Most AI research fails for one boring reason: the request was vague. Ask an AI to "research profitable niches" and you get a confident list of generalities — no sources, no sample sizes, no way to tell fact from guesswork. This prompt fixes the request itself. It turns any large language model or AI agent into a disciplined research delegate: you give it a decision you need to make, and it returns a structured brief with cited sources, a strict separation between verified facts and its own inferences, the strongest counter-evidence it can find, and a ranked shortlist tied to criteria you set in advance. It was built and refined during a real AI-agent project, where an agent used this exact structure to research content platforms and niches and came back with data the human could actually act on — including the finding that changed the whole strategy. Inside you get the complete copy-paste prompt, a short guide to filling in each field, and a filled example with sample output so you can see the expected format before you run it. It is for founders, creators, freelancers, and operators who use AI for research and are tired of answers they cannot verify, cite, or defend in a real decision.

What you get

- The complete copy-and-paste market-research delegation prompt

- Built-in rules for sources, fact/inference separation, counter-evidence, and no fabrication

- A ranked-shortlist output format scored against your constraints

- A filled example with sample output