Verified against ChatGPT · 2026-08-01
Turn a hard choice into a weighted decision matrix instead of a gut call
A three-step weighted-scoring prompt that confirms criteria weights before scoring, shows the justification behind every score, and checks whether the winning option would flip under a small weight change.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Help me decide between these options using a weighted decision matrix, not a narrative recommendation. DECISION Which project management tool should our 15-person team switch to? OPTIONS Linear, Asana, and staying on Trello CRITERIA THAT MATTER TO ME Ease of onboarding non-technical teammates, integration with our existing Slack and GitHub setup, and price at our headcount STEP 1 — WEIGHTS Propose a weight out of 10 for each criterion based on what I described above, and ask me to confirm or adjust before scoring anything. Do not invent a criterion I did not list unless you flag it separately as a suggested addition. STEP 2 — SCORING Once weights are confirmed, score every option against every criterion on a 1 to 5 scale with one sentence of justification per score, no unexplained numbers. Put this in a table: options as rows, criteria as columns, weighted total as the last column. STEP 3 — SENSITIVITY CHECK Tell me if the ranking is close. Specifically, name the smallest weight change on any single criterion that would flip the top choice. If the top two options are within 10% of each other's weighted total, say explicitly that the matrix is not decisive and name the real tiebreaker factor a spreadsheet cannot capture. Do not tell me what to choose outside of this structure. Show the matrix and let the numbers make the case, or explicitly say they do not. If your ChatGPT plan exposes a reasoning effort control for GPT-5.1, set it to high before running this. Keeping every score and the sensitivity check internally consistent benefits from extra thinking time, not the default setting.
Customize the highlighted detailsoptional — the prompt above already works
Why this works
Asking directly for a recommendation invites a persuasive-sounding narrative that can rationalize essentially any conclusion, while forcing per-criterion, per-option scores with a one-sentence justification for every cell makes the reasoning auditable line by line instead of hidden inside prose that reads confidently regardless of whether the underlying comparison actually supports it. Requiring weight confirmation before any scoring happens prevents a subtler bias: a model that scores first and derives weights to match afterward tends to unconsciously produce numbers that support whatever conclusion it would have narratively recommended anyway, rather than weights that reflect what you actually said mattered. The sensitivity check is the most load-bearing step here: a weighted matrix that flips its top choice with a small nudge to one weight is not actually decisive, it only looks decisive because it produced a single ranked number, and naming that fragility explicitly stops the output format itself, a number with two decimal places, from lending false confidence to what is genuinely a close call that should come down to a factor the matrix cannot quantify.
What you get back
Weights confirmed: onboarding ease 8, integrations 7, price 6. Scoring table shows Linear at 5.4 weighted total, Asana at 5.1, Trello at 3.2. Sensitivity check: Linear and Asana are within 6% of each other, so a one-point increase to the price weight alone would flip the top choice to Asana. The matrix calls this close, not decisive, and flags team familiarity with the tool as the real tiebreaker it cannot score.
Verified against
ChatGPT GPT-5.1 · 2026-08-01
Changelog
- 2026-08-01 — Initial publish, verified against ChatGPT GPT-5.1.
Building this for real?
This is a free starting point. If you'd rather have what Scult builds built and running for your business, that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
