Tag: ChatGPT presentations

14 May 2026
Featured image for Generative AI Presentation Storytelling: 3 Prompts That Turn Dry Data Into a Narrative
Executive Presentations Mary Beth Hazeldine 0 comment

Generative AI Presentation Storytelling: 3 Prompts That Turn Dry Data Into a Narrative

Quick Answer

Generative AI presentation storytelling works when the prompt forces the model into a narrative structure rather than a summary. The three prompts that consistently produce usable drafts are: the situation-complication-resolution prompt, the character-stake-shift prompt, and the data-to-decision prompt. Each forces the model to choose a narrative shape before it generates copy. Without that, AI produces summaries — and senior audiences disengage from summaries.

Hadiya had been a strategy lead in a global consulting firm for eleven years. Her team produced quarterly client decks for FTSE finance directors. In April she ran an experiment: she gave ChatGPT a 22-page client report and asked it to “write a presentation that tells the story of the data.” The model produced 14 slides. Polished bullets, neat headers, clean structure. Her partner read the draft and said, “This reads like a research summary. It doesn’t tell me anything I would remember after the meeting.”

Hadiya rewrote the deck by hand. The next month she tried again — different prompt. This time the draft was usable in 40 minutes. The difference was not the model. The difference was the structure she forced into the prompt before the model wrote a word.

If your AI-drafted decks read like summaries rather than stories

The model is not refusing to tell stories. It is defaulting to the structure most natural to a language model — paragraph-and-bullet summary — because the prompt did not ask for anything else.

Explore the Executive Prompt Pack →

Why generative AI defaults to summary, not story

Large language models are optimised for one task: predicting the next likely token given everything before it. When asked to “write a presentation,” the most likely structure across the training data is the summary deck — title, agenda, sections, bullets, conclusion. That structure dominates corporate output, so the model produces it by default.

A senior audience does not need the summary. They have read the pre-read; they have skimmed the report. What they need is the through-line — the question the data answers, the tension the analysis exposes, the decision that follows. None of that emerges from a prompt that says “write a presentation.”

The fix is not better writing on the model’s part. The fix is a prompt that names the narrative structure before the model generates a single word. Three prompts cover most senior-audience situations. Each one forces a different narrative shape into the output.

The 3 storytelling prompts for generative AI: situation-complication-resolution, character-stake-shift, and data-to-decision — with the use case for each shown as labelled cards

Prompt 1 — Situation, complication, resolution

Use this prompt when the audience needs to follow a logical chain from “where we were” to “where we are now” to “what we propose.” It is the structure underneath most McKinsey-style executive briefings, and it works because senior audiences are trained to listen for it.

The prompt skeleton:

PROMPT — Situation / Complication / Resolution

You are drafting a 12-slide executive presentation. Use the situation-complication-resolution structure. Slides 1–4: the situation (where the business was, supported by 3 specific data points from the source material). Slides 5–8: the complication (the new pressure or shift that disrupts the situation, supported by 2 data points and 1 named risk). Slides 9–12: the resolution (the recommendation, the expected outcome stated as a process commitment, the trip-wires, and the decision being asked of the audience). For each slide, write a 6-word headline and 3 supporting bullets of no more than 14 words each. Do not use abstract verbs (leverage, drive, enable). Use specific verbs from the source material.

The prompt does three things the default does not. It names the structure (situation-complication-resolution). It enforces evidence (specific data points from the source material). It bans the verbs that produce generic AI copy (leverage, drive, enable). The output reads as a deliberate piece of work, not a model’s average guess at what a presentation looks like.

The constraint that matters most is the verb ban. “Leverage” and “drive” are model-default verbs — they show up because they are common across the training data. Senior audiences register them as filler. A prompt that bans them forces the model to pull verbs from the source material instead. Those verbs are specific, sometimes technical, and almost always more credible.

When this prompt is the right choice

Use it for board updates, strategic proposals, and any presentation where the audience expects a logical progression from problem to recommendation. It is less effective for sales pitches, opening keynotes, or any setting where the audience needs an emotional hook before they engage with logic. For those, prompt 2 is stronger.

Prompt 2 — Character, stake, shift

The second prompt forces the model into a narrative shape: a person with something at stake, a moment when the situation changes, the decision that follows. It produces drafts that read like business stories rather than business summaries — useful for keynotes, all-hands briefings, conference talks, and any setting where the audience needs to feel the weight of the decision before they evaluate it.

PROMPT — Character / Stake / Shift

You are drafting a 10-slide presentation that opens with a real person facing a specific decision. Slide 1: name the person, their role, the moment, what was at stake. Slides 2–4: the situation as they understood it. Slide 5: the shift — the new information or moment that changed the calculation. Slides 6–8: how they responded, supported by evidence from the source material. Slide 9: what changed as a result. Slide 10: the decision the audience needs to make now. Use first or third person, not second person. No abstract verbs. No outcome guarantees — describe what the person did, not what was guaranteed to happen.

The “no outcome guarantees” line is critical. Generative AI defaults to outcome-promise language (“this approach delivered transformational results”) because that pattern is over-represented in marketing copy in the training data. Senior audiences are alert to outcome promises and discount the surrounding argument when they hear one. The prompt forces the model into process-commitment language instead.

The character requirement also blocks the model’s most common failure mode: opening with abstract market context. “In today’s rapidly evolving business environment” is the model’s default opener; it dies in the first 30 seconds in front of a senior audience. A real person at a real moment is the opposite.

Build executive slides in 25 minutes, not 3 hours

The Executive Prompt Pack — 71 prompts for ChatGPT and Copilot

  • 71 ready-to-use prompts for executive presentations — story, structure, opening, recommendation, risk, Q&A prep
  • Works in ChatGPT, Microsoft Copilot, and Claude — no separate setup
  • Copy-paste-and-fill format — replace the bracketed fields with your context, run the prompt
  • Includes the situation-complication-resolution and character-stake-shift prompts in full

The Executive Prompt Pack — £19.99, instant access, lifetime use.

Get the Executive Prompt Pack →

For busy professionals who want to create sharper, more strategic PowerPoint presentations.

When this prompt is the right choice

Use it for any presentation that opens with the audience cold — keynote, conference talk, sales pitch, internal kick-off — where the first 90 seconds need to earn the right to the rest. It is also the right prompt for change communications, where the human dimension is what carries the message past intellectual agreement into emotional acceptance.

Less suited to credit committee papers and quarterly board updates, where the audience already has the context and just wants the logic. For those, prompt 1.

Prompt 3 — Data to decision

The third prompt is for the situation senior professionals encounter most often: 30 pages of data that need to become a 12-slide deck that drives a single decision. Default AI prompts produce a “data summary deck” with a recommendation slide near the end. This prompt produces a “decision deck” with the data working as evidence, not as content.

PROMPT — Data to Decision

You are drafting a 12-slide decision deck. The audience must make a single decision at the end of the meeting. Slide 1: state the decision being asked of the audience in one sentence. Slide 2: the recommendation. Slides 3–6: the four most relevant data points that support the recommendation, one per slide. Each data slide must include the headline number, the source, the time period, and a one-sentence interpretation. Slides 7–9: the two or three counter-arguments and the response to each. Slide 10: the trip-wires that would force a re-vote. Slide 11: the resolution being put. Slide 12: the next decision point on the agenda. Do not include market context. Do not include backstory. Do not summarise — every slide must move the decision forward.

The instruction “do not include market context” sounds aggressive. It is necessary because market-context slides are the model’s most common form of padding. Senior audiences in a decision meeting do not need market context; they have it. A deck that opens with market context tells the audience the presenter does not know what they need.

The four-data-points constraint is also load-bearing. AI without a numeric constraint will produce 8–12 data points and trust the audience to pick the relevant ones. Senior audiences read that as analytical laziness. Four data points, with the analysis already done in the slide selection, reads as senior judgement.

For senior leaders running this prompt for the first time, the result is often disorienting — the deck looks shorter than expected, with no agenda slide, no executive summary, no closing thank-you. That is the point. It is a working document, not a conference talk. The room sees the work in the discipline of what was excluded.

Default AI Prompt vs Structured Storytelling Prompt comparison table showing the difference in opener, structure, evidence treatment and verb selection across both approaches

The editorial pass: making AI output sound like you

Even with a strong prompt, AI output reads as AI output without an editorial pass. The model produces text that is grammatically perfect, lexically broad, and tonally even — and that combination is exactly the signature senior audiences register as machine-drafted. A short editorial pass changes the read.

Four moves that take 15 minutes and remove most of the AI signature:

Replace three abstract verbs with specific ones from the source material. Search the draft for “leverage,” “drive,” “enable,” “optimise,” “transform” — replace each with the verb the source document uses. The shift from generic to specific lifts the credibility of the surrounding sentence.

Cut the opening adjective on every bullet. AI defaults to “robust framework,” “comprehensive analysis,” “strategic approach.” Senior audiences treat opening adjectives as filler. Cut them. The bullet reads sharper.

Add one specific number that did not come from the source material. A specific time or duration (“17 minutes into the meeting”), a specific date (“between October and December”), a specific small number (“three of the seven options”) — one of these per page anchors the reader and signals the writer was actually present in the analysis.

Rewrite the recommendation in your own voice. The recommendation slide is the one the audience remembers. AI’s default recommendation language sounds borrowed from a McKinsey report. Yours should not. Read the AI draft, close the file, write the recommendation from scratch. Compare. Use whichever sounds like you.

The editorial pass takes 15 minutes on a 12-slide deck. It is the difference between an AI-drafted deck and an AI-drafted deck the audience does not register as AI-drafted. For senior leaders integrating AI into their workflow, this pass is the discipline that separates time saved from credibility lost.

Want the longer story behind these prompts?

If narrative structure is the gap — not just the prompt — the Business Storytelling Mini-Course covers the frameworks behind these three prompts: situation-complication-resolution, character-stake-shift, and data-to-decision. £29, instant access.

Get the Business Storytelling Mini-Course →

Turn numbers into stories that move executive decisions.

Frequently asked questions

Which model produces the best storytelling drafts — ChatGPT, Copilot, or Claude?

For these three prompts, the difference between the major models is smaller than the difference between a structured prompt and an unstructured one. ChatGPT-5 and Claude Sonnet 4.6 produce slightly more usable drafts on the character-stake-shift prompt because both are stronger at narrative voice. Copilot is stronger on the data-to-decision prompt because it can pull from your own files. None of them produce decision-grade copy without the editorial pass.

How much source material should I paste into the prompt?

For the situation-complication-resolution and data-to-decision prompts, paste the full source — most modern models handle 50+ page documents in a single prompt. For the character-stake-shift prompt, paste only the section about the character and the moment, plus the surrounding context. Pasting more dilutes the focus and produces a draft that wanders. Quality of source material in produces quality of structure out.

Can I run all three prompts on the same source and pick the best draft?

You can, and senior leaders increasingly do. The three drafts read very differently and the comparison clarifies which structure suits the audience. Run all three, compare openers and recommendations, then pick one and apply the editorial pass. Total time: about 60 minutes for a 12-slide deck — substantially less than writing from scratch, and the structural variety is itself a useful reasoning tool.

Does this work for slides themselves, or just the narrative copy?

The prompts produce headline-and-bullet copy ready to drop into slide templates. The visual layout, charts, and design treatment still need to be done in PowerPoint or Keynote — generative AI image and chart output for executive presentations is not yet at a quality that survives a senior audience. The narrative copy is where the time saving sits; the visual layer remains a manual step.

The Winning Edge — weekly newsletter for senior presenters

One framework, one micro-story, one slide pattern — every Thursday morning, ten minutes’ read. For senior professionals presenting to boards, investment committees, and executive sponsors who want my best material before it appears anywhere else.

Subscribe to The Winning Edge →

Not ready for the full prompt pack? Start here: download the free Executive Presentation Checklist — a one-page reference for the structural questions every executive deck must answer before the meeting.

For the matched workflow article, see ChatGPT and Copilot together — the two-tool stack that builds executive decks faster than either alone.

Mary Beth Hazeldine — Owner & Managing Director, Winning Presentations Ltd. With 24 years of corporate banking experience at JPMorgan Chase, PwC, Royal Bank of Scotland, and Commerzbank, she advises senior professionals across financial services, healthcare, technology, and government on integrating AI into executive presentation workflows.

09 Feb 2026
Iterative AI presentation editing process showing refinement from first draft to final
AI Presentations Mary Beth Hazeldine 0 comment

The AI Editing Loop: Why Your First Output Is Never the Final One

The executive showed me her AI-generated presentation with a mix of pride and confusion. “It created this in 30 seconds,” she said. “But something’s off. I can’t use this.”

She was right. The slides were structurally sound. The content was accurate. The formatting was clean. And yet the whole thing felt generic, lifeless, obviously machine-made.

Her mistake wasn’t using AI. It was stopping after the first output.

I’ve watched this pattern hundreds of times since AI presentation tools became mainstream: professionals generate a deck, feel vaguely disappointed, and either abandon AI entirely or present something mediocre because “at least it saved time.”

Both responses miss the point. AI isn’t meant to produce finished work on the first try. It’s meant to produce raw material that you shape through iteration. The magic happens in the editing loop — the systematic process of refining AI output until it sounds like you, fits your audience, and makes the point you actually need to make.

Here’s how that loop works, and why skipping it guarantees underwhelming results.

Quick answer: AI presentation tools generate competent first drafts, not finished products. The editing loop — Generate → Evaluate → Refine → Elevate — transforms generic output into executive-ready content. Most professionals skip this loop entirely, which is why most AI presentations feel obviously AI-generated. Plan for 3-5 iteration cycles on any presentation that matters.

📋 Copy/Paste: The 4-Stage Editing Loop Prompts

Stage 1 (Generate): “Create a [X]-slide presentation on [topic] for [audience]. Include [key sections].”

Stage 2 (Evaluate): “Review this output. Identify: (1) generic phrases, (2) missing context, (3) weak recommendations, (4) sections a skeptic would challenge.”

Stage 3 (Refine): “Rewrite [specific section] to include [specific data/context]. Remove hedging. State the recommendation directly.”

Stage 4 (Elevate): [Manual] Add your opening story, the objection you know they’ll raise, and adjust the tone to match how you actually speak.

Use these prompts in sequence. Repeat Stage 3 until the output passes the “would I say this?” test.

Why AI’s First Output Always Disappoints

AI generates presentations by predicting what words and structures typically follow your prompt. This statistical approach produces content that is, by definition, average. It’s the most likely output given your input — which means it’s also the most generic.

Here’s what that looks like in practice:

The structure is predictable. AI defaults to formats it’s seen most often: intro → three points → conclusion. Agenda slide → content → summary. This isn’t wrong, but it’s not differentiated either.

The language is safe. AI avoids strong positions because strong positions are statistically riskier. You get “consider implementing” instead of “do this.” You get “may provide benefits” instead of “will increase revenue.”

The specifics are missing. AI doesn’t know your company’s situation, your audience’s concerns, or what happened in last week’s meeting. It fills these gaps with generalities that signal “I don’t actually know your context.”

The voice is neutral. AI has no personality because personality requires consistent preferences, and AI doesn’t have those. Every output sounds like every other output — competent but anonymous.

None of these are flaws in the technology. They’re features of how language models work. Understanding this is the first step toward using AI effectively.

For more on why AI presentations fall flat, see my article on why AI presentations fail.

The 4-Stage Editing Loop

The editing loop transforms AI output from generic to specific, from safe to compelling, from anonymous to unmistakably yours. Here’s how it works:

The 4-stage AI editing loop diagram showing generate, evaluate, refine, elevate

Stage 1: Generate (Accept the Rough Draft)

Your first prompt produces raw material, not a finished product. Treat it accordingly.

At this stage, you’re looking for structural bones: Does the overall flow make sense? Are the main sections in a logical order? Is the scope roughly right?

Don’t evaluate language, tone, or specifics yet. That comes later. Right now, you just need something to work with.

Stage 2: Evaluate (Find the Gaps)

Now examine the output with specific questions:

  • Where is it generic when it should be specific?
  • Where does it hedge when it should assert?
  • What context is missing that your audience needs?
  • Which sections feel thin or underdeveloped?
  • What would a skeptical executive challenge?

Write down your findings. You’ll use them in the next stage.

Stage 3: Refine (Targeted Iteration)

This is where most people go wrong. They either accept the first output or start over from scratch. Neither works.

Instead, iterate on specific elements:

“The executive summary is too vague. Rewrite it to lead with our 23% cost reduction and include the three-month implementation timeline.”

“Section 2 hedges too much. Make the recommendation more direct. We ARE recommending Option B, not ‘suggesting it might be worth considering.'”

“Add context about the Q3 budget discussions. The CFO needs to see how this connects to her stated priorities.”

Each refinement prompt targets one issue. Stack them until the output matches your requirements.

Stage 4: Elevate (Add What AI Can’t)

AI can’t add your judgment, your stories, your voice. This final stage is where you inject what makes the presentation yours:

  • Your opening anecdote from the last board meeting
  • The specific objection you know the CTO will raise
  • The phrase your CEO always uses when she’s aligned
  • Your perspective on why this matters beyond the numbers

These human elements transform competent AI output into compelling communication.

🎯 Master the AI Editing Loop

AI-Enhanced Presentation Mastery teaches the complete iteration system — from first prompt to final polish. You’ll learn exactly how to evaluate AI output, which refinement prompts work for different problems, and how to inject your expertise into every slide.

  • The 4-stage editing loop with real examples
  • 50+ refinement prompts for common issues
  • Quality checkpoints that catch AI weaknesses
  • Live Q&A sessions for your specific challenges

Join AI-Enhanced Presentation Mastery →

Self-study modules you complete at your own pace, plus live Q&A calls for direct feedback.

Iteration Prompts That Actually Work

Generic refinement requests produce generic improvements. Here are specific prompts that solve specific problems:

For vague content:

“Replace [general statement] with specific data from [source]. Include the exact percentage and timeframe.”

For weak recommendations:

“Strengthen the recommendation in section X. Remove hedging language. State clearly: we recommend [specific action] because [specific reason].”

For missing context:

“Add a paragraph connecting this proposal to [recent event/discussion]. Reference the specific concern [stakeholder] raised about [topic].”

For bland openings:

“Rewrite the opening slide to start with the outcome, not the process. Lead with what the audience gains, not what we did.”

For generic slide titles:

“Replace ‘Overview’ with a title that states the key point. Replace ‘Analysis’ with what the analysis concluded. Every title should be a complete thought.”

For corporate-speak:

“Simplify the language in section X. Remove jargon. Write as if explaining to an intelligent person outside our industry.”

For a complete AI presentation workflow, see my guide on AI-Enhanced Presentation Mastery.

How to Know When You’re Done

The editing loop can continue indefinitely. At some point, you need to stop. Here’s how to know when:

The “would I say this?” test.

Read each slide aloud. If any sentence sounds like something you’d never actually say, it needs another iteration. When everything sounds like you wrote it, you’re close.

The specificity check.

Count the generic phrases: “various stakeholders,” “multiple benefits,” “significant improvement.” Each one represents a gap. When you can’t find any, you’re done with refinement.

The skeptic test.

Imagine your toughest audience member reading each slide. What would they challenge? What would make them roll their eyes? Address those points, or acknowledge them if they can’t be fully addressed.

The “so what?” filter.

After each major point, ask “so what?” If the answer isn’t obvious from the content, add it. When every slide answers its own “so what,” you’re done.

The time-value calculation.

Additional iterations have diminishing returns. If an iteration improves the presentation by less than the time it takes, stop. Good enough on time beats perfect too late.

For more on making AI slides feel less generic, see my article on why AI-generated slides look generic.

What’s in AI-Enhanced Presentation Mastery: The complete 4-stage editing loop with templates, 50+ refinement prompts organised by problem type, quality checkpoints that catch AI weaknesses, and real examples from executive presentations. Self-paced modules plus live Q&A sessions for direct feedback.

The Editing Loop in Practice

Here’s what a real iteration sequence looks like:

First prompt: “Create a 10-slide presentation recommending we expand into the German market.”

First output: Generic market overview, safe recommendation, no specifics about our company or situation.

Iteration 1: “Add specific data: our current European revenue (€4.2M), German market size (€890M), and our three main competitors there.”

Iteration 2: “Strengthen the recommendation. We ARE recommending Germany. Remove ‘might consider’ and ‘could potentially.’ State the investment required (€1.2M) and expected ROI (18 months to break-even).”

Iteration 3: “Add the risk section the CFO will ask about. Include currency exposure, regulatory requirements, and the contingency if growth is slower than projected.”

Iteration 4: “Rewrite slide titles. Replace ‘Market Analysis’ with ‘Germany: €890M Market, 3 Vulnerable Competitors.’ Replace ‘Recommendation’ with ‘Invest €1.2M, Break Even in 18 Months.'”

Final elevation (manual): Add opening anecdote about the German distributor who approached us at the trade show. Add the CEO’s quote about European expansion from the last town hall. Adjust tone to match how I actually present.

Total time: 45 minutes. Result: A presentation that sounds like me, addresses my specific situation, and anticipates my audience’s concerns. That’s the editing loop in action.

The AI-Enhanced Presentation Mastery course includes dozens of examples like this, showing the full iteration sequence from generic first draft to polished final product.

If you want the complete system — the full prompt library, the evaluation frameworks, and live feedback on your specific presentations — it’s all in one place.

Join AI-Enhanced Presentation Mastery →

Frequently Asked Questions

How many iteration cycles should I plan for?

For important presentations, plan for 3-5 cycles. Quick internal updates might need only 1-2. The stakes of the presentation should determine your iteration investment. A board presentation deserves more refinement than a weekly team update.

Won’t all this iteration defeat the time savings of using AI?

Even with 4-5 iteration cycles, you’re typically faster than creating from scratch — and the quality is higher because AI handles the structural work while you focus on refinement. The time savings come from not staring at a blank page, not from accepting mediocre output.

Can I use the same refinement prompts for every presentation?

Some prompts work universally (strengthen recommendations, add specifics, improve titles). Others are situation-specific. Build a personal library of prompts that solve problems you encounter repeatedly, then adapt them for each presentation.

What if the AI keeps giving me the same generic output despite iteration?

This usually means your refinement prompts are too vague. Instead of “make it better,” specify exactly what’s wrong and exactly what you want. “Replace the generic market size statement with our internal estimate of £4.2M addressable market in Q1” gets better results than “add more specific data.”

📧 Weekly AI presentation insights: Subscribe to The Winning Edge →

Related: AI can help you structure difficult messages — like presenting cost cuts without destroying trust. And if presentation anxiety affects your delivery regardless of how good your slides are, see why your nervous system remembers that awful presentation from 2019.

About the Author

Mary Beth Hazeldine is the Owner & Managing Director of Winning Presentations. With 24 years in corporate banking and consulting, she has guided senior professionals through thousands of high-stakes presentations — and now teaches them how to leverage AI without sacrificing quality or authenticity.

Her AI-Enhanced Presentation Mastery course on Maven combines practical AI workflows with the executive communication standards she developed over two decades of corporate experience.

01 Feb 2026
Professional person looking frustrated at laptop screen showing AI-generated content that doesn't sound right
AI Presentations Mary Beth Hazeldine 0 comment

Why Your AI-Generated Executive Summary Always Sounds Wrong (The 30-Second Fix)

You asked ChatGPT to write your executive summary. It took 8 seconds. Then you spent 45 minutes rewriting it because it sounded like a press release written by a committee.

The sentences were technically correct. The structure was fine. But something was off. It didn’t sound like something you’d actually say to your CFO. It didn’t sound like something anyone would say to anyone.

This isn’t an AI problem. It’s a context problem. And it takes 30 seconds to fix.

Quick answer: AI-generated executive summaries sound wrong because the AI doesn’t know your audience, your relationship with them, or what decision you’re driving toward. It fills that gap with generic corporate language. The fix isn’t better editing—it’s better context injection. Before asking for content, give the AI three things: who’s reading, what they already know, and what you need them to do. This takes 30 seconds and transforms the output.

⚡ Presenting tomorrow? Copy this prompt:

AUDIENCE: [Who’s reading—role + what they care about]
KNOWLEDGE: [What they already know about this topic]
DECISION: [What action you need them to take]
TONE: [Formal/informal + your relationship]
CONSTRAINTS: [Word count, format, company style]

Write an executive summary for: [your topic]

Fill the 5 blanks. Paste into ChatGPT/Claude/Copilot. Watch the difference.

Why AI-Written Exec Summaries Sound “Off”

Last year, I watched a client—a VP at a major retailer—spend an entire afternoon fighting with ChatGPT.

She needed an executive summary for a board presentation on warehouse automation. ChatGPT gave her something that read like a Wikipedia entry crossed with a management consulting brochure. Phrases like “leveraging synergies” and “optimising operational efficiency” that no human being has ever said out loud to another human being.

She rewrote it. Fed it back. Asked for “more natural.” Got something slightly less robotic but still wrong. Three hours later, she wrote the whole thing herself.

“AI is supposed to save time,” she told me. “I would have been faster with a blank page.”

She wasn’t wrong. But she also wasn’t using the AI correctly. The problem wasn’t the tool—it was what she didn’t tell it.

Why does AI-generated content sound generic?

AI models are trained on vast amounts of text, which means they default to the most common patterns. Without specific context, they produce “average” corporate language—technically correct but lacking the specificity and voice that makes content feel human. The more context you provide about your audience, purpose, and constraints, the more specific (and useful) the summary output becomes.

The Context Gap (What AI Doesn’t Know)

When you ask AI to “draft an exec summary for my presentation,” here’s what the AI doesn’t know:

  • Who’s reading it — A board of directors? Your direct manager? External investors? Each requires completely different framing.
  • What they already know — Are they familiar with the project? New to it? Skeptical? Supportive?
  • What decision you need — Approval? Awareness? Budget? The summary should drive toward that outcome.
  • Your relationship with them — Formal? Informal? Do you have credibility or are you building it?
  • Your organisation’s voice — Every company has unwritten rules about how executives communicate.

Without this context, AI does what any reasonable system would do: it guesses. And it guesses conservatively, using the safest, most generic language possible.

That’s why the output sounds like it was written by someone who’s never met your audience. Because, in a sense, it was.

I’ve written extensively about how to structure executive summaries in my guide to the executive summary slide—but even the best structure falls flat if the voice is wrong.

The 30-Second Fix: Context Injection

Before you ask AI to write anything, spend 30 seconds injecting context. This is the single highest-leverage change you can make to AI-assisted presentation work.

The Context Injection Framework

Add this to the beginning of any prompt:

AUDIENCE: [Who will read this, their role, what they care about]

KNOWLEDGE LEVEL: [What they already know about this topic]

DECISION NEEDED: [What action you want them to take]

TONE: [Formal/informal, relationship context]

CONSTRAINTS: [Word count, format, company style]

That’s it. Five lines. Thirty seconds to write.

The difference in output quality is dramatic—not because the AI suddenly became smarter, but because you stopped asking it to read your mind.


Context injection framework showing five elements: Audience, Knowledge Level, Decision Needed, Tone, and Constraints

How do I make ChatGPT write in my voice?

Give it examples of your voice. Paste 2-3 paragraphs you’ve written previously and say: “Match this tone and style.” Combine this with context injection (audience, decision needed, constraints) and the output will be dramatically closer to how you actually communicate. The AI can mimic voice patterns—it just needs samples.

Master AI-Enhanced Executive Presentations

AI-Enhanced Presentation Mastery is the complete system for executives who want AI to accelerate their presentation work—not create more editing. Learn context injection, voice matching, and the full workflow that cuts creation time without sacrificing quality.

Enrol in AI-Enhanced Presentation Mastery → £249

Self-study modules + live Q&A calls. Study at your own pace.

Before and After Examples

Let me show you what context injection actually does.

Scenario: Q3 Results for the Board

Without context injection:

Prompt: “Write an executive summary for my Q3 results presentation.”

“This executive summary provides an overview of Q3 performance metrics and key achievements. The quarter demonstrated strong operational efficiency and strategic alignment with organisational objectives. Key highlights include revenue growth, cost optimisation initiatives, and enhanced customer satisfaction scores…”

Generic. Forgettable. Could be any company, any quarter, any presenter.

With context injection:

AUDIENCE: Board of directors, 8 people, mostly focused on margin and cash flow

KNOWLEDGE LEVEL: They saw Q2 results; know we’re behind on EBITDA target

DECISION NEEDED: Approve additional headcount for Q4 push

TONE: Direct, confident, acknowledge the miss but pivot to solution

CONSTRAINTS: 150 words max, no jargon, lead with the ask

Write an executive summary for my Q3 results presentation.

“Q3 EBITDA came in 12% below target. You’ll want to know why—and what we’re doing about it. The miss was driven by delayed product launches (now resolved) and higher-than-expected customer acquisition costs in APAC. Despite this, revenue grew 8% QoQ, and we’ve identified a clear path to recover margin in Q4. That path requires 6 additional headcount in sales engineering. This presentation makes the case for that investment and shows the projected impact on Q4 and FY targets. I’m asking for approval today so we can begin hiring next week.”

Same AI. Same request. Completely different output.

The second version sounds like a real person talking to real people about a real situation. It leads with the uncomfortable truth, acknowledges what the audience cares about, and drives toward a specific decision.

That’s what context injection does. It turns AI from a generic content machine into a tool that understands your specific communication challenge.

Want the complete library of context injection templates for every presentation type?

Get AI-Enhanced Presentation Mastery → £249

What context does AI need for executive presentations?

At minimum: who’s reading (role and what they care about), what they already know, and what decision you need. Adding tone guidance and constraints (word count, format) improves output further. The more specific your context, the less editing you’ll need. Think of it as briefing a smart but uninformed colleague—they need background before they can help.

Beyond Summaries: The Full Workflow

Context injection works for executive summaries, but it’s actually the foundation of a complete AI-assisted presentation workflow.

The Three-Layer Approach

Layer 1: Strategic Context (before any content)

Define your audience, decision, and constraints. This shapes everything that follows.

Layer 2: Structural Generation

Use AI to generate slide structures, not content. “Given this context, what are the 8 slides I need?” is a better prompt than “Write my presentation.”

Layer 3: Content Refinement

Generate content slide-by-slide, with context injection for each. Review and refine in passes, not all at once.

This approach typically cuts presentation creation time by 50-70%—not because AI writes everything, but because it handles the parts that don’t require your judgment while you focus on the parts that do.

I cover the full workflow in detail in my guide to using ChatGPT for PowerPoint presentations—including the specific prompts for each layer.

When AI Isn’t the Answer

Context injection dramatically improves AI output, but some elements of executive presentations still require human judgment:

  • Political navigation — AI doesn’t know that the CFO and COO are feuding, or that the CEO hates bullet points
  • Stakeholder relationships — The history between you and your audience shapes how you frame sensitive topics
  • Strategic ambiguity — Sometimes you need to be deliberately vague; AI defaults to clarity
  • Emotional calibration — Delivering bad news, building urgency, or inspiring action requires human touch

The goal isn’t to automate everything. It’s to automate the parts that don’t need you, so you can invest your judgment where it matters.

For more on the strategic side of executive presentations, see my article on AI for presentations.

The Compound Effect

Here’s what most people miss about AI-assisted presentations: the benefit compounds.

Once you have a context injection template for board presentations, you reuse it. Once you’ve trained AI on your voice with sample paragraphs, you can reference that conversation. Once you’ve built a library of prompts that work for your organisation’s style, every presentation gets faster.

The first presentation might save you 30 minutes. The tenth saves you 3 hours. The fiftieth is a completely different workflow—one where AI handles the scaffolding and you focus purely on strategic decisions and refinement.

That’s the real promise of AI for executive presentations. Not “AI writes your presentation.” But “AI handles the 80% that doesn’t need your brain, so your brain can focus on the 20% that does.”

Stop Fighting With AI. Start Collaborating.

AI-Enhanced Presentation Mastery teaches you the complete workflow: context injection templates, voice matching techniques, structural generation, and the refinement process that produces executive-ready output. Self-study modules you can complete at your own pace, plus live Q&A calls for personalised guidance.

Enrol in AI-Enhanced Presentation Mastery → £249

Created from 24 years of executive presentation experience combined with systematic AI workflow development.

Frequently Asked Questions

Will this work with any AI tool (ChatGPT, Claude, Copilot)?

Yes. Context injection is model-agnostic—it works with ChatGPT, Claude, Copilot, Gemini, and any other large language model. The principle is the same: AI produces better output when you give it better input. The specific prompts in AI-Enhanced Presentation Mastery are tested across multiple tools so you can use whichever your organisation prefers.

How long does it take to learn the context injection method?

The basic framework takes about 15 minutes to understand and apply. You’ll see improved output immediately. Mastering the nuances—when to add more context, how to iterate, how to build reusable templates—takes longer, typically 2-3 weeks of regular practice. The course accelerates this with pre-built templates and worked examples.

What if my company has a specific presentation style?

That’s actually ideal. Feed the AI examples of presentations your company has approved. Include style guidelines in your context injection. The more specific you are about organisational norms, the better the output matches. Many course participants create company-specific template libraries they reuse across their teams.

Is this different from prompt engineering courses?

Yes. General prompt engineering teaches principles that apply across use cases. AI-Enhanced Presentation Mastery is specifically designed for executive presentations—the context injection frameworks, the structural prompts, the refinement workflows are all built for the specific challenge of creating high-stakes business presentations. It’s specialised, not general.

📬 Get Weekly AI + Presentation Insights

Join professionals who receive my weekly newsletter on AI-enhanced presentations, executive communication, and high-stakes delivery.

Subscribe to The Winning Edge →

Related reading:

📋 Free Resource: 10 Essential AI Prompts for Presentations

Not ready for the full course? Start with my free prompt library—10 tested prompts for common presentation tasks, including context injection templates you can use immediately.

Get the Free AI Prompts →

Your Next Step

The next time you need an executive summary, don’t start with “Write an executive summary.”

Start with 30 seconds of context injection. Tell the AI who’s reading, what they know, and what decision you need.

Watch what happens to the output.

And if you want the complete system—not just context injection, but the full workflow that transforms how you create executive presentations—AI-Enhanced Presentation Mastery will show you how.

AI is a tool. The question is whether you’re using it as a content generator or a thought partner. Context injection is the difference.

About the Author

Mary Beth Hazeldine is the Owner & Managing Director of Winning Presentations. With 24 years of corporate banking experience at JPMorgan Chase, PwC, Royal Bank of Scotland, and Commerzbank, she has delivered hundreds of high-stakes executive presentations—and now teaches professionals how to use AI to create them more efficiently.

A qualified clinical hypnotherapist and NLP practitioner, Mary Beth combines executive communication expertise with systematic AI workflow development. She has helped senior professionals and teams transform their presentation process.

Book a discovery call | View services

19 Jan 2026
Why AI-generated slides look generic - the framework-first fix for executive-quality presentations
AI Presentations Mary Beth Hazeldine 0 comment

Why Your AI-Generated Slides Look Generic (And How to Fix It)

Quick answer: Your AI-generated slides look generic because you’re asking AI to do the thinking for you. The tool isn’t broken—the input is. When you prompt AI without a clear framework (structure, audience, decision point), it defaults to safe, templated output. The fix isn’t better prompts. It’s building your presentation framework first, then using AI to accelerate execution.

This fixes the endless cycle of generate → cringe → delete → redo that wastes hours and leaves you with slides you’re embarrassed to present.

⚡ Need to fix generic AI slides right now? Do this before your next prompt:

Step 1: Write your main message in one sentence (what do you want them to decide/believe?)

Step 2: List your 3 supporting points in order of importance

Step 3: Identify your audience’s #1 objection

Step 4: NOW prompt AI with this structure—watch the output transform

The £2M Pitch That AI Almost Ruined

A client came to me last year in a panic. She’d used AI to create her investor pitch deck—Gamma for the slides, ChatGPT for the script. The output looked polished. Professional fonts, clean layouts, smooth transitions.

The investors passed in under five minutes.

“It felt like every other pitch we’ve seen this month,” one told her. “Nothing stood out.”

That’s the trap. AI-generated slides look generic not because the tools are bad, but because they’re designed to be safe. They optimise for “acceptable to everyone” rather than “compelling to your specific audience.”

Six weeks later, we rebuilt her deck using a framework-first approach. Same information. Same AI tools for execution. Different result: £2.1M raised.

The AI didn’t change. Her input did.

⭐ Master the Framework That Makes AI Output Executive-Ready

Stop fighting with prompts. Learn the structure-first methodology that transforms any AI tool from “generic template generator” to “presentation accelerator.”

In this live cohort course:

  • The Decision-First Framework for AI-enhanced presentations
  • How to brief AI tools so they produce executive-quality output
  • Live feedback on your actual presentations
  • Templates that work with Copilot, Gamma, ChatGPT, and any future tool

Join AI-Enhanced Presentation Mastery →

Live cohort course with Mary Beth Hazeldine. Limited seats per session. Framework-first methodology tested across banking, consulting, and FTSE 100 environments.

If you have an investor pitch, board deck, or QBR in the next 2–3 weeks, this will pay for itself immediately.

Why Every AI Tool Produces Generic Output

Here’s what most people don’t understand about AI presentation tools: they’re trained on millions of slides, which means they’ve learned to produce the average of all those slides.

Average is, by definition, generic.

When you prompt Copilot with “Create a presentation about Q3 results,” it generates what a Q3 presentation typically looks like—across thousands of companies, industries, and contexts. It doesn’t know your audience is a skeptical CFO. It doesn’t know your Q3 results contain a critical pivot point. It doesn’t know the board has seen 47 similar presentations this month.

So it gives you:

  • Safe bullet points that could apply to any company
  • Stock imagery that signals “corporate presentation”
  • Slide titles like “Overview” and “Key Takeaways” that tell the audience nothing
  • A structure that builds to a conclusion (when executives want conclusions first)

This isn’t a flaw in the AI. It’s working exactly as designed. The problem is the input, not the tool.

If you’ve tried fixing generic Copilot slides with better prompts, you’ve probably noticed: better prompts help marginally. They don’t solve the core problem.

The Framework-First Method That Changes Everything

The executives I’ve trained over 24 years in banking don’t start with slides. They don’t start with AI prompts. They start with a framework.

Framework-first means answering these questions before you touch any tool:

1. What’s the one decision I need from this audience?

Not “inform them about Q3.” A specific decision: “Approve the £500K investment in the new system.”

2. What’s their biggest objection or concern?

A CFO worries about ROI. A board worries about risk. A client worries about implementation. Name it.

3. What evidence will overcome that objection?

Not all your data. The specific proof points that address their specific concern.

4. What’s the logical flow that leads to yes?

Decision → Impact → Risk mitigation → Evidence. This is the executive presentation structure that actually works.

Once you have this framework, AI becomes extraordinarily useful. You’re not asking it to think for you. You’re asking it to execute your thinking faster.

Instead of prompting: “Create a presentation about our new CRM system”

Prompt with framework: “Create a 6-slide presentation for our CFO requesting £500K for a CRM upgrade. Main message: this investment pays back in 14 months through reduced customer churn. Address the objection that implementation will disrupt Q4 sales. Structure: recommendation first, then ROI evidence, then risk mitigation, then timeline.”

The output from the second prompt is unrecognisable from the first—even though it’s the same AI tool.

Want to master framework-first AI presentations? AI-Enhanced Presentation Mastery is a live cohort course that teaches the complete methodology—with feedback on your actual presentations. See upcoming sessions →

Before and After: Same Tool, Different Input

Here’s what the framework-first difference looks like in practice:

BEFORE (prompt-first approach):

Prompt:

“Create a presentation about implementing a new project management system”

AI Output:

  • Slide 1: Title slide with generic stock image
  • Slide 2: “Agenda” (why do executives need an agenda for 8 slides?)
  • Slide 3: “Current Challenges” (vague bullet points)
  • Slide 4: “Proposed Solution” (feature list)
  • Slide 5: “Benefits” (generic claims)
  • Slide 6: “Implementation Timeline” (Gantt chart)
  • Slide 7: “Budget Overview” (numbers without context)
  • Slide 8: “Next Steps” / “Questions?”

AFTER (framework-first approach):

Framework completed first:

Decision: Approve £85K for project management system. Audience: COO + Finance Director. Main objection: disruption to current workflow. Key evidence: 23% productivity gain from pilot team.

Prompt:

“Create a 6-slide executive presentation requesting £85K budget approval for a project management system. Lead with the recommendation and expected ROI. Address workflow disruption concerns by showing pilot results. Include risk mitigation. Audience is COO and Finance Director who value efficiency metrics.”

AI Output:

  • Slide 1: “Recommendation: Approve £85K—Expected 340% ROI in 18 months”
  • Slide 2: Pilot results showing 23% productivity gain
  • Slide 3: Workflow disruption mitigation plan
  • Slide 4: Financial breakdown with payback timeline
  • Slide 5: Risk assessment with contingencies
  • Slide 6: Decision requested + implementation start date

Same AI. Same topic. Completely different output. The difference is worth thousands in approved budgets and closed deals. Learning to create framework-first presentations can transform how decision-makers perceive your proposals—and your readiness for senior roles.


Framework-first vs prompt-first approach comparison showing how the same AI tool produces generic versus executive-quality slides based on input quality

⭐ Stop Producing Slides That Look Like Everyone Else’s

The framework-first methodology works with any AI tool—because it fixes the input, not the technology. Learn it once, apply it forever.

What you’ll master:

  • The 4-question framework that transforms AI output
  • Executive presentation structures that work across industries
  • How to brief any AI tool for professional results
  • Live practice with real-time feedback

Join the Next Cohort →

Live sessions + async practice. Includes templates, frameworks, and direct feedback on your presentations.

Which AI Tool Actually Matters? (Hint: None of Them)

People ask me constantly: “Should I use Copilot or Gamma? Is ChatGPT better than Claude for slides? What about Beautiful.ai?”

The honest answer: it barely matters.

Every major AI tool can produce executive-quality slides—if you give it executive-quality input. And every tool will produce generic output if you give it generic prompts.

The tools will keep changing. Copilot will update. New competitors will launch. GPT-6 will arrive. But the framework-first methodology stays constant because it’s based on how humans make decisions, not how AI generates content.

This is why I teach frameworks that are tool-agnostic. My clients use the same methodology whether they’re in Copilot, Gamma, or building slides manually. The AI presentation workflow accelerates execution, but the thinking happens before any tool is opened.

What to ask instead of “which tool is best?”:

  • “Do I have a clear decision I’m asking for?”
  • “Have I identified my audience’s main objection?”
  • “Do I know the evidence that overcomes that objection?”
  • “Is my structure decision-first or conclusion-last?”

Answer those questions, and any AI tool will serve you well.

Ready to master framework-first presentations? AI-Enhanced Presentation Mastery teaches the complete system—70% framework thinking, 30% AI execution. Works with any tool, now and in the future. View course details →

Related: Once your slides are executive-ready, make sure your structure and delivery match. Read Executive Presentation Structure: The Format That Gets Instant Buy-In and How to Stop Saying Um (Without Sounding Robotic).

Common Questions About AI-Generated Slides

Why do AI presentations look so generic?

AI tools are trained on millions of slides, so they produce the statistical average of all presentations. Average means generic. The tool optimises for “safe and acceptable” rather than “compelling for your specific audience.” To get non-generic output, you must provide specific input: the decision you need, the objection you’re addressing, and the evidence that overcomes it.

How do I make AI-generated slides look professional?

The secret isn’t in the prompts—it’s in the framework you create before prompting. Define your one key decision, your audience’s main concern, and your supporting evidence structure. Then prompt AI with this specific context. The same tool that produces generic bullet points will produce executive-ready slides when given framework-quality input.

What’s wrong with AI presentation tools?

Nothing is wrong with the tools. Copilot, Gamma, ChatGPT, and others are all capable of producing excellent output. The problem is how most people use them—asking AI to think instead of asking AI to execute. When you do the strategic thinking first (framework) and use AI for tactical execution (slides), the results transform completely.

⭐ Create Presentations That Don’t Look AI-Generated

Learn the methodology that makes AI your presentation accelerator—not your presentation liability.

Inside the course:

  • The Decision-First Framework (works with any AI tool)
  • Executive presentation templates with prompting guides
  • Live cohort sessions with direct feedback
  • How to brief AI for boardroom-quality output

Enroll in AI-Enhanced Presentation Mastery →

Live cohort format with Mary Beth Hazeldine. Framework-first methodology developed from 24 years in corporate banking and executive coaching.

FAQ

Which AI tool is best for presentations?

The tool matters far less than the input. Copilot, Gamma, ChatGPT, Beautiful.ai, and Canva’s AI features can all produce excellent presentations—if you give them framework-quality input. Choose based on what integrates with your workflow (Copilot for Microsoft users, Gamma for standalone, etc.), not based on which “produces the best slides.” They all produce generic slides with generic prompts.

Can AI really create executive-quality slides?

Yes—but only when you provide executive-quality thinking first. AI excels at execution: formatting, visual consistency, generating variations quickly. It struggles with strategy: understanding your specific audience, identifying the key decision, structuring for persuasion. Do the strategy yourself, use AI for execution, and the output will impress executives.

How long does the framework-first approach take?

About 10-15 minutes of structured thinking before you open any tool. This feels slower initially but dramatically reduces total time. You eliminate the “generate, delete, regenerate” cycle that wastes hours. Most of my clients report cutting total presentation creation time by 40-60% once the framework-first approach becomes habit.

Will this work with Copilot/Gamma/ChatGPT?

The framework-first methodology works with any AI tool because it focuses on input quality, not tool features. I’ve tested it extensively with Copilot, Gamma, ChatGPT, Claude, and several others. The specific prompting syntax varies slightly by tool, but the core framework remains identical. Learn the framework once, adapt to any tool.

📧 The Winning Edge Newsletter

Weekly insights on AI-enhanced presentations, executive communication, and framework-first thinking. Practical techniques from 24 years in corporate banking—no AI hype, just what actually works.

Subscribe Free →

Your Next Step

Your AI-generated slides look generic because AI is doing what it’s designed to do: produce safe, average output. The fix isn’t a better tool or better prompts. It’s better input.

Before your next presentation, take 10 minutes to answer the framework questions: What decision do you need? What’s the main objection? What evidence overcomes it? What’s the logical structure?

Then prompt AI with that framework. The output will transform—and so will how your audience responds.

If you want to master the complete framework-first methodology with live feedback and executive-tested templates, join AI-Enhanced Presentation Mastery.

About the Author

Mary Beth Hazeldine is the Owner & Managing Director of Winning Presentations and a former corporate banker with 24 years of experience at JPMorgan Chase, PwC, Royal Bank of Scotland, and Commerzbank. She has trained thousands of executives on high-stakes presentation skills and helped clients secure more than £250 million in funding and budget approvals.

Mary Beth is also a qualified clinical hypnotherapist and NLP practitioner, specialising in helping professionals overcome presentation anxiety. She developed the framework-first AI methodology after seeing countless executives struggle with generic AI output—and discovering that the fix was strategic thinking, not better technology.

Book a discovery call | View services

21 Dec 2025
How to use AI for presentations - complete guide to saving hours and creating better slides with AI tools
AI Tools, Presentation Skills, Productivity Mary Beth Hazeldine 0 comment

How to Use AI for Presentations: Save Hours and Create Better Slides

A practical guide to using AI for presentations — with 50+ prompts, proven frameworks, and a complete workflow from a presentation skills trainer

If you want to learn how to use AI for presentations effectively, you’re in the right place. Most professionals are either ignoring AI completely or using it badly — getting generic content that sounds like a robot wrote it.

There’s a better way.

Last month, I watched a senior consultant spend an entire Sunday preparing a 20-minute client presentation. Research. Structure. Slides. Rewrites. More rewrites. Eight hours for twenty minutes of content.

The following week, I helped another consultant prepare a similar presentation. We used AI strategically throughout the process.

Total time: 90 minutes. And honestly? The second presentation was better.

This isn’t about AI replacing your skills. It’s about AI amplifying them — so you create better presentations in a fraction of the time. After 24 years of corporate presenting and training over 5,000 executives, I’ve developed a systematic approach to using AI for presentations that actually works.

🎁 Free Download: Get my Executive Presentation Checklist — includes the AI prompts I use for rapid presentation preparation.

Why Most People Use AI for Presentations Wrong

Here’s what traditional presentation preparation looks like:

  1. Stare at blank slides
  2. Write too much content
  3. Reorganize everything
  4. Cut half of it
  5. Redesign slides
  6. Practice and realize the structure doesn’t flow
  7. Reorganize again
  8. Run out of time
  9. Deliver something “good enough”

Sound familiar?

Now here’s what most people do when they try using AI for presentations: they ask ChatGPT to “write a presentation about X” and get generic, bloated content that sounds nothing like them.

The problem isn’t AI. It’s how they’re using it.

AI works when you use it for specific tasks within a proven framework — not as a magic button that does everything.

Related: Microsoft Copilot for Presentations: What Works and What Doesn’t

AI presentation tools workflow showing how to use AI for research, structure, content, and Q&A preparation

The Right Way to Use AI for Presentations

AI changes presentation preparation completely — but not by doing the work for you. It accelerates every step of a proven process:

  • Research that took 2 hours now takes 15 minutes
  • First drafts that took an afternoon now take 20 minutes
  • Anticipating questions becomes systematic, not guesswork
  • Structure emerges quickly instead of through painful iteration

The result? Better presentations in less time. And when you’re well-prepared with a solid structure, you naturally feel more confident delivering it.

Here’s the framework I teach:

Step 1: Start With Structure (Before You Touch AI)

Before you use any AI tool, you need to know what you’re building. I use a simple 3-part framework that works for any presentation:

  • Opening: Hook them in 30 seconds with a problem, question, or surprising fact
  • Body: 3-5 key points maximum (one idea per slide)
  • Close: Clear call to action or key takeaway

This takes 5 minutes to sketch out — and it transforms how you use AI because now you have specific sections to fill, not a blank page.

Related: Presentation Structure: The 3-Part Framework That Works Every Time

Step 2: Use AI for Research and Content Generation

Now AI becomes powerful. Instead of “write me a presentation,” you prompt:

  • “Give me 5 compelling statistics about [topic] that would surprise a senior executive”
  • “What are the 3 strongest counterarguments to [my recommendation] and how would I address them?”
  • “Write a 2-sentence opening hook for a presentation about [topic] to [audience]”

Specific prompts = useful outputs. Generic prompts = generic garbage.

Step 3: Use AI for Q&A Preparation

This is where AI saves the most stress. Prompt:

“I’m presenting [topic] to [audience]. What are the 10 toughest questions they might ask, and give me a 2-sentence answer for each.”

You’ll walk in prepared for questions you never would have anticipated.

Step 4: Refine (Don’t Use Raw AI Output)

Raw AI content sounds like AI. Your job is to:

  • Add your stories and examples
  • Cut the filler words AI loves
  • Adjust the tone to sound like you
  • Verify any facts or statistics

AI does the heavy lifting. You add the human elements that make presentations land.

Related: 10 ChatGPT Prompts for Better Presentations

AI for presentations time savings - preparation reduced from 6-8 hours to 90 minutes with AI workflow

Want the Complete AI Presentation System?

My AI-Enhanced Presentation Mastery course gives you the full framework — 50+ tested prompts, proven structures for any presentation type, and live coaching to apply it to your specific work.

What’s included:

  • 4 weeks of structured curriculum (frameworks + AI tools)
  • 50+ copy-paste AI prompts for research, structure, content, and Q&A
  • 2 live coaching sessions with personalized feedback
  • Community access for peer support
  • Lifetime access and all future updates

January cohort: £249 (increases to £499 in April)

Only 60 seats. Early bird ends December 31st.

See the full curriculum →

Executive Resource

Stop Writing AI Prompts From Scratch

The Executive Prompt Pack gives you 50 battle-tested prompts for executive-level presentations — board updates, budget requests, investor briefs, and Q&A preparation. Built for PowerPoint Copilot and ChatGPT.

Get the Executive Prompt Pack →

Used by executives preparing for board briefings, budget requests, and high-stakes presentations.

Best AI Tools for Presentations in 2025

You don’t need expensive tools to use AI for presentations effectively. Here’s what actually works:

For Research and Content

ChatGPT (Free or Plus): Best for brainstorming, research synthesis, and generating first drafts. The free version works fine for most tasks.

Claude: Better for longer, more nuanced content. Excellent for refining messaging and anticipating objections.

Perplexity: Best for research with sources. Use when you need verified facts and statistics.

For Slides

Microsoft Copilot for PowerPoint: Creates slides from prompts or documents. Good for first drafts, but requires heavy editing. Best if you’re already in the Microsoft ecosystem.

Gamma: Creates beautiful presentations from prompts. Better design than Copilot, but less control over structure.

Your existing tools + AI-generated content: Often the best approach. Use AI to create the content, then build slides in whatever tool you already know.

Related: Microsoft Copilot for PowerPoint: Complete Guide

My Recommendation

Start with ChatGPT or Claude for content, and your existing slide tool. Don’t add complexity until you’ve mastered the fundamentals. The prompts matter more than the tools.

Complete AI Presentation Workflow: Step by Step

Here’s exactly how I use AI for presentations — the same workflow I teach in my course:

Phase 1: Preparation (15 minutes)

  1. Define your audience and their key concerns
  2. Clarify your one main message (if they remember one thing, what is it?)
  3. Sketch the 3-part structure: hook, 3-5 key points, close

Phase 2: AI-Assisted Content Creation (30-45 minutes)

  1. Use AI for research: statistics, examples, counterarguments
  2. Generate first draft content for each section
  3. Create your opening hook (test 3-5 options)
  4. Prepare Q&A responses for tough questions

Phase 3: Refinement (30 minutes)

  1. Add your personal stories and examples
  2. Cut anything that doesn’t serve your main message
  3. Adjust tone to sound like you
  4. Verify facts and statistics

Phase 4: Slides (20-30 minutes)

  1. One idea per slide
  2. Minimal text (your words, not the slides, do the work)
  3. Use AI-generated content as speaker notes, not slide text

Total time: 90 minutes to 2 hours for a presentation that used to take 6-8 hours.

“The AI workflow alone was worth the course fee. I used to spend entire weekends preparing for Monday presentations. Now I do it in a couple of hours on Friday afternoon. The prompts are incredibly specific and practical.”

— James T., Product Manager

Common Mistakes When Using AI for Presentations

Avoid these errors that make AI-generated presentations sound robotic:

Mistake 1: Using AI output without editing. Raw AI content is generic. Always add your voice, stories, and specific examples.

Mistake 2: Prompting too broadly. “Write me a presentation” gives you garbage. “Write a 2-sentence hook for [specific audience] about [specific topic]” gives you gold.

Mistake 3: Skipping the structure step. AI can’t read your mind about what the presentation needs to accomplish. Define structure first, then use AI to fill sections.

Mistake 4: Trusting AI facts without verification. AI makes things up. Always verify statistics, quotes, and specific claims.

Mistake 5: Putting AI text directly on slides. AI-generated text belongs in your speaker notes or script, not on the slides your audience sees.

Related: The 10 Presentation Mistakes That Kill Your Credibility

“I was skeptical about AI for presentations — I thought it would make everything sound generic. But Mary Beth’s approach is different. The AI accelerates the slow parts (research, first drafts, Q&A prep) while you keep control of what matters (story, strategy, voice). My presentations are better AND faster now.”

— Rachel K., Strategy Consultant

AI Presentation Prompts That Actually Work

Here are 10 prompts from my collection of 50+ that I use regularly:

For Research

1. “Give me 5 surprising statistics about [topic] that would make a [job title] pay attention. Include sources.”

2. “What are the 3 biggest misconceptions about [topic] that my audience of [description] probably believes?”

For Structure

3. “I need to present [topic] to [audience] in [X] minutes. Give me a structured outline with timing for each section.”

4. “What’s the most compelling order to present these 5 points: [list points]? Explain your reasoning.”

For Opening Hooks

5. “Write 5 different opening hooks for a presentation about [topic] to [audience]. Include: a surprising statistic, a provocative question, a brief story, a counterintuitive statement, and a vivid scenario.”

For Q&A Preparation

6. “I’m presenting [recommendation] to [audience]. What are the 10 toughest questions they might ask? Give me a confident 2-sentence response for each.”

7. “What are the strongest objections to [my proposal] and how would I address each one?”

For Storytelling

8. “Help me turn this data point [insert data] into a brief story that illustrates why it matters to [audience].”

For Slides

9. “Reduce this paragraph to a 6-word slide headline that captures the key message: [paste paragraph]”

10. “What visual or diagram would best illustrate this concept: [describe concept]?”

The full course includes 50+ prompts across research, structure, storytelling, slides, and Q&A — plus the context for when and how to use each one.

Frequently Asked Questions About Using AI for Presentations

Can AI create an entire presentation for me?

Technically yes, but you shouldn’t let it. AI-generated presentations without human refinement sound generic and miss the nuances of your specific audience and message. Use AI for the time-consuming parts (research, first drafts, Q&A prep) and add the human elements yourself (stories, insights, your voice).

What’s the best AI tool for presentations?

For content creation, ChatGPT and Claude are both excellent — and free tiers work fine. For slides, Microsoft Copilot works if you’re already in PowerPoint; Gamma creates better-looking slides but with less control. Start with AI for content + your existing slide tool before adding new platforms.

How do I make AI-generated content sound like me?

Three techniques: First, give AI examples of your previous writing and ask it to match the tone. Second, always edit AI output to add your specific stories and examples. Third, read the content aloud — if it doesn’t sound like something you’d say, rewrite it until it does.

Will my audience know I used AI?

Not if you use it correctly. When you use AI for research and first drafts, then add your own stories, examples, and voice, the result is distinctly yours. The only presentations that “sound like AI” are ones where someone used raw AI output without refinement.

How much time can AI really save on presentations?

In my experience and my students’ experience: 60-70%. A presentation that took 6-8 hours typically takes 2-3 hours with a proper AI workflow. The biggest time savings come from research (AI synthesizes information faster), first drafts (no more staring at blank pages), and Q&A prep (systematic instead of guesswork).

“I was preparing a board presentation and dreading the usual weekend of work. Used the Week 3 prompts and had a solid draft in 45 minutes. The frameworks from Week 1 meant I knew exactly what to include. Game changer.”

— David L., Finance Director

AI-Enhanced Presentation Mastery course overview - 4 weeks covering structure, storytelling, AI tools, and delivery

Learn the Complete AI Presentation System

This article covers the fundamentals — but there’s much more to master.

My AI-Enhanced Presentation Mastery course gives you the complete system:

Week 1: Structure That Works Every Time

Proven frameworks for client pitches, board updates, team meetings, and keynotes. The foundation that makes AI useful (instead of a source of generic content).

Week 2: Storytelling That Connects

How to turn data into compelling narratives. Finding stories in “boring” business content. The emotional arc that keeps audiences engaged.

Week 3: AI-Powered Preparation

50+ prompts for research, structure, storytelling, and slides. My complete workflow for client presentations. How to refine AI output so it sounds like you.

Week 4: Delivery and Executive Presence

Present your well-prepared content with confidence. Handle Q&A (including “I don’t know”). Virtual and in-person techniques.

Your Next Step: Master AI for Presentations

You now have a complete framework for using AI to create better presentations in less time. But knowledge isn’t transformation — implementation is.

Choose your path:

🎁 START FREE: Download the Executive Presentation Checklist — includes AI prompts for rapid preparation.

📘 GO DEEPER (£39): Get Presentations with AI: The Complete Prompt Collection — 50+ prompts with examples and use cases.

🎓 GET THE FULL SYSTEM (£249): Join AI-Enhanced Presentation Mastery — 4 weeks of curriculum, live coaching, community, and personalized feedback. Early bird ends December 31st.

AI is changing how presentations get made. The professionals who master this now will have a significant advantage over those still spending weekends on slide decks.

AI-Enhanced Presentation Mastery

Proven frameworks + 50+ AI prompts + Live coaching

£249 £499

Early bird ends December 31st • 60 seats • Full refund guarantee

Enroll Now →

Mary Beth Hazeldine is Managing Director of Winning Presentations, with 24 years of corporate banking experience at JPMorgan, PwC, Royal Bank of Scotland, and Commerzbank. She’s trained over 5,000 executives in presentation skills and specializes in AI-powered presentation techniques — testing every method on real client work before teaching it.

15 Dec 2025
Why AI presentations fail - the hidden problem with AI-generated slides and how to fix them
AI Tools, Presentation Skills Mary Beth Hazeldine 0 comment

Why AI Presentations Fail (And How to Fix Them)

📅 Updated: December 2025

Why AI presentations fail - the hidden problem with AI-generated slides and how to fix them

Why AI Presentations Fail (And How to Fix Them)

Want a structured framework for this?

71 structured prompts for ChatGPT and Copilot — covering board decks, investor pitches, quarterly reviews, and strategy presentations.

Explore the Prompt Pack →

Quick Answer

AI presentations fail because they optimise for speed, not persuasion. Tools like Copilot, ChatGPT, and Gamma generate slides in seconds — but the output is generic, forgettable, and often counterproductive. The fix isn’t avoiding AI; it’s using frameworks first (AVP, 132 Rule, S.E.E. Formula) and AI second. This article explains why most AI-generated presentations underperform and the 4-step system to make yours actually work.

Executive Resource

Stop Writing AI Prompts From Scratch

The Executive Prompt Pack (£19.99) gives you 71 tested prompts for executive-level presentations — board updates, budget requests, investor briefs, and Q&A preparation. Built for PowerPoint Copilot and ChatGPT.

Get the Executive Prompt Pack →

Designed for executives preparing for board briefings, budget requests, and high-stakes presentations.

🎁 FREE DOWNLOAD

Executive Presentation Checklist

The 12-point framework that makes AI presentations actually persuade. Complete this BEFORE you prompt any AI tool.

Download Free Checklist →

No spam. Unsubscribe anytime.

AI presentation tools promise to save you hours. And they do — if you measure success by how fast you create slides.

But speed isn’t the goal. Persuasion is. Decisions are. Results are.

And by those measures, most AI presentations fail spectacularly.

I’ve trained executives on presentations for more than 16 years. In the last two years, I’ve watched AI tools transform how people create slides — and I’ve seen the results. The presentations are faster to create. They’re also worse at persuading.

Here’s what’s actually happening, and how to fix it.

The 5 Reasons AI Presentations Fail

1. AI Optimises for Completeness, Not Clarity

Ask ChatGPT or Copilot to create a presentation about your product, and you’ll get comprehensive slides covering every feature, benefit, and use case.

The problem? Comprehensive isn’t persuasive.

Human attention is limited. The best presentations focus ruthlessly on 2-3 key messages. AI doesn’t know which messages matter most to YOUR audience in THIS context. So it includes everything — which means nothing stands out.

The result: Your audience remembers nothing. The decision gets delayed. You’ve saved 4 hours of creation time and lost 4 weeks of momentum.

2. AI Can’t Read the Room

A CFO cares about ROI and risk. A technical buyer cares about integration and security. A CEO cares about strategic fit and competitive advantage.

AI doesn’t know who’s in the room. It generates generic content for a generic audience — which resonates with no one specifically.

I recently reviewed a sales deck created with Copilot for a client pitching a private equity firm. Beautifully formatted. Professionally structured. And completely wrong for the audience — they wanted 3 slides on financial returns, not 15 slides on product features. The deal went to a competitor who understood what the audience actually wanted.

The result: The AI presentation looked professional but felt tone-deaf.

3. AI Produces “Correct” But Forgettable Content

AI-generated text is grammatically perfect and factually accurate. It’s also utterly forgettable.

Why? Because AI optimises for the average of all presentations it’s trained on. It produces the most statistically likely content — which is, by definition, the most generic.

Great presentations aren’t average. They have a point of view. They take a stance. They make you think. AI doesn’t do that — unless you specifically prompt it to, and most people don’t.

The result: Your slides look like everyone else’s slides. In a competitive pitch, you blend in when you need to stand out.

5 reasons AI presentations fail - completeness over clarity, generic content, no audience awareness, missing structure, false confidence

4. AI Skips the Strategic Thinking

The hardest part of a presentation isn’t making slides. It’s deciding what to say.

What’s your core message? What action do you want? What objections will arise? What story ties it together?

AI tools skip this entirely. They jump straight to slide creation — which is like writing a novel by generating sentences without knowing the plot.

When I work with clients, we spend 70% of our time on strategy and 30% on slides. AI inverts this ratio. You spend 5 minutes prompting and get 20 slides — none of which answer the fundamental question: “Why should this audience care?”

5. AI Creates False Confidence

This might be the most dangerous failure mode.

When you struggle to create a presentation manually, you’re forced to think. You wrestle with structure. You cut slides that don’t work. You refine your message through iteration.

AI eliminates that productive struggle. You get a polished-looking deck in minutes and assume it’s ready. But “looks professional” isn’t the same as “will persuade.”

I’ve seen executives walk into board meetings with AI-generated decks that looked beautiful and completely failed to land. They trusted the tool instead of testing the thinking.

📄
Fix Your AI Presentations

Start with the free checklist — complete this BEFORE prompting any AI tool. It’s the strategic thinking AI can’t do for you.

Download Free Checklist →

If you want a structured approach to executive presentations, the Executive Prompt Pack gives you everything you need — £19.99, instant access.

The Hidden Costs of Failed AI Presentations

When AI presentations fail, the costs are real — even if they’re invisible.

Lost revenue: A SaaS company I worked with had a 23% close rate with AI-generated decks. We restructured their pitch around the AVP framework (Action-Value-Proof) and their close rate hit 34%. On an £8M pipeline, that’s an £880K swing — from changing how they presented the same product.

Wasted time: The promise of AI is saving time. But if your AI presentation requires 3 follow-up meetings to clarify what you meant, you’ve saved nothing. I’ve seen teams spend 4 hours “perfecting” AI output that would have taken 90 minutes to create properly from scratch.

Career stagnation: The executives who rely on AI for high-stakes presentations often plateau. They’re not developing the strategic thinking that separates good from great. Meanwhile, colleagues who understand frameworks and audience psychology advance faster.

I worked with a director at a major consulting firm who’d been passed over twice for partner. His presentations were technically solid but forgettable. After applying the AVP framework to his next client pitch, the feedback was: “That’s the clearest we’ve ever seen our strategy articulated.” He made partner 8 months later.

Decision paralysis: Generic AI presentations don’t drive decisions. They create more questions. “Can we schedule a follow-up to clarify…?” is the sound of an AI presentation failing.

Related: Best Copilot PowerPoint Prompts That Actually Work

How to Make AI Presentations Actually Work

AI isn’t the problem. Using AI without frameworks is the problem.

Here’s the 4-step approach that transforms AI from a liability into a genuine advantage:

Step 1: Start With Frameworks, Not Prompts

Before you touch any AI tool, answer these questions:

  • What’s the ONE action you want? (Not three actions. One.)
  • What’s the core value proposition for THIS audience?
  • What proof will they find credible?

This is the AVP framework: Action-Value-Proof. It takes 10 minutes to complete and makes your AI prompts 10x more effective.

Step 2: Use the 132 Rule for Structure

The 132 Rule: 1 message, 3 supporting points, 2 minutes maximum per section.

AI generates endless content. The 132 Rule forces focus. Before you prompt, decide your one message and three supporting points. Then prompt AI to develop ONLY those — not everything it thinks might be relevant.

Step 3: Prompt for Specificity, Not Completeness

Bad prompt: “Create a presentation about our product for potential customers.”

Better prompt: “Create 5 slides for a CFO audience. Core message: Our platform reduces month-end close from 12 days to 4. Focus on: (1) time savings, (2) error reduction, (3) ROI within 6 months. Tone: Direct, data-driven, no fluff.”

The difference? The second prompt embeds your strategic thinking into the AI request. You’re using AI as an execution tool, not a thinking tool.

Step 4: Apply the S.E.E. Formula to Proof

AI-generated proof is generic: “Companies see significant improvements…”

The S.E.E. Formula makes proof memorable: Story-Evidence-Emotion.

  • Story: “Acme Corp’s finance team was drowning in manual reconciliation…”
  • Evidence: “Within 90 days, they reduced close time from 12 days to 4.”
  • Emotion: “Their CFO told me it was the first time she left work before 7pm during month-end.”

AI can help you draft this — but only after YOU identify which story, what evidence, and what emotional hook matters for this audience.

Related: Executive Presentation Template: 12 Slides That Command the Room

The 4-step framework for AI presentations that work - AVP, 132 Rule, Specific Prompts, S.E.E. Formula

Who Gets AI Presentations Right — And Wrong

In my experience, AI presentations work for:

  • People who already know how to present — They use AI to execute faster, not to think for them
  • Internal updates with low stakes — When “good enough” is actually good enough
  • First drafts that will be heavily edited — AI as starting point, not final product

AI presentations fail for:

  • High-stakes pitches — Board meetings, investor presentations, competitive deals
  • Audiences you don’t understand well — AI can’t compensate for missing audience insight
  • People who skip the strategic thinking — Garbage in, garbage out

The professionals pulling ahead use AI as a strategic execution tool, not a content generator. They apply frameworks first, then use AI to execute 10x faster.

Stop Prompting AI Without a Framework

The Executive Prompt Pack (£19.99, instant access) gives you 71 tested prompts for every executive presentation scenario — built for PowerPoint Copilot and ChatGPT. Stop generating generic slides. Start producing board-ready decks that land decisions.

  • 71 prompts covering board updates, budget requests, investor briefs, and Q&A preparation
  • Frameworks embedded in every prompt — AI executes your strategy, not a generic template
  • Built for executives presenting at board and leadership level

Get the Executive Prompt Pack →

Instant digital download. Works with PowerPoint Copilot and ChatGPT.

Frequently Asked Questions

What are the disadvantages of AI presentations?

The main disadvantages are: generic content that doesn’t resonate with specific audiences, missing strategic structure, false confidence from polished-looking slides that don’t actually persuade, and skipping the thinking work that makes presentations effective. AI optimises for completeness and speed, not for the focus and audience awareness that drive decisions.

Why do AI-generated slides fail?

AI-generated slides fail because they produce statistically average content — the most likely output based on training data. Great presentations aren’t average. They have a point of view, focus ruthlessly on 2-3 key messages, and tailor everything to the specific audience. AI can’t do that thinking for you.

Is Copilot good for presentations?

Copilot is excellent for presentations — if you use it correctly. The tool itself is powerful. The problem is how people use it. When you apply frameworks like AVP (Action-Value-Proof) before prompting, Copilot becomes a massive time-saver. When you skip frameworks and just prompt, you get fast garbage. The tool is only as good as the thinking you bring to it.

How do I make AI presentations better?

Four steps: (1) Use the AVP framework to clarify your action, value proposition, and proof before touching AI. (2) Apply the 132 Rule — 1 message, 3 supporting points, 2 minutes per section. (3) Prompt for specificity, not completeness — tell AI exactly what to focus on. (4) Use the S.E.E. Formula (Story-Evidence-Emotion) to make proof memorable. This approach takes 25 extra minutes upfront but saves hours of follow-up and dramatically improves results.
>

© Winning Presentations Ltd. All rights reserved 2025