Use AI to Draft Slide Deck Outlines

Why Deck Work Feels Slower Than It Should

Most slide work does not fail because people lack ideas. It fails because the story is fuzzy. Teams jump straight into design, fill slides with disconnected facts, and end up rewriting the narrative three times before the deck is useful.

AI is most helpful at the stage before design: shaping the argument, organizing the flow, and surfacing the questions each slide needs to answer.

This tutorial shows you how to use AI to draft a slide deck outline that is clear enough for design work and strong enough for decision-making. You are not asking the model to make your final presentation. You are using it to build the backbone of the deck faster.

What a Good Deck Outline Actually Does

A useful outline does four things:

• defines the audience
• clarifies the goal of the deck
• sequences the story
• states what each slide must prove

Without those four pieces, slides become decoration. With them, slides become evidence.

That is why the outline matters more than most people think. It is the difference between "a bunch of slides" and "an argument a room can follow."

Step 1: Start With the Decision, Not the Topic

Bad prompt:

Help me make a presentation about onboarding.

Better prompt:

I need a 10-slide outline for a presentation to leadership.
Goal: recommend changes to our onboarding flow.
Audience: VP Product, Head of Support, and Growth lead.
Desired outcome: approval for a 6-week experiment.

The model needs the decision context. A slide deck about "onboarding" could be a performance review, a process explanation, a strategy pitch, or a project update. Those are different decks.

Before drafting, answer:

• who is this for?
• what do I want them to believe?
• what action do I want afterward?

Step 2: Ask for a Narrative, Not Just a Slide List

Your prompt should request a story structure.

Use:

Create a slide deck outline with:
1. opening context
2. problem definition
3. evidence
4. options considered
5. recommendation
6. next steps

For each slide, include:
- slide title
- objective
- key points
- suggested proof or data

That format pushes the model to think in terms of argument, not filler.

If you skip the "objective" field, the output usually becomes generic. If you include it, every slide has a reason to exist.

Step 3: Add the Source Material Before You Refine

Once you have the basic shape, feed in the actual inputs:

• rough notes
• research summaries
• metrics
• customer quotes
• project constraints

Prompt:

Revise the outline using these inputs.
Prioritize evidence that supports the recommendation.
Remove weak or repetitive slides.
Flag any gap where the deck needs more proof.

AI works best here as an organizer and gap finder. It can tell you where the story is thin before you waste time polishing weak slides.

Step 4: Create Two Versions of the Outline

Most decks need two layers:

• the full working outline for the builder
• the compressed executive outline for the reviewer

Ask for both:

Create:
1. a full outline with detail for the deck owner
2. a one-line-per-slide version for executive review

This keeps the thinking deep while making the deck easy to review quickly.

It also helps you spot over-complexity. If a slide cannot be summarized in one line, the slide probably is not ready yet.

Step 5: Stress-Test the Flow Before Opening Slides

Use AI to challenge the outline before design begins.

Good follow-up prompts:

What would a skeptical executive question in this outline?

Which slides feel redundant or weak?

What evidence is missing for the recommendation to feel credible?

This turns the model into a cheap first reviewer. It is especially useful when you are close to the material and can no longer see the gaps clearly.

Step 6: Hand the Outline to Design With Clear Instructions

Once the logic is stable, create a design-ready brief:

Turn this outline into a slide production brief.
For each slide, include:
- slide purpose
- suggested chart or visual
- level of detail
- tone
- presenter note if needed

That keeps design focused on presentation quality instead of re-litigating the story.

Common Mistakes

• asking for a whole deck before defining the audience
• using AI to generate polished slide copy too early
• skipping the evidence gaps
• accepting a linear list of slides that has no argument underneath
• designing before the recommendation is clear

The model should help you think before it helps you polish.

What To Do Next

• Use Turn Raw Notes Into Clear Reports when the source material is still messy
• Use Write Better Briefs So AI Gives Better Answers if your prompt quality is the real bottleneck
• Use Prepare for Meetings with AI when the deck is meant to support a live conversation