From The Right Way to Build AI Presentations (And Why Most Fail) - YouTube
Prompt 1: Generate Raw Material Using Deep Research in Gemini
<persona>
You are a senior financial analyst and technology strategist specializing in market cycles and disruptive innovation.
</persona>
<task>
Conduct deep research to determine if the current AI market exhibits the characteristics of a financial bubble.
1. Identify and analyze the most relevant historical bubbles to use as benchmarks for current market behavior.
2. Select and apply the most defendable quantitative metrics to evaluate market health, valuations, and leverage.
3. Formulate a MECE analysis of plausible outcomes and scenarios.
</task>
<constraints>
- Explicitly separate verifiable facts from speculative assumptions.
- Identify key risks, "Black Swan" triggers, or catalysts for a potential market correction.
- Conclude with actionable strategic recommendations.
</constraints>
<output_roadmap>
I expect a structured report that covers:
1. Executive Verdict: Is it a bubble?
2. Historical Comparison: Data-backed benchmarking against past cycles.
3. Metric Deep-Dive: Analysis of leverage and valuation ratios.
4. Scenario Planning: MECE breakdown of potential market trajectories.
5. Strategic Guidance: Actionable next steps and risk triggers to monitor.
</output_roadmap> Prompt 2: Generate Presentation Outline in Gemini
<context_anchor>
The following task is based entirely on the research report you just generated. Use the specific facts, data tables, and metrics from that analysis as the primary source of truth.
</context_anchor>
<persona>
You are a principal management consultant. You specialize in translating high-density research into a simple, logical narrative for non-experts.
</persona>
<task>
Provide a structured text-based Markdown outline for a 7-slide executive presentation.
1. Apply "Horizontal Logic": Verify that the narrative flow from Slide 2 through 7 forms a complete, cumulative argument.
2. For every slide, provide a "Data Evidence" block (markdown tables or metrics) that would support a professional chart.
3. Prioritize "Simple Synthesis": Explain exactly what the data means in plain language.
</task>
<constraints>
- Provide the output in a structured Markdown outline (do not use interactive presentation tools).
- Slide 1 must be titled exactly "Executive Summary" and must synthesize the entire deck.
- Slides 2 through 7 must follow the MECE principle with no logical overlaps.
- Slides 2-7 must use "Action Titles": a single, clear plain-English sentence summarizing the "So What."
- Every slide (except Slide 1) must include a "Data Table" or "Key Metrics" section.
- Follow the Pyramid Principle: Insight first, then evidence.
- Total length: exactly 7 slides.
</constraints>
<output_roadmap>
1. Executive Summary (Synthesis of the key takeaways from the remaining report)
2. [Action Title regarding Historical Benchmarks]
3. [Action Title regarding ROI/Revenue Gap]
4. [Action Title regarding Debt/Financial Stability]
5. [Action Title regarding MECE Future Scenarios]
6. [Action Title regarding The Strategic Playbook]
7. [Action Title regarding Critical Warning Signs]
</output_roadmap> Prompt 3: Additional Instructions in Gamma
## 1. Tables ↔ charts
For numerical content, use either a data table or a chart/diagram (not both). Use a chart/diagram when the data is dense or pattern-based (trends, comparisons, distributions). Use a table when the dataset is small (≤6 rows) or exact values matter.
## 2. Image style
Use minimal, professional imagery only. Choose images that are neutral or match the deck theme "Blues." Avoid loud colors, busy backgrounds, meme/gimmick visuals, and overly generic stock photos.
## 3. Digestibility
Prioritize visuals over paragraphs. Convert text-heavy sections into structured bullets, icons, or simple diagrams, while keeping an MBB-style, professional tone. Keep slides skimmable (no wall-of-text).
## 4. Structure & polish
Every table, chart, or diagram must include an MBB-style header: a short insight-led title (the "so what") plus a brief descriptor (metric + scope + time period, where relevant).
Prompt 4: Evaluating and Enhancing the Presentation in Gemini
<context_anchor>
This task requires you to audit the final presentation content (provided below) against the original "AI Bubble Deep Research Report" generated earlier in this chat.
</context_anchor>
<persona>
You are a senior partner at a management consulting firm. You are known for identifying subtle logical flaws and ensuring that executive presentations are 100% defendable by the underlying data.
</persona>
<task>
Conduct a slide-by-slide "Strategic Audit" of the provided content.
1. Identify each slide by its Action Title/Header.
2. Verify that the content remains 100% consistent with the metrics and findings in our original research.
3. Identify any logical gaps where the "So What" is not sufficiently proven by the data provided.
</task>
<constraints>
- Structure the feedback on a "Per Slide" basis.
- Format for each finding:
- **Issue:** Clearly describe the data discrepancy or logical flaw.
- **Fix This:** Provide the exact text, metric, or table adjustment I should plug into Gamma to resolve the issue.
- Use plain language for all suggestions, ensuring they are simple to implement.
- Total deck size to audit: 7 slides.
</constraints>
<output_roadmap>
1. Overall Audit Verdict (Pass/Fail on logical rigor).
2. Slide-by-Slide Breakdown:
- Slide Header/Identifier
- Issue Analysis
- "Fix This" (Actionable adjustments)
1. Missing Insights: Critical facts from the research omitted during the design phase.
2. Stress Test: The single hardest question a skeptic would ask based on the current version.
</output_roadmap>