sense.monster

Collect the dots. Connect the dots.

AI prompt generator for independent fact research without narrative.

Pick an example to watch the tool run end to end in one click - handy if you just want a fast test.

What do you want to investigate? Each choice builds prompts designed for that kind of subject - an event needs different questions than a media outlet or a corporation.

Pick the domain that best fits your subject. Each domain has prompts calibrated for its specific patterns and actors.

Enter the specific subject. Be specific - "Vioxx withdrawal 2004" beats just "Vioxx". Or pick a ready example below.

0 / 4000

Examples:

Choose the angle the AI should take on your subject. For best results, run every angle in sequence inside the same chat.

Include web search instruction

AI models have a knowledge cutoff date - they don't know about events from the last months. If your topic has had significant developments in the past 24 months (new revelations, court rulings, declassified documents, policy changes), keep this on. It adds an instruction asking the AI to search the internet for the latest updates before responding. For older, well-documented topics you can turn it off.

Paste this prompt into any AI (ChatGPT, Gemini, Claude, etc.)

How to use these prompts

Choose your AI chat. Free AI chats work well for these prompts. Gemini performs slightly better than ChatGPT in its free version. For deeper and more accurate results, paid versions and more advanced models (GPT-4, Gemini Advanced, Claude Pro) produce noticeably better quality.

Ask follow-up questions. These prompts are starting points. After each response, ask the AI to expand on specific entries, provide more detail on a person or event, or cover areas it missed.

Read the prompt before sending it. Feel free to adjust the wording, add or remove categories, or change the focus. These are templates, not commandments.

Use more than one AI service. Every AI has its own biases and blind spots. Run the same prompt through ChatGPT, Gemini, Claude, and others. Compare the results. What one misses, another may catch.

Older events produce better results. These prompts work best on topics from 3+ years ago, where documents have been declassified and investigations concluded. Recent events are harder to assess. Systemic topics like "media ownership consolidation" work just as well.

Always verify the results independently. AI can hallucinate facts, misattribute sources, or get dates wrong. Treat every response as a research lead, not as a finished product. If it says "court filing" - find that document.

How NOT to use these prompts

Coincidences can still be coincidences. Some prompts deliberately surface suspicious timing and financial patterns. Resist the urge to connect every dot - that's pattern-matching bias.

If you're collecting "I told you so" moments, you're using this wrong. This tool is not ammunition. If the results consistently make you feel validated - that is a sign you have fallen into the same trap from the other side. Genuine inquiry is uncomfortable. That is confirmation bias.

No AI response is a definitive list of facts. AI models miss key elements - because of safety filters, training data limitations, or simply running out of space. This is a starting point for research, never the final word.

This tool does not draw conclusions for you. The prompt generator is a shovel, not a map - it helps you dig further into documented facts, nothing more. Before you interpret, read the Important concepts and the articles in the Topics section.