Ever wished you could get top-tier consulting analysis without the six-figure price tag? I know I have.
Well, I just stumbled upon an incredible post from an AI professional who shared a couple of mind-blowing prompts designed to make ChatGPT perform like a professional research consultant. I was blown away by how structured and powerful these are!
Here’s the breakdown from the original poster:
💡 For a Quick Analysis
For those fast queries, the expert shared a simple but effective template. It’s all about giving the AI a clear role and focused sources.
Here’s the prompt:
“Act like [expert, for example “McKinsey consultant”]. Search on the web for these [sources, for example a link] to answer the following [your topic].”
🚀 For a Full-Blown Research Report
This is where my jaw really dropped. For a truly deep research report, the creator shared a monster of a prompt. It’s basically a complete project brief that forces the AI to be rigorous, cite its sources, and deliver a professional-grade output.
Check this out:
“Role: Research analyst. Show a brief plan, then execute with citations.
Goal: [ONE-LINE GOAL].
Scope: Geo=[GEOGRAPHY]; Time=[START–END] history; [FUTURE YEARS]=forecasts only (with citations). Units=[CURRENCY/UNITS]. Include=[IN]; Exclude=[OUT].
Metrics: [METRIC 1]; [METRIC 2]; [METRIC 3]; [ADD MORE].
Sources (priority): [STATS OFFICE/MINISTRY] > [REGULATOR/INDUSTRY BODY] > [INTL DB: OECD/World Bank/etc.] > [TOP ANALYSTS]. Prefer originals; avoid PR/blogs.
Method: Use date filters. For each figure include value, unit, year/period, source title/org, pub date, URL. If conflicts, show both, explain, and choose. Add ≤25-word quotes w/ page.
Outputs: 10-bullet summary; assumptions/definitions; gaps/risks + next steps.
Guardrails: Only citable info. Show publication date vs data period. Label forecasts. No invented numbers.”
I think this is an absolute game-changer. The level of detail in that second prompt is what separates basic AI use from professional-grade results. It tells the AI exactly what to do, what sources to trust, and how to format the output.
This is seriously powerful stuff. The author goes into even more detail in a blog post, so be sure to check out the original LinkedIn post for the full scoop!