. You open a blank document. You type your name. Your job. Maybe the project that's eating your week. And then the cursor just sits there, blinking, because everything else that actually matters about how you work is stuff you have never once put into words.
That's the wall. Not laziness, not writer's block. The problem is that the most useful things about you are the things you'd never think to write down, because to you they're just the water you swim in.
Context mining is how you get past that wall.
What context mining actually is
Context mining is a guided process for surfacing the personal context an AI needs to actually know you. Instead of a blank page, you answer curated questions across the durable dimensions of who you are, and your answers become the raw material for your context anchors.
That's the whole idea. The rest of this is just what makes it work when a blank document doesn't.
Why the blank page fails
Ask someone to describe themselves and they'll give you the résumé: title, industry, a project or two. All true, all nearly useless to a model, because none of it is the stuff that makes your work sound like yours.
The valuable context lives one layer down. You think in analogies and reach for a story before you reach for a stat. You'd rather be told the hard thing straight than be managed around it. You're three weeks from a move and every decision is quietly running through that filter. None of that shows up on a blank page, because you've never had a reason to say it out loud. It's too obvious to you to notice, and it's exactly what an AI can't guess.
So the blank page doesn't just make you work harder. It makes you write down the wrong things — the facts you happen to remember instead of the truths that actually shape how you operate. You end up describing the outside of yourself and leaving the engine unmentioned.
What makes it mining
Mining isn't collecting what's lying on the surface. It's going after what's buried, in a way that's structured enough to actually surface it. Context mining earns the name by working the same way.
It's guided. You answer questions instead of inventing structure. A good question does half the work for you — it points a light at one specific thing and asks you to say what's there, so you're never stuck on "tell me about yourself" with a blank stare.
It goes after the non-obvious. The questions are built to reach the stuff you'd never volunteer. Not "what's your job" but the shape of how you think, the words you'd never be caught using, the thing you're carrying this season that colors everything else.
It's structured across dimensions. You don't mine "who you are" as one impossible lump. You mine it one durable dimension at a time — your values, your voice, your situation, your history, your people — so each pass has a clear edge and a clear question.
It produces raw material, not a finished essay. Your job is to answer honestly. Turning those answers into a clean, reusable document is the next step, not your problem to solve mid-sentence. Mining gets the ore out of the ground; refining it into an anchor comes after.
What context mining is not
It's not a survey. A survey collects the facts it already knows it wants. Mining is after the durable, hard-to-see material a form would never think to ask for.
It's not a brain-dump. A brain-dump gives you a pile of whatever surfaced, in no particular order, still shapeless. Mining is directed — each question has an edge, so what comes out is already sorted into the dimension it belongs to.
It's not you performing for a chatbot. You're not trying to sound impressive or complete. The best answers are the plain, slightly boring true ones, because that's the layer a model actually needs.
It's not a one-time setup. Context goes stale. You finish the project, you move cities, your priorities shift and your voice sharpens, and the answers you gave in January want a light re-mine by summer.
The six dimensions you mine
"Who you are" is too big to mine in one pass, so RUMO breaks it into six — each one a durable question worth answering about yourself.
Together they're the seam: what you stand for, how you sound, what you've lived, where you are now, where you've been, and who's in your orbit. There's a full tour of the six here if you want it. You don't have to mine all six to start — one honest pass beats a blank chat every time, and it's one of the most durable ways to give AI context because it's the material you own.
Why the questions do the work
Hand a model nothing and it writes for the average of everyone who ever typed a request like yours. The average sounds like no one, which is why the output comes back competent and faceless and not quite you.
The questions fix that at the source. A prompt tells the model what to do right now. Mining tells it who it's doing it for — it's the difference between giving your AI instructions and giving it direction. Answer the questions once and the model stops guessing about you and starts working from the real thing. It didn't get smarter. It got informed.
Where RUMO fits
I built RUMO because I kept hitting the blank page myself. I knew my AI would work better if it actually knew me, and I'd sit down to write it all out and get about as far as my name before the shape of the thing defeated me. The parts I most needed to hand over were the parts I'd never once thought to say.
So I made the process I wished existed: guided questions across the six dimensions that pull the durable stuff out of you, then compile your answers into structured context anchors you can drop into any AI you open. The Personal Constitution — your values, mined and made into a document — is free, and it's the one to run first.
Start with one dimension
Go back to that blinking cursor on the blank page, the one that knew you needed to do this and had no idea where to begin. The reason it stalled wasn't you. It's that nobody handed you the questions.
So let someone hand you the questions. Pick one dimension, answer honestly, and let the process turn what you'd never think to write down into something an AI can finally read.
Start with the free one. Thirty minutes this afternoon, and the next thing you ask your AI gets answered like it actually mined who you are.
Frequently Asked Questions
- What is context mining?
- Context mining is the guided process of surfacing the personal context an AI actually needs to know you. Instead of facing a blank page, you answer curated questions across six durable dimensions of who you are, and your answers become the raw material for your context anchors. The term was coined by RUMO, and the questions are built to pull out the things you'd never think to write down on your own.
- Why can't I just write my own AI context in a blank document?
- You can, but a blank document assumes you already know what to write and how to structure it — and the most useful context is usually the stuff that's too obvious to you to notice. You don't write down that you always think in analogies or that you hate being managed, because to you it's just water. Context mining works the other way around: it asks the questions that surface what you'd never volunteer, so the picture ends up complete instead of whatever happened to come to mind.
- Is context mining the same as filling out a survey?
- No. A survey collects facts it already knows it wants — name, role, industry. Context mining is after the durable, hard-to-see stuff: how you actually think, what you sound like, what you're carrying this season, who's in your orbit. The questions are designed to make you say the thing you didn't know was worth saying, then turn those answers into structured documents an AI can use.
- What does context mining produce?
- Context anchors — short, structured documents that each capture one durable part of who you are. Context mining is the extraction; the anchor is the refined result. You mine across six dimensions and walk away with reusable documents you can drop into any AI tool, instead of a pile of notes you have to shape yourself.
- How do I start context mining?
- Pick one dimension and answer its questions honestly — your values are the best place to start. RUMO's guided process walks you through the questions and compiles your answers into a finished context anchor. The free Personal Constitution is the fastest way to feel what mining one dimension does, and most people finish it in about thirty minutes.
