AI Research & Analysis Revision
AI Research & Analysis Revision
Your Personal Intelligence Analyst
Framework: 3 Deep-Dive Techniques (Actionable Summary, ELI5, Cross-Examination)
The core idea
Generic "summarise this" prompts waste AI's real research power. Day 3 teaches three targeted techniques instead: the Actionable Summary (bullet points a specific role actually needs to act on), ELI5 (explaining dense or technical sections in plain language), and Cross-Examination (asking the AI to find weaknesses, hidden assumptions, and unanswered questions in a document — useful before trusting a proposal or contract).
The other half of Day 3 is trust and verification. AI can hallucinate facts and misattribute sources with total confidence, so the Three-Layer Verification Protocol exists to catch that before it costs you — independently locating and checking any statistic, quote, or clause the AI surfaces, rather than taking it at face value.
A toolkit matters here too: ChatPDF for questioning a single uploaded document, Perplexity for live web research with citations, Gemini's YouTube Shortcut for timestamped video summaries, and NotebookLM for querying across many documents you've accumulated over time.
Key terms
| Term | Definition |
|---|---|
| Actionable Summary | A summary technique targeted at what a specific role needs to do, not a generic overview |
| ELI5 | Explain Like I'm 5 -- simplifying dense or technical content into plain language |
| Cross-Examination | Asking AI to surface weaknesses, assumptions, and gaps in a document rather than just describing it |
| Three-Layer Verification Protocol | A structured approach to independently verifying AI-surfaced facts before relying on them |
Common mistakes to avoid
- Trusting a specific AI-cited statistic or source without opening and checking the original
- Using the same generic summary technique regardless of who the summary is actually for
- Skipping Cross-Examination on high-stakes documents like contracts or investment proposals
- Forgetting that even a confidently-worded, specific-sounding AI citation can still be wrong
Before the exam, make sure you can...
- Name and briefly explain all 3 Deep-Dive Research Techniques
- Know which tool fits which research task (ChatPDF vs Perplexity vs NotebookLM vs Gemini)
- Understand why independent verification matters even when AI cites a specific source
- Recall the target format for a professional Executive Summary (200 words, standard template)