How to Turn a Product Spec Dump Into a Three-Line Scope

Anyone drowning in a product spec dump can turn it into a three-line scope in 8 minutes: paste the ramble, force exactly three lines for user, outcome, and out-of-scope, and ban brochure adjectives. Run it in ChatGPT, Claude, or Gemini before the dump becomes a twelve-page PRD costume.

A product spec three-line scope is a peer-ready lock, not a feature brochure and not a ticket title. If the model keeps “delightful seamless experiences,” you never cut the dump. User, outcome, out-of-scope beats a paragraph of vision.

Teams ship fog when kickoff decks treat adjectives as scope. A three-line lock lets engineering reject extras without a debate club. If a stakeholder cannot point to the Out of scope line, they are still negotiating a brochure.

This is not a messy note into a one-page brief. A one-page brief still has sections. This is three lines only. It is not a feature request into a one-line ticket title. A ticket title is one line. This keeps user, outcome, and out-of-scope. It is not a same-day decision note with two options. A decision note compares options. This locks scope before options expand.

What you need

One messy PRD, Slack dump, or meeting notes paste. The real user noun if you have it (or UNKNOWN). ChatGPT, Claude, or Gemini. Eight minutes and one cold peer scan from someone who did not write the dump.

Banned before you start: delightful, seamless, innovative, next-gen, robust, scalable platform, empower users, best-in-class, leverage AI, “phase 2 TBD” as fake out-of-scope, and turning three lines back into a brochure.

Print the ban list next to your laptop. Models recycle the same praise words when the dump is long. Your job is to make those words illegal before the first draft lands.

Turn a product spec dump into three-line scope

1. Paste the dump. Name the user or write UNKNOWN

Paste the full ramble. Do not pre-clean it into polite bullets. The mess is the input.

If product pasted three personas, pick the one who touches the screen this week. Secondary personas wait for a later scope lock. Stacking four users into one line is how UNKNOWN becomes a crowd.

Instruct ChatGPT, Claude, or Gemini: “Do not write a PRD. Do not write a brochure. Output exactly three lines labeled User, Outcome, and Out of scope. Ban delightful, seamless, innovative, robust, empower, and best-in-class. If the user is not named, write UNKNOWN on the User line. Then stop.”

Good user line: “User: warehouse ops lead on night shift.” Bad: “User: anyone seeking delightful inventory magic.” Magic is brochure.

2. Force exactly three lines — user, outcome, out-of-scope

Shape:

Outcome must survive a sticky note. If you need a slide to explain the outcome, it is still a vision sentence. Rewrite until a new teammate can repeat it after hearing it once. User: who does the work. Outcome: what changes in one concrete result. Out of scope: what this pass will not build.

Instruct: “Output exactly three lines. Outcome must name a measurable or observable result, not a vibe. Out of scope must name at least one real cut, not ‘phase 2.’ No fourth line. No vision paragraph.”

Good outcome: “Outcome: night-shift lead can flag a damaged pallet in under 30 seconds from the floor app.” Bad: “Outcome: transform warehouse excellence.” Excellence is fog.

Keep out-of-scope honest

If the dump never names a cut, write “Out of scope: UNKNOWN — need one real cut from PM” rather than inventing “AI recommendations” the team never discussed. Invented cuts become fake commitments.

3. Ban brochure adjectives, then strip extras

Tell the model: “Delete delightful, seamless, innovative, next-gen, robust, scalable platform, empower, leverage, and best-in-class. If a line needs those words to stand, rewrite the line with a concrete noun.”

When the model sneaks “delightfully simple” back in, do not debate tone. Delete the adjective and demand a noun. Tone workshops are how dumps return through the side door.

ChatGPT will add a fourth “Success metrics” line. Claude will soften outcome into “help users feel confident.” Gemini will invent stakeholders you did not paste. Cut all three gifts.

Hard fail: more than three lines; missing Out of scope; brochure adjectives; UNKNOWN upgraded into a fake persona with a backstory.

4. One cold peer scan

Hand the three lines to a peer who did not write the dump. Ask: “Can you reject a feature request using only Out of scope?” If they cannot, the cut is fake.

Ask the peer to invent one feature request that should die. They must kill it using only your Out of scope line. If they need the original dump, your cut is theater.

Read the three lines out loud in under twenty seconds. If you need a slide to explain them, you still have a dump.

5. One human pass, then lock

Paste the three lines at the top of the ticket or doc. Do not let the model re-expand them into a vision memo.

Check traps: brochure adjectives, fourth line, fake persona, out-of-scope as “TBD,” outcome as vibe. Fix: “Replace only the line I mark. Keep exactly three lines. No adjectives from the ban list.”

One pass. If delightful still sits on Outcome, you never wrote a three-line scope. Stop.

What good looks like

Trigger: You are about to paste a two-thousand-word PRD dump into a kickoff without a scope lock.

Input: Spec dump, user noun or UNKNOWN, ChatGPT/Claude/Gemini, cold peer scan.

Output: Exactly three lines — User, Outcome, Out of scope — concrete nouns, no brochure adjectives, one real cut.

Stop: After one cold peer scan. No fourth line. No vision re-expand.

Takeaways

- Paste the dump. Force User, Outcome, Out of scope. Exactly three lines.

- Ban brochure adjectives. UNKNOWN beats a fake persona.

- Out of scope needs a real cut, not phase-2 theater.

- One cold peer scan. Then lock. Do not re-expand.

Keep the three lines pinned at the top of the epic. New comments go under the lock, not inside it. Re-running the model for “richer language” is how you undo the morning.