13 GPT-5 Prompts for Consulting Slide Decks

Why Most AI-Generated Slides Still Look Like a College Project

GPT-5 can write a 40-page strategy deck in seconds. The problem? Without the right prompts, it produces generic bullet-point walls that no partner would approve and no client would sit through.

The difference between a forgettable deck and one that wins the engagement comes down to prompt engineering. We spent 6 weeks testing 50+ prompt variations on real consulting deliverables — strategy reviews, market sizing, due diligence readouts, and board presentations — then distilled the 13 prompts that consistently produced partner-ready output.

---

How We Tested These Prompts

Methodology:

Baseline (manual process):

---

Prompt #1 — The MECE Framework Generator

This prompt forces GPT-5 to think in mutually exclusive, collectively exhaustive buckets before writing a single slide.

The prompt:

```

Act as a McKinsey engagement manager. I need a MECE framework for [TOPIC].

Break it into 3-5 mutually exclusive categories that collectively cover the

entire problem space. For each category, provide:

1) A slide title (max 8 words, action-oriented)

2) The governing thought (one sentence that states the "so what")

3) 3-4 supporting data points or arguments

Format as a slide outline, not paragraphs.

```

**Why it works:** The "engagement manager" role plus explicit MECE instruction eliminates the rambling, overlapping frameworks GPT-5 defaults to. The "governing thought" requirement forces every slide to have a clear takeaway.

**Time saved:** Framework creation drops from 45 minutes to 3 minutes.

---

Prompt #2 — Executive Summary With the Pyramid Principle

The prompt:

```

Write an executive summary for a consulting presentation on [TOPIC].

Use the Pyramid Principle: lead with the recommendation, then group

supporting arguments into 3 pillars, each backed by 2-3 evidence points.

Total length: 150-200 words. Tone: assertive, not hedging.

Replace "we believe" with "our analysis shows."

Replace "it seems" with "the data indicates."

```

**Why it works:** The explicit Pyramid Principle instruction plus banned phrases eliminates the wishy-washy summaries GPT-5 defaults to. Partners want conclusions, not explorations.

---

Prompt #3 — Market Sizing Slide (TAM/SAM/SOM)

The prompt:

```

Create a market sizing analysis for [MARKET] using the TAM/SAM/SOM framework.

For each layer, provide: the dollar figure, the methodology (top-down or

bottom-up), key assumptions, and a one-line "sanity check" explaining why

the number is credible. Format as 3 distinct slide outlines. Include a

fourth slide showing the methodology waterfall from TAM to SOM.

```

**Why it works:** Asking for methodology and sanity checks forces GPT-5 to show its math instead of hallucinating round numbers. The waterfall slide gives partners the visual they always request.

---

Prompt #4 — The Situation-Complication-Resolution Slide

This was our highest-rated prompt. It produced partner-ready output 78% of the time on first generation.

The prompt:

```

Structure a consulting slide using the SCR framework for [TOPIC]:

Then write a slide title that captures the resolution in under 10 words.

Add 3 bullet points for the slide body, each starting with a bold verb.

```

**Why it works:** SCR is the backbone of McKinsey-style storytelling. The "urgency driver" instruction prevents generic complication statements. Bold verb bullets match real deck formatting.

**Time saved:** Full slide creation from 3 hours to 18 minutes (average across 12 test decks).

---

Keep Reading

6 AI Tools That Cut Consulting Deliverables 60%Best AI Tools for Freelancers (2026)

Prompt #5 — Competitive Landscape Matrix

The prompt:

```

Create a competitive landscape analysis for [COMPANY] in [INDUSTRY].

Format as a 2x2 matrix with [AXIS 1] on the X-axis and [AXIS 2] on the Y-axis.

Place 5-7 competitors in the appropriate quadrants with 1-line justifications.

Then write the "so what" insight: which quadrant is most attractive and why.

Include a recommended strategic move based on the white space.

```

---

Prompt #6 — Data Storytelling Narrative

The prompt:

```

I have the following data: [PASTE DATA]. Write a consulting-style data narrative

for a slide. Structure: 1) Headline that states the insight (not the topic),

2) 2-3 sentences explaining the trend, 3) One "so what" sentence connecting

to business impact. The headline should pass the "so what" test — if a partner

reads only the headline, they should understand the implication.

```

**Why it works:** The "headline states the insight, not the topic" instruction is the single biggest quality lever. It transforms "Q3 Revenue Analysis" into "Revenue grew 23% but margin compression signals pricing pressure."

---

Prompt #7 — Risk Assessment Slide

The prompt:

```

Create a risk assessment for [PROJECT/INITIATIVE]. List 5 risks in a table with

columns: Risk, Likelihood (High/Medium/Low), Impact (High/Medium/Low),

Mitigation Strategy, Owner. Sort by combined risk score (likelihood × impact)

descending. Add a summary sentence identifying the top risk and recommended

immediate action.

```

---

Prompt #8 — Implementation Roadmap

The prompt:

```

Create a phased implementation roadmap for [INITIATIVE] over [TIMEFRAME].

Structure in 3 phases: Quick Wins (0-90 days), Build (90-180 days),

Scale (180-365 days). For each phase list: 3-4 key activities, 1 measurable

milestone, required resources, and primary risk. Format as a Gantt-style

text outline that a designer can visualize.

```

---

Prompt #9 — Stakeholder Alignment Slide

The prompt:

```

Map the stakeholder landscape for [INITIATIVE]. Create a 2x2 matrix:

Influence (High/Low) vs Support (High/Low). Place 5-7 stakeholders with

their current position and recommended engagement strategy for each quadrant.

Write a one-paragraph "political read" summarizing the key dynamics and

potential blockers.

```

---

Prompt #10 — Financial Impact Waterfall

The prompt:

```

Create a financial impact analysis for [INITIATIVE] formatted as a waterfall chart

narrative. Start with current state revenue/cost, then show each lever:

[LEVER 1], [LEVER 2], [LEVER 3], ending at projected state. For each lever,

provide the dollar impact, confidence level, and key assumption.

Write the slide headline as "[X]M opportunity through [primary lever]."

```

---

Prompt #11 — The "Appendix Killer"

The prompt:

```

I need to create appendix slides for [TOPIC]. For each of the following

sub-topics: [LIST], create a one-slide summary with: title, 3 key data points,

source citation, and one "if asked" talking point. Format so each slide is

self-contained and can answer a partner's deep-dive question without

additional context.

```

**Why it works:** Appendix slides are time sinks. This prompt generates defensible backup slides in minutes that would normally take hours of research.

---

Prompt #12 — Client-Ready Recommendation Slide

The prompt:

```

Write a recommendation slide for [CLIENT] on [TOPIC]. Structure:

1) Headline: "[Action verb] [specific initiative] to [quantified outcome]"

2) 3 supporting arguments, each with a bold label and one supporting sentence

3) Next steps: 3 immediate actions with owners and deadlines

4) Investment required vs expected ROI in one line

Tone: confident, not salesy. This is a trusted advisor speaking.

```

---

Prompt #13 — The Full Deck Skeleton

The prompt:

```

Create a 12-slide consulting deck outline for [ENGAGEMENT TYPE] on [TOPIC].

Use this structure:

For each slide, write: the title, governing thought, and 3-4 content bullets.

Ensure the deck tells a coherent story from diagnostic to prescription.

```

**Why it works:** This creates a complete first draft skeleton in under 2 minutes. Fill in client-specific data and you have a reviewable deck in 30 minutes instead of 4 hours.

---

Pro Tips for Maximum Quality

**Chain your prompts:** Use Prompt #13 first to get the skeleton, then apply specific prompts (#1, #4, #6) to individual slides that need the most polish.

**Feed context first:** Before any prompt, paste 2-3 paragraphs of client background. GPT-5's output quality jumps dramatically with domain context.

**Use "Act as" consistently:** "Act as a McKinsey engagement manager" produces measurably better output than "Act as a consultant." Specificity in role-setting directly impacts structure quality.

**Set formatting constraints:** Always specify word counts, bullet counts, and format requirements. Without constraints, GPT-5 defaults to verbose paragraphs that don't map to slides.

---

FAQ

Can GPT-5 fully replace a consulting analyst for slide creation?

Not yet. GPT-5 handles 70-80% of first-draft work — frameworks, structures, data narratives — but still needs human judgment for client-specific nuance, political sensitivity, and design polish. Think of it as cutting analyst time by 60%, not eliminating the role.

Which prompt should I start with if I'm new to AI slide creation?

Start with Prompt #4 (SCR framework). It's the most versatile, works for any topic, and produced partner-ready output 78% of the time in our testing. Once comfortable, layer in Prompt #13 for full deck skeletons.

Do these prompts work with Claude or Gemini, or only GPT-5?

All 13 prompts work across major models, but GPT-5 scored highest on structure quality and MECE compliance. Claude 4 was a close second for narrative quality. Gemini 2.5 Pro handled data-heavy prompts well but struggled with consulting-specific formatting.

How do I handle confidential client data when using these prompts?

Never paste raw client data into public AI models. Use anonymized placeholders ("[CLIENT]", "[REVENUE FIGURE]") in prompts, generate the structure, then fill in real data locally. For sensitive engagements, use your firm's private AI deployment or Azure OpenAI with enterprise data protection.