AI presentation prompts: PowerPoint and Google Slides examples

Using AI to structure a presentation It doesn't mean clicking a button and finding yourself a ready deck. It means driving a language model with precise instructions — the brief, the audience, the number of slides — and getting a solid structure to work on in minutes instead of hours. The real advantage is not the design: it is the content of slides, the one that often locks for half an hour in front of the white screen.

In this guide you will find ready-made examples of AI prompt for the most common presentation types, a step-by-step workflow, and errors that turn a good prompt into useless output. The angle is deliberately on the content — texts, outline, speaker notes — not on the graphic design.

Why AI prompts for presentations must be different from others

When you use ChatGPT prompt to write a text or analyze a document, the output is linear: a block of text that you read from start to finish. A presentation has a hierarchical structure — deck title, slide titles, bullet, notes. and must work for a specific audience at a specific time.

This changes the prompt logic. You must specify:

• Type of presentation: pitch, update, training, client proposal, keynote
• Number of slides or duration: without explicit limit, AI always produces too much
• Public: investors, colleagues, clients, managers, technical team
• Final objective: approval, purchase, training, alignment
• Output format: slide + bullet title or full text or only outline

Without these items, you get a generic presentation on any topic — which could be anyone's.

4-step workflow with ready prompts

The procedure is also applicable to presentations created with Kimi AI: public, expected decision and sources remain part of the brief. Before entrusting the product with recurring works, verify the Kimi app plans and limits.

The most efficient way to use AI for a presentation is not to ask "create me a presentation on X". It's to build slide by slide, phase by phase.

1. Brief. Describe the full context to AI: topic, audience, duration, goal, tone, any constraints. This message is the foundation of everything that follows.
2. Outline. Ask for the structure: title of each slide and communicative goal of each. No texts yet. Correct here if the structure does not convince you.
3. Slide lyrics. For each slide (or group), ask for definitive title + 3-4 concise bullet. Add your real data at this stage.
4. Rapporteur's notes. For each slide, ask for the speaker's notes: key point, example, transition to the next slide. Max 80 words per slide.

Brief prompt (phase 1)

I need to create a [TYPE] presentation.
Audience: [WHO THEY ARE].
Duration: [X minutes / X slides].
Goal: [WHAT VIEWERS SHOULD GAIN].
Tone: [formal / direct / narrative].
Context: [2-3 lines about the product/service/project].
Before proceeding, confirm the structure we will use.

Outline prompt (phase 2)

Create the presentation outline. For each slide:
- Number and title
- The slide's communication goal (one line)
- Suggested content type (text, data, chart, image)
Total: [N] slides. Do not write the text yet.

Slide Text Prompt (phase 3)

Write the content for slide [N]: "[TITLE]".
Format: H2 heading + 3 bullets of max 12 words each.
Data to include: [insert your actual figures here].
Tone: [consistent with the brief].
No generic phrases: each bullet must be specific and verifiable.

Prompt for speaker notes (phase 4)

Write the speaker notes for slide [N]: "[TITLE]".
Max 80 words. Structure: key point (1 sentence) → concrete example → transition to the next slide.
Tone: spoken, not formal like a written text.

Prompt for presentation type

Each type of presentation has a different narrative logic. Here are the starting prompts for the most frequent cases.

Prompt for high impact specific slides

Some slides make the difference between a convincing presentation and one that is forgotten. Here are the prompts for the three most critical.

Opening Slide (hook)

Create the opening slide for a presentation on [topic].
Goal: capture attention in 30 seconds.
Options to give me: 3 title variations (shocking data / provocative question / scenario).
No "today I will present X": go straight to the problem or tension.

Slide "Why Now"

Create the slide that answers "why act now on [problem/opportunity]".
Include: 2 actual market signals (which I will provide) + 1 risk of delaying.
Max 4 bullets. No artificial urgency.

Slide call to action final

Create the final slide with the call to action for [type of presentation].
Audience: [who is watching].
Specific CTA: [what they should do: approve, sign, reply by X].
Add a concrete next step with a realistic date or deadline.

Practical examples

Consulting firm — client proposal. A partner in a professional studio uses this workflow: paste the client brief (a Word page) into the prompt, ask for the 8-slide outline, then generate the slide texts per slide by manually entering the quote numbers. Time: 25 minutes for the first draft instead of 3 hours. Human review focuses on prices and nuances of relationship, not structure.

Head of HR — Onboarding Training. The training module on business processes was updated once a year because it required too much work. With training prompts, each update starts from a regenerated outline in 10 minutes. AI produces slide texts, HR adds internal examples and specific policies. The module updates quarterly.

These cases reflect a common pattern: AI deals with the Structure and Standard Texts, the professional brings the specific context and real data that AI cannot invent. To understand how to extend this approach to other documents, read the guide on AI prompt for documents.

Mistakes to Avoid

• Prompt without defined audience. "Create a presentation about AI" produces something for no one. It always specifies who is watching and what they already know about the topic.
• Asking for everything in a prompt. "Create me 20 complete slides with texts and notes" returns long and poor output. One slide at a time, or a block of 5, keeps the quality high.
• Accepting AI numbers. If AI writes "the market is worth 4.2 billion", either that data you provided or don't use it. Presentations with fabricated data deflate at the first audience demand.
• Ignoring the speaker's notes. Slides are visual support, not teleprompter. The speaker's notes are where the narrative lives. Asking AI is often more useful than optimizing the bullets.
• Using the same prompt for different audiences. A same presentation for the board and the operating team has completely different structures, tones and levels of detail. Adapt the prompt, not just the content.

How to apply it in the company

AI for presentations becomes a structural advantage when it stops being an individual experiment and becomes a shared process.

A concrete approach for an SME: create an internal document with the approved template prompt for the most used 3-4 types of presentation (pitch, update, customer proposal). Each template already includes the company's standard brief —sector, type audience, tone.so the user just needs to add the case-specific data. Time savings multiply on the team, not the individual.

For those who use ChatGPT for content already in other departments, integrating presentations into the stream is natural: same tool, same prompt scheme, same human review logic on the output.

If you want to explore tools that also generate the visual design of slides — not just text — the guide on Best AI Tools covers the main options with a practical comparison.

For teams that use Claude instead of ChatGPT, the above prompts work without modification: the syntax is identical, they slightly change the output format preferences (Claude tends to structure markdowns better).

Conclusion

The AI prompt for presentations They work when you stop asking AI to "make the presentation" and start using it as a collaborator who needs a precise brief. The 4-step workflow — briefs, outline, texts, notes — reduces the time on the first draft from hours to minutes and focuses human work where it is worth: the real data, the relationship tone, the final call to action.

AI becomes useful when it enters processes. Giallo Studio helps SMEs and teams build concrete, measurable, and sustainable automations — including business document and presentation production flows.