Stay-or-Go Questionnaire

A sample AI Analysis

What the optional AI Analysis writes, shown exactly as it was generated for a fictional person who answered this questionnaire.

Who it is for

This fits the long stretch where the question has stopped being an occasional thought and become the weather — a year of half-deciding, a case you argue with yourself on the drive home, a conversation you've had with friends more often than with your partner. It suits you if you'd rather stop making it in fragments and set the whole thing down in one place. It isn't built for a decision you have to make this week. And if fear for your safety is part of what's keeping you, that needs support beyond a questionnaire, not a longer look at your own answers.

What you get out of it

Seventeen sections, and the order is part of the design: it opens by asking where you are today and closes by asking where you stand after answering. Twenty-seven of the 128 questions are marked Core, so the full shape of it is within reach well before you have answered all of it. What you come away with:

• A ledger kept in three parts. What you bring, what your partner brings, and the dynamic between you get separate sections, so a pattern the two of you make together doesn't quietly get filed under one person's name.
• A relationship broken into its actual parts. Communication, emotional intimacy, physical and sexual intimacy, money and the household, and the world outside the two of you each get a section — so "it's not working" resolves into which parts, and which are fine.
• An honest account of the history. How you got here, what resentment and hurt are still being carried, and what the two of you have already attempted — which is the difference between a thing untried and a thing tried and not held.
• Your own standard, written down. Sections on deal-breakers and non-negotiables, on the future the two of you actually share, and on what good would look like — a description specific enough to measure a relationship against, rather than a mood.
• Both costs, weighed in the same place. One section holds the cost of staying next to the cost of leaving, so neither gets assessed on its own — which is how one of them ends up looking smaller than it is.

What’s inside

17 sections, in this order. The first is free.

1. 01Where I Am TodayFree
2. 02How We Got Here
3. 03What I Bring to This Relationship
4. 04What My Partner Brings
5. 05The Dynamic Between Us
6. 06Communication
7. 07Emotional Intimacy
8. 08Physical and Sexual Intimacy
9. 09Money, Household, and Daily Life
10. 10Family, Friends, and the World Outside
11. 11Resentments and Wounds
12. 12What We've Already Tried
13. 13Deal-Breakers and Non-Negotiables
14. 14Shared Vision
15. 15What Good Looks Like
16. 16The Cost of Staying, The Cost of Leaving
17. 17Where I Stand After Answering

How it compares

Questions people ask

Will the Stay-or-Go Questionnaire tell me whether to stay or leave?

No. There's no verdict at the end and no recommendation — the final section asks where you stand, and that sentence is yours to write. What it does is make sure the answer isn't resting on the two or three parts of the relationship that happen to be loudest this month, by taking you through seventeen of them.

How is this different from making a pros and cons list?

A list gets written by whichever mood you're in when you sit down, and it tends to come out at the same six items every time. This is 128 questions across seventeen sections, and the sections you'd have left off your own list are in it: what you bring, what's already been tried, what good would actually look like.

Why does it ask what I bring, when the problem is my partner?

Because a decision made on half the ledger tends not to hold. There is a section for what your partner brings, and a separate one for the dynamic between you — the part neither of you produces alone. Keeping the three apart is what stops a pattern being mistaken for a person, or a person for a pattern.

What if I already know my answer and just can't act on it?

Then the later sections are probably where the value is. What you've already tried, your own deal-breakers, and the cost of staying set beside the cost of leaving are less about finding a position than about seeing what's holding one in place. Sometimes the obstacle turns out not to be the relationship.

Should I take this if my relationship is basically fine?

Probably not this one. It assumes the question is already live — that some part of you has been carrying it long enough to want it written down. If nothing is wrong and you're just curious how things are actually going, the Relationship Satisfaction Quiz is shorter and better aimed at that.

What if I'm afraid of my partner, or leaving feels unsafe?

Then a questionnaire is the wrong first step, and answering one honestly where you live may not be safe either. The National Domestic Violence Hotline is 1-800-799-7233, or thehotline.org, and they help with planning, not only with crises. That is a conversation to have with a person.

Can I give this questionnaire as a gift?

Yes. Sign in, open your dashboard and choose "Gift access" on this questionnaire, then enter the person's email address. They get an email with a link to claim it, and it arrives as soon as you have paid, so it is never late. Their answers are theirs: nothing they write is shown to you.

The optional AI Analysis

Answering the questionnaire is the part that does the work. The AI Analysis is what you can do afterwards: it reads the answers you wrote and gives them back to you as structured reflection notes — an overview of the whole questionnaire, then one entry for every section you answered. Your unlock includes 3, and it is genuinely optional. Nothing about the questionnaire depends on running one.

Each entry has three parts, and they are doing different jobs rather than restating each other:

What you were clear about

The things your answers state plainly, named back to you in a sentence or two. This is the part people underestimate: a lot of what you know about your relationship is already in what you wrote, and has never been said in one place.

The patterns underneath

What you do repeatedly, rather than what you did once. The same behaviour surfacing in sections you would never have put side by side; an answer near the end sitting against one from the beginning; the thing you circled several times without ever settling it. Where a questionnaire asks you to sum yourself up at the close, the analysis also reads that summary against what your earlier answers actually describe — and says so, gently, when the two do not quite meet.

A few things to sit with

Two to four short prompts. Not instructions, not advice, and not a plan — questions worth carrying into the next week, or into a conversation with someone qualified to have it.

• Each questionnaire has its own analysis prompt, written for it. The five questionnaires do not share one general-purpose instruction with the title swapped. Each has its own, built around what that instrument is actually asking and what its sections are for — so the Stay-or-Go analysis reads for the shape of a decision, and the Long-Distance one reads for whether a plan and a timeline hold up. It is the part of this that took the longest and is the hardest to copy.
• It reads your answers, not your category. What reaches the model is what you wrote, the question each answer belongs to, and the section it sits in — in the order you answered them. There is no profile, no demographics, and nothing about you beyond your own words.
• It is written as reflection notes, not as advice. Considered and unhurried — closer to notes someone made while reading your answers carefully than to anything that answers back. It does not tell you what to do, and it is not a diagnosis.
• Answer only the Core questions and you still get one. The analysis covers what you actually answered, so the shortest honest path through the questionnaire produces a real analysis rather than a thin one.
• You can change your mind. Edit your answers, resubmit, and run it again — each run costs one credit. Resubmitting answers you have not changed costs nothing.
• Your answers reach the model without your name, email address, or any identifier for your account. Anything you wrote is sent as you wrote it, so leave out names you would rather not include. Your answers are stored encrypted, and nobody is reading them alongside your identity to improve anything.
• On a Partner questionnaire you each get your own. Two people can run an analysis on the same set of answers and neither sees the other's — they are generated and stored separately.

Every analysis also carries the questionnaire's own safety note, with its resources, whether or not anything in your answers suggested it was needed. That is deliberate: a resource that only appears when something decides you need it is a resource that can fail to appear.

Guides on this