How to Ask for a Research Brief (and Actually Get Something Useful Back)

If you've ever asked someone — a colleague, an assistant, or an AI tool — to "do some research on X" and gotten back either a wall of unstructured text or a shallow list of bullet points, the problem usually isn't the researcher's ability. It's the request.

A good research brief is really a contract between the person asking and the person (or AI) writing. The clearer the contract, the better the output. Here's what actually matters when you're framing a research request, whether you're doing it yourself or outsourcing it.

1. State the decision, not just the topic

"Tell me about electric vehicle batteries" produces something generic. "I'm deciding whether to recommend lithium iron phosphate or nickel-based batteries for a mid-range delivery van fleet, and I need the tradeoffs in cost, lifespan, and cold-weather performance" produces something you can actually act on. Whenever possible, name the decision the research is supposed to support. That single sentence reshapes everything downstream: what gets included, what gets cut, and what "useful" even means.

2. Specify the audience

A brief written for a technical co-founder reads differently than one written for a board member or a curious teenager. If you don't specify, the writer has to guess — and guesses default to a generic, middle-of-the-road tone that serves nobody well. Say who's going to read it and what they already know.

3. Give a length and depth expectation

"Brief" means different things to different people. Some want 300 words they can skim in two minutes; others want 1,500 words with real depth. Naming a rough word count or "depth level" (overview vs. deep-dive) saves a round of disappointment.

4. Flag what kind of knowledge you need

This matters more than people think. If you need live statistics, breaking news, or figures that change month to month, say so explicitly — and make sure whoever (or whatever) is writing the brief actually has access to current data. A brief built from general/established knowledge is genuinely useful for explaining how something works, summarizing established debates, or structuring a topic you're new to. It is the wrong tool for "what's the current price of X" or "what happened in the news yesterday." Knowing which kind of question you're asking saves everyone time.

5. Name the sub-questions you actually care about

Instead of one broad topic, list the 3-5 things you're actually trying to resolve. "Explain how carbon offset markets work" is vague. "Explain how carbon offset markets work, who verifies the offsets, and what the main criticism of the model is" gives the writer a skeleton to fill in — and gives you a brief that hits the things you'll actually be asked about later.

The pattern

Decision → audience → depth → knowledge type → sub-questions. Five short lines of context, and the brief that comes back is dramatically more useful than one built from a bare topic alone. This applies whether you're briefing a research assistant, delegating to a colleague, or ordering a written brief from a service.

---

Lucid Desk writes custom research briefs (800-1200 words, from general/established knowledge, not live data) for $18, delivered within minutes of payment. If you give us the five things above in the order form, you'll get a brief worth reading. See the offer.