Research-to-slides workflow
Turn a long report or research notes into a short, decision-focused deck with one message per slide and evidence in the notes.
Fill in your details
Values are inserted into every prompt below. Nothing you type leaves your browser.
4 field(s) left — unfilled variables stay as {{placeholders}}.
1Extract the messages
The most common mistake is one slide per section of the report. Slides should follow the audience's decision, not the report's structure.
Here is research material: {{research_text}} Audience: {{audience}} Decision they need to make: {{key_decision}} List the 5-8 findings that matter most for this decision, each as a single sentence with the supporting evidence (quote or number and where it came from). Rank them by importance to the decision. Then list findings that are interesting but not relevant to this decision; these go in the appendix.2Storyline the deck
Read only the slide titles in order: they should tell the whole story on their own.
Create a {{slide_count}}-slide storyline. Slide 1: the answer or recommendation. Then the findings that support it, in an order that builds the argument. Final slide: the decision needed and options. For each slide: action title (a full-sentence takeaway), the evidence or chart that proves it, and speaker notes with sources. Then write the titles alone as a list so I can check they tell the story by themselves.3Build and check
Paste into your deck tool. For each chart, check numbers against the source. Remove any slide whose title the audience wouldn't miss.
Example input
A 30-page user research report on mobile usage; product leadership deciding between a native app and mobile web.
Example output
Titles only 1. Improve mobile web now; revisit a native app in Q3 2027 2. 68% of mobile sessions start from search or email links, not a home screen 3. The top mobile complaint is slow checkout, which an app wouldn't fix by itself ... 8. Decision today: approve a 2-quarter mobile web performance programme
Troubleshooting
- Findings lack evidence: ask the model to quote the source passage for each, and drop any it can't support.
Known limitations
- Summaries may lose nuance (sample sizes, caveats); keep them in speaker notes or the appendix.