Instagram

Verified against ChatGPT · 2026-08-12

Design a Stories poll that actually decides something, instead of collecting opinions nobody acts on

Writes the question and two options for an Instagram Stories poll sticker built to produce a real, actionable answer to a decision you're actually making, rather than a generic engagement-bait either/or.

ChatGPT (GPT-5.1)3 fillable variables

The prompt

Ready to copy — highlighted parts are example details you can swap.

You are writing a Stories poll sticker — question plus two option labels — and it has to be built around a decision that will actually change based on the result, not a generic engagement prompt where the answer changes nothing regardless of which side wins.

THE ACTUAL DECISION THIS POLL RESOLVES
Whether the next glaze restock should happen weekday evening or weekend morning.

WHAT HAPPENS DIFFERENTLY DEPENDING ON THE RESULT
If evening wins, restock goes live Thursday 7pm; if morning wins, it goes live Saturday 9am — no other option is on the table.

AUDIENCE CONTEXT NEEDED TO ANSWER
Viewer needs to know the restock is happening this week at all — mention that in the frame text above the poll.

RULES
Write the question so that both options are genuinely plausible outcomes of the actual decision, not a lopsided pair where one option is obviously correct — a poll where 95% of people will pick the same option isn't collecting information, it's a vote of approval dressed as a choice, and it should be flagged as such rather than written as if it were a real toss-up. Keep the question itself short enough to read in the two seconds a Stories viewer typically spends per frame, front-loading the actual choice rather than a setup sentence before it. Write the two option labels (each capped at the sticker's short character limit) so that someone who has not read any caption or context, only the Story frame itself, could answer meaningfully — if answering correctly requires information from a post they haven't seen, include that context directly in the Story frame's other text, not assumed as background knowledge. State explicitly, in the output, what you will actually do with each of the two possible results — if the answer would be 'nothing changes either way,' say so plainly instead of shipping the poll anyway.

WHAT NOT TO DO
Do not default to a poll about opinions on something already decided ('team pizza or team tacos?' when the menu is already finalized) just because it's an easy engagement format — every poll produced here should trace back to the stated decision. Do not write a question with three or more effective options crammed into two labels ('cats, dogs, or both?' squeezed into a two-option sticker) — if the real decision has more than two live options, say so and recommend a different sticker or a follow-up poll instead of distorting the question to fit two boxes.

OUTPUT FORMAT
1. The poll question text.
2. The two option labels.
3. One line stating what you will actually do for each possible result.
4. If the decision genuinely doesn't fit a binary poll, a note saying so instead of forcing it.

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

Requiring the model to check whether both options are genuinely plausible, rather than accepting any two-option framing as valid, targets a specific and very common failure mode: asked to 'write an engaging poll,' GPT-5.1 defaults to whatever pairing sounds catchiest as a question (this-or-that framing, playful opposites) without checking whether the underlying choice is actually live — producing a poll that looks like audience research but functions as a popularity contest with a foregone conclusion, since one option is obviously going to win regardless of framing. Forcing an explicit statement of what happens for each possible result is the mechanism that keeps the poll honest: a model that has to write down 'if X wins, we do Y' before finalizing the question can't quietly ship a poll where the honest answer is 'nothing changes either way,' because that gap becomes visible in the output rather than staying an unexamined assumption. The rule against cramming three-plus real options into a two-option sticker addresses a structural mismatch the model won't flag on its own: a Stories poll sticker only supports two labels, so a model asked to poll a genuinely three-way decision will usually just pick the two most common options and silently drop the third, producing a poll that resolves a different, narrower question than the one actually being decided — naming this explicitly as something to flag rather than paper over keeps the tool from being used for a decision shape it structurally can't represent.

What you get back

Poll question: "Restock drops Thursday evening or Saturday morning?" Options: "Thurs 7pm" / "Sat 9am". Result plan: Thursday-evening win schedules the drop for 7pm that day; Saturday-morning win schedules it for 9am — no third option exists, so this fits the two-label format cleanly.

Verified against

ChatGPT GPT-5.1 · 2026-08-12

Changelog

  • 2026-08-12 Initial publish, verified against ChatGPT GPT-5.1.

Need this built into your business?

If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.

EXPLORE WHAT SCULT BUILDS
All Instagram prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY