🔬 Master Deep Research Mode: The AI Feature Most People Waste

The Real Power of AI Research Tools

ChatGPT Deep Research, Gemini Deep Research, and Claude Research can compress weeks of manual investigation into hours of quality analysis. But most people barely scratch the surface of what these tools can do.

I was excited when this deeper analysis was introduced; however, I sometimes found the output to be a little vanilla.

Let’s run through how to use them best.

Understanding What Each Tool Does Best

ChatGPT Deep Research

• Strength: Most thorough analysis and strongest reasoning

• Best for: Complex business decisions, competitive analysis, market entry strategies

• Quirks: Creates very long reports, can be slow, and has an expensive pro tier

• Sweet spot: When you need genuinely deep investigation with solid recommendations

Gemini Deep Research

• Strength: Shows research plan upfront, good at asking clarifying questions

• Best for: Structured analysis, academic-style research, policy investigations

• Quirks: Sometimes adds unnecessary detail, can be generic

• Sweet spot: When you want transparency in the research process

Perplexity

• Strength: Clean, well-formatted output with excellent source citations

• Best for: Quick overviews, fact-checking, staying current on trends

• Quirks: Less comprehensive than ChatGPT, more summary than analysis

• Sweet spot: When you need fast, reliable information that's easy to scan

Claude Research

• Strength: Uses many sources quickly, sound synthesis

• Best for: Rapid overviews, trend analysis, literature reviews

• Quirks: Results can feel surface-level, and have limited free access

• Sweet spot: When you need broad coverage fast

The Setup That Makes All the Difference

The quality of your research output depends heavily on the context you provide upfront.

Think of it as briefing a human (which is the key to using many AI tools).

Essential Context to Include:

Your Business Situation

• Industry and business model

• Company size and stage

• Geographic focus

• Current challenges

Your Specific Goal

• What you're trying to write/build/achieve (supply the end goal)

• Timeline for implementation

• Success metrics you care about

Your Constraints

• Budget limitations

• Technical capabilities

• Regulatory requirements

• Team bandwidth

Example of Good Context:

"I run a 50-person B2B SaaS company serving mid-market manufacturing. We're considering expanding to the EU market in Q3. I must understand VAT compliance requirements, typical implementation timelines, and costs. Our constraint is a limited legal budget—we need practical guidance on acting quickly."

Getting Better Sources and Analysis

Directing Source Quality:

• Ask for "recent data from the last 18 months"

• Request "independent analysis rather than company marketing materials"

• Specify "user reviews and employee insights over press releases"

• Ask for "government data and industry reports from established firms"

• Ask for “sources backed by large consulting companies (McKinsey, Bain, BCG, etc.)

Improving Analysis Depth:

• Request specific frameworks: "Use Porter's Five Forces" or "Apply SWOT analysis"

• Ask for contrarian viewpoints: "What are the strongest arguments against this approach?"

• Demand evidence: "Support each recommendation with specific data points"

The Research Plan Review Process

Most tools will share their research approach if you ask. This prevents wasted time investigating the wrong things.

Simple addition to any prompt: 

Before starting your research, please share your planned approach and key questions you'll investigate. I'll provide feedback to ensure we focus on the most important areas.

• Are they covering aspects of the decision you need to make?

• Are they using the proper evaluation criteria?

• Are they planning to investigate sources you trust?

• Is anything obviously missing?

Output Formatting That Actually Helps

Default outputs are often hard to use. Request specific formatting:

For Quick Decision-Making:

• "Start with a one-page executive summary"

• "Use bullet points and tables instead of long paragraphs"

• "Bold the key insights and recommendations"

For Implementation:

• "Include a prioritized action plan with timelines"

• "Provide templates or examples I can adapt"

• "List the top 3 risks and how to mitigate them"

For Validation:

• "Show me which sources support each major conclusion"

• "Highlight where expert opinions conflict"

• "Identify the biggest assumptions I should test"

The Iteration Approach

For me, I expect the first try to be a little average. The best research comes from building on initial findings and iterating.

• Phase 1: Get a broad overview with basic context

• Phase 2: Dive deeper into the most relevant areas

• Phase 3: Focus on specific implementation questions

This prevents information overload while ensuring you get actionable insights.

Maximizing Value from Long Reports

ChatGPT can generate 10,000+ word reports. Don't try to read them linearly.

Instead:

• Read the executive summary first

• Scan section headers to understand the structure

• Jump to the sections most relevant to your decision

• Use the AI to create shorter summaries for specific audiences

Pro tip: Feed the full report back into the AI and ask for "key insights for [your specific role]" or "implementation checklist based on this research."

Common Upgrades Worth Making

From Generic to Specific:

• Instead of: "Research project management tools"

• Try: "Compare Asana vs Linear for a 20-person product team, focusing on roadmap planning and developer workflow integration"

From Academic to Actionable:

• Instead of: "Analyze market trends in fintech"

• Try: "Identify the top 3 fintech trends that will impact small business lending in 2025, with specific implications for our product roadmap"

From One-Shot to Conversational:

• Start with overview questions

• Follow up with specific deep dives

• End with implementation planning

Stay up-to-date with AI

The Rundown is the most trusted AI newsletter in the world, with 1,000,000+ readers and exclusive interviews with AI leaders like Mark Zuckerberg, Demis Hassibis, Mustafa Suleyman, and more.

Their expert research team spends all day learning what’s new in AI and talking with industry experts, then distills the most important developments into one free email every morning.

Plus, complete the quiz after signing up and they’ll recommend the best AI tools, guides, and courses – tailored to your needs.

Sign up to start learning.

Getting Started

Pick one business question you've been researching manually. Use the context framework above to brief your chosen AI tool properly.

Compare the results to your usual research process in terms of:

• Time invested vs insights gained

• Quality of sources and analysis

• Actionability of recommendations

The improvement should be substantial.

Most people use these tools as sophisticated search engines, but they're actually more like having a research analyst available 24/7.

They’re good but they can be crappy unless managed well.

Until next week,

Adam