Research

Verified against ChatGPT · 2026-08-14

Build a research presentation storyline that survives an audience skipping straight to slide 12

Structures a research presentation around a single argument thread with a clear slide-by-slide storyline, so the deck holds up even for someone who jumps ahead or only reads the headlines.

ChatGPT (GPT-5.1)4 fillable variables

The prompt

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

Build a slide-by-slide storyline for a research presentation from the findings below — not the slide content itself yet, the structural outline of what each slide argues and how it connects to the one before it.

RESEARCH FINDINGS
Flexible scheduling raised retention 12% overall, effect was concentrated in employees with caregiving responsibilities, no effect on productivity metrics, cost to implement was under $50/employee/year.

AUDIENCE
The exec team, who are primarily worried about cost and whether this creates a fairness perception issue between employees.

CORE ARGUMENT I WANT THE DECK TO MAKE
We should offer flexible scheduling as an opt-in for all employees, not restrict it to caregivers, because the cost is low and restricting it risks a fairness backlash.

TIME LIMIT
10 minutes total, including questions.

Start by stating the core argument as a single sentence a viewer should be able to state back after seeing the deck, then build every slide as a step that either establishes a piece of evidence for that argument or addresses the most likely objection to it — cut any finding from Flexible scheduling raised retention 12% overall, effect was concentrated in employees with caregiving responsibilities, no effect on productivity metrics, cost to implement was under $50/employee/year. that's true and interesting but doesn't serve the core argument, even if it took real effort to uncover.

For each slide, write its headline as the actual claim of that slide, not a topic label — "Retention improved 12% under flexible scheduling" is a slide headline, "Retention Results" is not, because a reader skimming headlines alone should be able to reconstruct the deck's argument without reading the supporting content on any slide.

Assume someone in the room will jump straight to a specific slide or ask about it out of order — for each slide, note in one line whether it depends on a claim established earlier that would need a quick recap if someone jumped ahead, so the presenter knows which slides are safe to present out of sequence and which need a bridge sentence.

Given the time limit, mark which slides are core (the argument breaks without them) versus supporting (useful but cuttable if time runs short) — a presenter needs an honest answer to "what do I cut if I'm running long," not a deck where every slide claims to be essential.

OUTPUT FORMAT
For each slide: number, claim-headline, core or supporting, and a one-line note on whether it depends on an earlier slide's claim. End with the one-sentence core argument the deck should leave the audience able to repeat.

Customize

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

Why this works

Requiring slide headlines to state the actual claim rather than a topic label directly targets the most common structural weakness in research decks — a title like 'Retention Results' asks the audience to hold the interpretation in their head until they've processed the chart beneath it, while a claim-headline front-loads the interpretation so the argument survives even if someone only reads the headline row, which is closer to how a skimming or distracted audience actually consumes a deck. The instruction to cut findings that don't serve the stated core argument, even effortful ones, counters GPT-5.1's default bias toward exhaustive inclusion when it has a rich set of findings available — a model building a deck from a full findings list will tend to give every real finding a slide out of a sense that leaving out something true wastes the research, when a presentation's job is persuasion toward one argument, not a complete report of everything discovered. Marking each slide's dependency on an earlier claim addresses a real structural risk specific to live presentations that a slide deck built for linear reading ignores: executives frequently jump to a slide that interests them or interrupt with a question out of the planned order, and a deck built without accounting for this leaves the presenter unable to answer a mid-deck question without silently backfilling context the audience hasn't seen yet — flagging dependencies in advance lets the presenter prepare a one-line bridge rather than getting caught improvising it live. The core-versus-supporting split tied to the actual time limit gives an honest answer to a pressure a presenter genuinely faces (running long and needing to cut on the fly) rather than a deck that implicitly claims every slide is equally load-bearing, which is what happens by default when a model is simply asked to "turn these findings into a deck" without being told time is genuinely constrained.

What you get back

Slide 3: 'Flexible scheduling costs under $50/employee/year to implement' — core, no dependency on earlier slides, safe to present out of order. Slide 5: 'The 12% retention gain is concentrated among caregivers, not evenly spread' — supporting but load-bearing for the fairness argument; depends on slide 3's cost figure being established first if asked why not just restrict it to caregivers. Core argument: offer flexible scheduling to all employees as opt-in, since the cost is low and restricting it by role risks a fairness backlash without saving meaningfully on cost.

Verified against

ChatGPT GPT-5.1 · 2026-08-14

Changelog

  • 2026-08-14 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 Research 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