Prompt
Along

You're a practical build partner, not a coach. I have five free working days and I want to use them to make one thing I'd love to have made, for home or for work. It doesn't have to be software and it doesn't have to be impressive. It has to be something I'd actually finish.

Ask me six questions, one at a time. Ask one, then stop and wait for my answer. In your own words, cover: what I already make or fix for people without being asked; what I keep meaning to sort out; who it would be for; what tools I already have; what I'd enjoy even if it didn't work; and what "finished" would look like on Friday. If I get stuck, give me one short example and ask again. I can say "pass".

Don't rate or comment on my answers. Go straight to the next question. Keep every reply under 60 words until the end. No praise, no headings, no bullet points. No suggestions until all six are answered. Plain British English. Don't ask for names, confidential work or anything private; if I start sharing it, ask me to use a placeholder.

Then reflect back the two or three things you heard me care about most, a sentence each. Ask me to name the one thing I'd build. If I can't, offer three rough starting points and ask me to pick one and put it in my own words. Don't choose for me. Then ask what I'd do in the first hour of day one. Once I've said it, write a five-line brief in my words: what I'm building, who it's for, what done looks like, what I won't build, and my first hour. Then stop. Start with question one only.

I want you to know when I work badly and when I work well, so you're more useful to me. Don't diagnose me, label me or tell me what I seem to be feeling. This isn't therapy. It's working conditions.

Ask me seven questions, one at a time. Ask one, stop, and wait for my answer. Cover: a recent time I put off something that mattered; what happened just before; what I reached for instead (my inbox, a meeting, more research, my phone, you); what about the task was uncertain; a recent time I worked unusually well; what was different; and which of those conditions I could change. I can say "pass". One short follow-up per question at most.

Don't comment on my answers, just ask the next question. Keep every reply under 40 words until the end. No praise, no bullet points. Plain British English.

At the end, ask me to finish three sentences in my own words: "I tend to put things off when..." "The first sign is..." "I work best when..." Tighten my wording only after I've agreed the meaning. Then tell me one specific thing you'll do differently when I bring you a task I've clearly been avoiding.

Then ask if I'd like you to remember those three sentences. Only say you've saved them if you genuinely have a memory feature and have used it. If you don't have one, or I say no, tell me in one line where I could keep them so they're in front of you next time (custom instructions, saved information, a project, or a note I paste in at the start of a chat). Start with question one only.

I'm avoiding a decision. Help me look at it without making it for me. First ask me to describe it in one or two sentences, without names or anything confidential. Then ask me these four questions, addressed to me, exactly one at a time, waiting for my answer each time.

One: procrastination isn't the problem, so what's the uncertainty in front of you that's making you want to leave the room? Two: have you given this any space, or only more attention? (Space means stepping away and letting it sit. Attention means more research, another opinion, another framework.) Three: how much of what you're waiting for is real information that would change your move, and how much are you waiting so you don't have to feel uncertain? Four: what would you do if you weren't thinking about how it looked?

Don't diagnose me, interpret my answers, reassure me or add questions of your own. Keep every reply under 40 words. No bullet points. Plain British English. If I say I don't know, or pass, ask the same question once more in plainer words and ask for my best guess. Then move on.

At the end, don't summarise. Ask me to choose one move: a deliberate pause with a date I'll come back on; one named piece of missing information I'll go and get; or the smallest useful thing I could do before the end of tomorrow. Don't choose for me. When I've chosen, ask me once whether this move could be another way of avoiding it. Then ask me to confirm or change it, and stop. Start by asking for the description only.

Help my team design one small AI experiment we could run this week. Not a pilot, not a rollout. An experiment with four conditions: safe to get wrong; finished by Friday; something we'll observe, chosen before we start; done together and talked about afterwards. If I don't manage a team, treat "team" as the two or three people I work with most.

First ask me two things, one at a time: which AI tools we're allowed to use at work, and what kinds of information we're allowed to put into them. If I'm not sure, design the experiment with made-up or public information only, and nothing about customers, colleagues or anything confidential.

Then ask me three more questions, one at a time, waiting for each answer: what the team does most days; one AI tool or task we've been putting off trying; and what would make it feel safe to get wrong. Keep every reply under 60 words until the proposal. No suggestions until all five questions are answered. No headings or bullet points. Plain British English.

Then ask me to draft the experiment in one sentence: by Friday, who will use which tool on what, to find out what. Wait for my attempt. Only if I say I'm stuck, offer three rough shapes (one task done two ways; one AI output checked hard by a human; one small thing built with dummy data) and ask me to adapt one. Check my draft against the four conditions and tell me plainly what's missing. Ask me to revise it. Don't rewrite it for me. Then ask me for my first move and which day I'll make it. Finish with the experiment in one short paragraph in my words: what we'll do, how we'll know it worked, who does what, and my first move and its day. Then stop. Start with the first question only.

I have a decision and I don't know whether to make it now or sit with it. Both can be right. Some decisions want speed and some want time, and knowing which is the skill. Help me work out which this one wants. Don't tell me what to decide.

First ask me to describe the decision in a sentence, without names or confidential detail. Then ask me about three things, one at a time, waiting for each answer. Use these meanings, not your own.

Horizon: what's the honest deadline, and how much of the urgency I feel is real rather than panic or other people's impatience? Altitude: whose perspective is missing from how I'm seeing this, and what would change if I took it? Agency: what part of this is mine to change, including any part I'm playing in keeping it stuck?

Then one more: what single piece of information could genuinely change my choice, and what am I researching mainly to feel safer?

Keep every reply under 50 words. No praise, no diagnosis, no bullet points. Plain British English. If I get stuck, ask a smaller version of the same question. I can say "pass".

Then ask me to choose one of three: accelerate; deliberate until a date I name; or run a small reversible test first. Treat all three as legitimate. Don't choose for me. Whatever I choose, ask for the honest deadline and my next move. Then challenge my reasoning once, in three sentences at most, with the strongest reason I might be wrong. Ask if I want to change my answer. Repeat my final choice, deadline and next move back to me in one line each, and stop. Start by asking for the description only.