Define the core message and evidence for each slide

Author: AILesson8 min setupTested with:ChatGPTReviewed: 2026-08-28

Quick answer

Turn an outline and source pack into one-message slides with traceable support. Provide: Presentation brief, Outline or current slides, Evidence pack. Expected result: A slide-by-slide evidence map with titles, visuals, speaker notes, and content gaps.

Add your context

Your text stays in this browser. AILesson Prompts does not send it to a model or server.

Presentation briefOutline or current slidesEvidence pack

Your prompt

Unfilled fields remain visible as placeholders, so you can still copy and edit the prompt

Create a slide-by-slide message and evidence map. Do not invent evidence or design a finished deck.

Audience, decision, time, setting, action, and slide limit:
[brief]

Numbered outline, sequence, and required content:
[outline]

Source IDs, facts, dates, methods, limitations, quotes, and permissions:
[evidence]

For each proposed slide define exactly one audience-relevant takeaway written as a complete claim, then a short assertion-style title that does not overstate it. Map every factual claim to source ID and preserve sample, date, uncertainty, correlation, and comparison basis. Distinguish evidence, interpretation, recommendation, and open question. If support is absent, mark Evidence gap rather than creating a fact.

Specify the most suitable evidence form—number, quote, table, chart, process, example, or none—and explain why; never recommend a chart when the available data cannot support it. Give minimal on-slide content, speaker-only context, transition purpose, and what can be removed. Check narrative dependency, duplicated claims, unsupported precision, unreadable density, accessibility, source permissions, and whether the final request follows from the evidence. End with a source coverage ledger and questions that must be answered before production.

Private by defaultPrompt assembly happens locally in your browser. Avoid placing confidential information into any AI service unless your organization allows it.

From input to outcome

A worked example

See how concrete context turns this recipe into a usable result

Actual input

Presentation brief: Seven-minute internal presentation to the library director and operations manager. Decide whether to run a larger reminder-text test, not whether to deploy permanently. Maximum five content slides plus title. Desired action: approve a four-week randomized test and name an operations owner.
Outline or current slides: 1 problem: expired reservations; 2 pilot result; 3 patron feedback; 4 staff workload; 5 proposed next test; 6 decision request.
Evidence pack: S1 circulation report, 1 Jan-30 Jun 2026: 1,820 reservations expired uncollected, 12% of all reservations. S2 pilot log: 214 opted-in adults, 61% collected before expiry over six weeks; prior-period 198 opted-in adults, 54%; periods not randomized and seasonal mix differed. S3 opt-out log: 11 opted out; 3 late-message reports. S4 staff diary: reply handling averaged about 25 minutes/day across six weeks; self-recorded, no control period. S5 test brief draft: randomize 400 consenting adults, four weeks; primary measure collection before expiry; guardrails opt-out and staff minutes; owner unassigned. Internal data only; no patron quotes cleared.

Example output

Slide 1—“Twelve percent of reservations expired uncollected in Jan-Jun.” Claim/evidence: 1,820 and 12%, S1, exact period. Visual: two-number card, not a trend chart because no monthly series is supplied. Speaker only: define expiry; gap: cost or patron impact is unknown.

Slide 2—“Collection was 61% during the pilot versus 54% in the prior period.” Evidence: S2. Visual: two bars with n=214/n=198 and visible note “non-randomized periods; no causal conclusion.” Speaker interpretation: difference justifies a stronger test, not rollout.

Slide 3—“The pilot revealed patron and workload guardrails that a larger test must measure.” Evidence: S3 has 11 opt-outs/3 late reports; S4 has self-recorded ~25 staff-minutes/day. Visual: three annotated measures, no patron quote. Preserve weak measurement limitation. This combines outline items 3-4 under one readiness message.

Slide 4—“A four-week randomized test can answer the effect question with explicit safeguards.” Recommendation based on S5, not established evidence. Show process: consent → randomize 400 → compare collection; guardrails opt-out/staff minutes. Gaps: sample rationale, message timing, privacy/accessibility review, stopping thresholds.

Slide 5—“Decision requested: approve test design work and assign an operations owner.” The supplied evidence supports testing, not permanent deployment. On-slide: approve/refine/decline plus owner field; speaker states unanswered design gates.

Ledger: S1→1; S2→2; S3/S4→3; S5→4-5. No source unused; no cleared quotes. Remove separate title-only pages for patron feedback and workload to fit five slides. Before production, obtain source permissions classification, sample-size rationale, guardrail thresholds, and decision authority for the test.

Why this works

1
One claim per slide makes evidence and narrative purpose easier to inspect
2
A source ledger prevents polished visuals from detaching numbers and conclusions from their limits

Check the result

Does each slide contain one claim supported by named evidence or marked as a gap
Are screen content and speaker context separated at an appropriate density
Do the sequence and final request follow logically from the evidence

Use it with confidence

Frequently asked questions

Practical answers about when to use this recipe, what to provide, and where human review still matters

What should I prepare before using “Define the core message and evidence for each slide”?

For “Define the core message and evidence for each slide,” prepare Presentation brief, Outline or current slides, and Evidence pack. Replace placeholders only with information you can verify. If a detail is unknown, preserve that uncertainty explicitly instead of asking the model to infer it.

When is the “Define the core message and evidence for each slide” result not ready to use?

The result is not ready if it does not yet deliver the stated outcome—A slide-by-slide evidence map with titles, visuals, speaker notes, and content gaps—from the supplied evidence, or if it relies on unresolved assumptions, missing approvals, or invented details. Use the checks as release gates: revise the source inputs or assign a named, authorized reviewer instead of polishing an unsupported output.

Which AI tools have recorded tests for “Define the core message and evidence for each slide”?

The published test record for “Define the core message and evidence for each slide” lists ChatGPT as of 2026-08-28. This confirms recorded runs, not guaranteed compatibility or identical results in later product versions. For another tool or version, keep every constraint visible and repeat the result checks before use.

More ways to explore