Last verified: Claude Sonnet 5.5 ยท October 2026
You asked a yes or no question and got four paragraphs, a bulleted overview, three caveats and a summary of what it had just said. The answer was in there. You had to dig. This page gives you a prompt that builds a brevity block: the answer first, the length you actually want, and the padding named so it can be cut.
Paste this into a fresh Claude or ChatGPT chat. It asks how long your answers should be for each kind of question and what padding you hate, then hands back a brevity block you paste at the top of any chat.
You're going to build my BREVITY BLOCK: a reusable instruction block I'll paste into any AI chat, so answers come back at the length I want, with the answer first. Interview me first, one question at a time. Where I don't know, propose a default and move on: 1. What do I ask AI most, and how long should a good answer be for each kind (a quick question, a decision, a draft, an explanation)? 2. What padding annoys me most? Offer: an opening line that restates my question, a bulleted overview before the answer, caveats nobody asked for, a closing summary, offers of more help. 3. When do I want more detail? (when I ask for it, on hard topics, for drafts.) 4. Do I want the AI to ask one clarifying question rather than answer a vague request at length? 5. Paste one answer that was too long. Show me where the answer started and what could have gone. Then output my BREVITY BLOCK as one copy-paste block containing: (a) a length rule per kind of request, in sentences or words, (b) 'answer first': the first sentence is the answer, (c) the named list of padding to leave out, (d) when to give more and how I signal it, (e) a closing self-check: could this reply lose its first or last paragraph without losing anything? If yes, cut it, then append 'Length check: pass' or say what you cut. Keep it under 200 words. After the block, give me the one short phrase that works best to shorten a single reply on the spot.
Short is not always better. These habits keep the brevity where it helps and let you pull more detail when you want it.
Here is the block the generator hands back, for our running example company, the fictional consultancy Fenwick & Cole.
USE: Quick questions, decisions and short drafts for a consultancy partner who is usually between meetings. LENGTH: A factual or yes/no question gets one to two sentences. A decision gets one short paragraph with the recommendation first. A draft or a requested explanation gets as much as it needs and no more. ANSWER FIRST: The first sentence is the answer. Reasoning comes after it, only if it changes what I would do. LEAVE OUT: restating my question, an overview before the answer, caveats I did not ask about, a closing summary, and offers of further help. MORE DETAIL: Give it when I ask, when the topic is genuinely hard, or when I write 'expand'. UNCLEAR REQUESTS: Ask one clarifying question instead of answering at length on a guess. SELF-CHECK: before replying, ask whether the reply could lose its first or last paragraph without losing anything. If so, cut it. Then append 'Length check: pass', or say what you cut.
Models tend to be thorough and to look helpful, so they add context, caveats and a summary by default. Without a length instruction, thorough is the default setting, whatever you actually asked.
Tell it the length for each kind of question, put the answer in the first sentence, and name the padding to leave out. Store that in your project or custom instructions so you do not retype it. Naming what to cut works better than just saying 'be concise'.
Not by themselves. A short answer drops padding, not reasoning, provided you say reasoning is welcome when it changes your decision. For hard problems, ask for more detail explicitly or let the block say when to expand.
You can, and it helps a little, but it is vague: the AI has to guess how short and what to cut. A length per kind of question and a named list of padding gives it something it can follow.
AI output arrives with curly quotes, arrows, long dashes and non-breaking spaces. Pasted into code, a CSV, a CMS or an email tool, they become broken strings, odd characters or failed imports.
Open the fixThe AI writes competent but generic prose: wrong register, hedging, buzzwords you would never use, sometimes American spelling. Someone rewrites every draft to sound like the company, which cancels out the time saved.
Open the fixThe AI follows your rules for the first few replies, then drifts: the format slips, the banned words return, the tone changes. In a long chat you end up repeating yourself.
Open the fixThe brevity block you just generated is the kind of rule an AI Brain stores as always-on context, tuned on your real work and kept current as new problems appear. Paste it and this chat answers at the length you want; installed, every chat starts that way, because the rule lives in the system and not in your memory of what to fix.
What else does AI get wrong that you fix by hand? Browse the other fixes
Free founder tools
Each tool gives you a first-pass answer in your browser. No email gate and nothing stored.