Prompting for Research & Analysis

AI is arguably most powerful as a research and analysis tool. It can help you understand complex topics faster than any other method — but only if you ask the right questions. This lesson covers the prompting techniques that turn AI into a genuine research partner.

The Research Prompt Hierarchy

There are three types of research prompts, each serving a different need:

Level 1 — Information gathering: "What do I need to know about X?" Level 2 — Analysis: "What does X mean for my situation?" Level 3 — Synthesis: "Given all of this, what should I do?"

Most people only use Level 1. The real power is in Levels 2 and 3.

Understanding Complex Topics Fast

"I need to understand [complex topic] in depth.

My background: [what you already know]
Why I need this: [your specific use case]
Time available: I need to be conversant enough to [specific goal] by [when]

Structure your explanation as:
1. Core concept (what it actually is, in plain terms)
2. Why it matters (the real-world stakes)
3. Key mechanisms (how it works — just enough to be useful)
4. Most important subtopics I need to understand
5. Common misconceptions I should avoid
6. The 3 most important things to know if I only have 10 minutes

Assume I'm intelligent but new to this domain. Use analogies freely."

Competitive Analysis

"Conduct a competitive analysis of [company/product] vs [competitor 1] and [competitor 2].

Analyze across these dimensions:
1. Core product/service offering
2. Target customer and positioning
3. Pricing model and price points
4. Key strengths and weaknesses
5. Customer reviews and common complaints
6. Recent strategic moves or product changes
7. Where each is vulnerable

Format: Start with a summary table for quick reference, then go deep on each dimension.

Focus especially on: [specific aspect most relevant to your situation]"

Market Research

"Research the [industry/market] landscape for a company evaluating
whether to enter this market.

Cover:
1. Market size and growth rate (current and projected to 2028)
2. Key players and their market share
3. Major trends driving the market
4. Customer pain points that are currently underserved
5. Regulatory or structural barriers to entry
6. Three specific opportunities that new entrants are successfully exploiting
7. Three reasons most new entrants fail

Be specific with numbers where they exist. Note when information is uncertain or rapidly changing."

Document Analysis

"Analyze this [document type] and extract the key information:

[paste document]

Extract:
1. The main argument or purpose (1-2 sentences)
2. Key claims or findings (bullet list, each under 20 words)
3. Supporting evidence for each claim
4. Assumptions the author is making
5. What's missing or what questions this document raises
6. Any potential biases or conflicts of interest in the framing

For a [legal document / research paper / business report / contract]:
Also flag: [any clauses/claims that need special attention]"

Technical Research

"Research [technical topic] from an implementation perspective.

I need to understand:
1. What problem does this solve and when is it the right choice?
2. How does it work (at the level of understanding needed to implement it)?
3. The main implementation patterns and which to use when
4. Common pitfalls and how to avoid them
5. Performance characteristics and limits
6. Comparison to [alternatives] — when would you pick each?

Assume I'm a software engineer who will implement this.
Use code examples where they aid understanding.
Cite specific versions where behavior differs."

Synthesizing Multiple Sources

"I'm going to share information from multiple sources on [topic].
Synthesize them for me.

Source 1: [paste]
Source 2: [paste]
Source 3: [paste]

After reading all sources:
1. What are the points of agreement across sources?
2. Where do they contradict each other?
3. What's the most plausible interpretation given the conflicts?
4. What questions remain unanswered that I should research further?
5. Your overall synthesis in 3-4 sentences

Don't just summarize each source separately — find the through-line."

Fact-Checking and Verification Prompts

"Critically evaluate these claims:

[list of claims]

For each claim:
1. How confident are you in its accuracy (0-100%)?
2. What's your basis for that confidence?
3. What would need to be true for it to be wrong?
4. Any important caveats or context that changes the interpretation?

Be honest about uncertainty. I'd rather know 'I'm not sure' than have you guess."

Decision Research Framework

"Help me research the decision: [describe the decision you're facing]

I need to understand:
1. What information do I actually need to make this decision well?
2. For each piece of information: what are reliable sources?
3. What are the most important uncertainties that could change my choice?
4. What does research say about similar decisions made in similar situations?
5. What are the key questions I should ask [relevant experts]?
6. What would I regret not researching before deciding?

Start by telling me what I should research, then help me research it."

Handling AI Research Limitations

AI research has real limitations you must account for:

Knowledge cutoff: AI training data has a cutoff date. For current events, prices, or recent research, verify with current sources.

Hallucination: AI can generate plausible-sounding but incorrect facts, especially for specific statistics, citations, and less-documented topics.

Verification prompt:

"For each specific claim, statistic, or citation in your previous response:
- Rate your confidence (high/medium/low)
- Note if it should be verified with a primary source before I use it
- Flag anything you're less certain about

I will independently verify any claim rated medium or low."

The 'steelman' technique — force consideration of counterarguments:

"You've argued for [position]. Now steelman the opposing view.
What are the strongest arguments AGAINST the conclusion you just reached?
What evidence would change your assessment?
After presenting both sides, give your final view with explicit uncertainty ratings."

Building a Research Brief

For comprehensive research on any topic, use this prompt as a starting template:

"You are a senior research analyst. Create a research brief on [topic].

Purpose: [What decision or task will this brief inform?]
Audience: [Who will read this — their expertise level and what they care about]
Depth required: [overview / standard / deep dive]

Include:
- Executive summary (5 bullet points max)
- Background and context
- Current state of the topic
- Key debates or open questions
- Implications for [specific use case]
- Knowledge gaps and recommended further sources
- Bottom-line assessment

Flag all claims that are uncertain or require verification."

Next lesson: Building Your Prompt Library — how to systematize everything you've learned for maximum productivity.