What Is a Story Bank?

There's a story you tell. You know the one. The people who love you can see it coming from three sentences out, and they let you tell it anyway, because it's yours and it's good and it has a shape now that a decade of retelling carved into it.

Now open your AI tool of choice and ask it to draft something in your name. Watch what shows up instead: a nameless client, a lesson learned "early in my career," an anecdote with no city in it and no weather. Your best material, the stuff a room full of people has already voted for, appears nowhere.

That's not a model problem. That's an inventory problem. The stories live in you, and nothing you've ever done has put them anywhere a machine can reach.

That's the job of a Story Bank.

What is a Story Bank?

A Story Bank is a short document that holds the stories you tell more than once and the phrases you always reach for — the hiring mistake you still think about, the line your grandfather used, the moment on a pier that explains half your decisions since. It's one of the six context anchors, alongside your Personal Constitution, Writing Codex, State of the Union, Timeline, and Influence Roster.

The other anchors tell an AI who you are, how you sound, and where you're standing. The Story Bank is different. It's the raw material — the lived evidence a model pulls from when it needs to sound like it has actually been somewhere with you.

Why your AI writes like it never lived anywhere

When a model needs an example, it invents one. Not because it's lazy. Plausible is all it has, because you never gave it true. So it reaches for the statistical average of every business anecdote ever written, and the average is a composite with no proper nouns, nothing you could be cross-examined on.

The thing is, specificity is where believability lives. "A client taught me the value of listening" is furniture. The version with a name, a city, and weather in it has a pulse. The difference between the two is whether the writer had somewhere real to reach.

A model with your Story Bank loaded stops inventing. Asked for an example, it pulls the real one — the actual city, the actual kitchen, the line your mother actually said — and the draft reads like someone who has lived alongside you, because in a narrow, useful sense, it has.

Isn't a Writing Codex enough?

No, and the gap between them is worth being precise about, because they get collapsed constantly.

A Writing Codex captures how you sound: your rhythm, your sentence habits, the words you'd never use. It's the instrument. The Story Bank is the songbook — what you actually have to say, accumulated over a life.

Load only the codex, and you get something strange: well-formed emptiness. The cadence is yours, the punctuation is yours, and the content is nobody's. I've written about why AI defaults to the smooth average when it has no reference for your voice. The same collapse happens one layer down with material. No bank, and the model averages its way through your life story too.

You need both. One teaches the how. The other supplies the what.

What actually goes in one?

Three layers, and the first has a filter built in.

The stories you retell. Run the repeat test: anything you've told more than once (at dinner, in a job interview, to your kid in a rental car) has already survived an audition in front of a live audience. Retelling is curation you did without noticing. Each entry gets a few plain sentences: what happened, where, and why it stays with you. Write it the way you say it out loud, not the way you'd polish it for print.

The phrases you reach for. The toast you always give. The thing your mom said when you were about to do something dumb. The expression your friends would put on your tombstone. Signature language is material, and it's the fastest thing in the file for an AI to use well.

The patterns. This layer you can't write on day one, which is part of the fun. After a dozen entries, read the whole bank in one sitting. Shapes emerge that no single story shows.

The part you can't predict

That third layer is where the Story Bank stops being a filing cabinet and starts being a mirror.

A new entry never lands in an empty file. It lands next to everything you've already banked, and proximity does something no single story can: the themes start showing their shape. You notice that half your stories are about walking away at the right moment. Or that every phrase you love came from the same person. Or that the settings keep repeating, and the settings are trying to tell you something. The stories usually know it before you do.

So the bank pays you twice. Once as material for the writing, and again as a map of your own recurrences — the themes you circle, the wells you keep drawing from. An AI that knows those patterns doesn't just retell your stories. It knows which one you'd reach for next.

Where to start

Tonight, at whatever table you eat at, notice the story you tell. That one goes in first.

Then give it twenty minutes: list five more that pass the repeat test, write each in a few flat sentences, and add two or three phrases that are unmistakably yours. That's a working Story Bank. Feed it as material surfaces, and if the excavating feels hard, that's what context mining exists for — the questions do the digging. You can build yours at /anchors/story-bank.

And then, some ordinary morning, you'll ask an AI to help you write something that needs an example: a talk, a post, a hard email, a eulogy. The draft will come back carrying the story your family can see coming from three sentences out. It won't have to invent a thing.

The good stories wait to be told again. The bank just makes sure they're there when you come looking.