Skip to content
How Premise works

Built from your words, and you can check every step.

Premise is one short loop you don’t have to keep up with. Each part starts from what you actually said and did, marks what’s still a guess, and builds on the last pass instead of starting over.

01The loop

One short loop you repeat. Nothing to maintain.

Every session moves through the same six steps. Each pass starts from what you actually said and did last time, so the guidance builds on itself instead of starting over.

Signal01
You share what’s unclear and what’s in the way.
Sensemaking02
Patterns reflected; evidence split from assumption.
Options03
Two to four realistic paths, with honest tradeoffs.
Experiment04
One small action with an if-then plan.
Reflection05
Mark it: tried it, unclear, or didn’t happen.
Profile06
What you learn updates a profile you can correct.
02Direction Snapshot

Your first read: your own evidence, sorted and shown.

After a short intake, Premise reflects back what it heard and separates evidence from assumptions to check. No score, no personality label; every line is something you can correct.

Direction Snapshot · Maya
Direction Snapshot
What I heard

You like making confusing things clearer for people, but you're worried a “helping” path won't pay, and you feel behind friends who already have a plan.

Evidence & signals
Strong signal

re-explained a chemistry unit to two friends “because the teacher made it harder than it was”

Evidence

rewrote the robotics-club sign-up steps so people stopped asking the same questions

Assumption · to check

that “helping work doesn't pay” (a worry, not a checked fact)

Paths that fit

Learning design

Fits the “make it clearer” thread; keeps creative and stable options open.

UX / product

Same instinct, stronger income range; more competitive entry.

One thing to try this week

Make one short explainer for someone who's stuck.

About an hour, low stakes. It tests whether the pull is teaching, design, or just helping, before you commit to a label.

03Mentor chat

Advice that has to earn its place before you see it.

The mentor reflects ambivalence, asks before sensitive areas, and marks uncertainty out loud. A recommendation only appears once it can explain why it fits, the evidence it used, the tradeoff, and one concrete next step.

Mentor · Maya
You

everyone already seems to have a plan. i like helping people but i'm scared a “helping” path won't pay, and my family can't cover much. i don't know where to start.

That's a real place to start from. Let's separate what you've actually shown from what's still just worry, and find one small thing worth testing.

You

okay. i guess i do end up explaining things to people a lot. is that anything?

Yes, and you've done it more than once, which is the part worth paying attention to.

Why this fits
You keep making a confusing thing clear for one specific person, twice now, unprompted.
What your words show
Strong signal
re-explained a chemistry unit to two friends “because the teacher made it harder than it was”
Evidence
rewrote the robotics-club sign-up steps so people stopped asking the same questions
What it tests
Whether the pull is the teaching, the design, or just being useful, before you commit to a label.
Next step
This week: pick one person, one thing they're stuck on, and write the explainer you wish someone had handed you.
04Living Profile

A profile you can see, question, and correct.

Everything Premise believes about your direction is stored as typed records (evidence, assumptions, open questions), kept strictly separate. Assumptions are marked “to check,” never treated as fact, and any line is yours to fix.

Living Profile · Maya
At a glance

You keep gravitating toward making confusing things clearer for people, and we're testing where the energy actually is before picking a direction.

The record
Strong signal

re-explained a whole chemistry unit to two friends, unprompted

Inference · tentative

the thread looks less like “math” and more like making a confusing thing clear for one person

Assumption · to check

that “helping work doesn't pay,” marked to check against real ranges

Open question

would clarity work be the job, or stay a thing you do on the side?

Teach one tiny thing
ChosenPlannedTried

Pick one person stuck on one thing this week. Make a 5-minute explainer. ~1 hour, low stakes.

05Action experiments

One small, real thing to try, with a clear plan and a fallback.

No “do research” and no invented passion projects. You get a single low-stakes action sized to your time, with steps and a fallback, so trying it tells you something either way.

Action experiment

Teach one tiny thing to one person this week.

~1 hr
Why it fitsYou keep helping people understand things.
What it testsWhether the energy is teaching, design, or just helping.
The planIf they get it faster than before, note what you changed.
06Reflection loop

Mark how it went, and the next step actually changes.

After you try the action, you mark the outcome, and Premise shows what that changes about what to test next. Reflection that changes nothing reads as homework, so it never does.

How did it go?
Tried itUnclearDidn’t happen
This changes what we test next

If it felt good, we go deeper on teaching. If it didn't, we keep the thread and try a different route: same theme, new step.

Where it draws the line

Some things it won't do, on purpose.

The same gate runs on every output: it names why it fits you, the evidence it used, what it tests, the tradeoffs, and a concrete next step, or you never see it.

  • No admissions chances

    It never predicts your odds of getting in anywhere.

  • No essay writing

    It won’t draft or edit application essays. Your words stay yours.

  • No fake projects

    No “start a nonprofit to look good.” Only real things worth trying.

  • No personality labels

    No types, no destiny claims, only evidence you can check.

Start where you are

Begin with the part that feels unclear right now.

The free loop does real work on its own: one snapshot, a few mentor replies, one action, one reflection. No card to start.

You don't have to know what you want yet. Start from the part that feels unclear, and let the loop do the rest, one honest step at a time.