01 / START HERE
A useful way to approach it.
A useful research brief starts with a decision, not a request to find everything about a topic. Specify what someone will do differently when the research is finished. That makes it possible to choose sources, stop collecting information and explain the result to the person making the decision.
Use this template before asking an AI assistant to research a business question. Replace every bracketed field, especially the audience, country and time period. Keep the final brief to one or two pages and separate what you already know from what you still need to establish. The download is editable Markdown; it works in a text editor or document tool.
02 / THE PROCESS
Go from question to next step.
- 01
Write the decision sentence
Complete: We need to decide whether to [action] for [audience] by [date]. Add the alternatives, including doing nothing. Research that cannot change the decision belongs outside this brief.
- 02
Choose a bounded scope
Specify the geography, customer segment, period and product category. List exclusions such as primary interviews, private company data or legal advice. A boundary makes the work easier to review and prevents unrelated findings from taking over.
- 03
Define the evidence plan
For each question, name the preferred source type and the claim it should support. Official documentation can establish a published feature; a small review sample can surface concerns but cannot establish how common they are.
- 04
Set an output contract
Ask for a decision summary, a dated evidence table, conflicting findings, limitations and next steps. Each material claim needs its supporting source or an explicit unknown label. State whether the assistant has web access or must use only supplied material.
- 05
Review before starting
Have the decision owner check the question and success criteria. Remove personal or confidential information you do not need to share. If sources are missing, revise the brief instead of asking the assistant to fill the gap with a plausible answer.
03 / YOUR STARTING POINT
A prompt you can make your own.
Replace the bracketed fields. Add only information you have permission to share. This is a reusable text prompt; it does not run a task from this page.
Help me prepare a business research brief. Decision: [action and alternatives]. Audience and country: [segment, United States]. Deadline and period: [dates]. Questions: [three answerable questions]. Sources available: [public URLs or pasted source material]. Out of scope: [exclusions]. Deliver a one-page brief with a decision statement, scope, source plan, evidence table fields, stopping rule, unknowns and next validation steps. Use only accessible sources or supplied material. Do not invent citations, market figures or customer findings. Mark assumptions separately from supported facts. Ask for missing inputs before drawing a conclusion.
04 / SEE THE SHAPE OF IT
Example brief: deciding whether to offer a paid onboarding session
Illustrative planning example for a fictional business. These are research questions, not measured customer results.
| Brief field | Example entry | What to check |
|---|---|---|
| Decision | Test a paid onboarding session or keep self-service | Who owns the decision and trial budget? |
| Audience | New U.S. customers of a small scheduling service | Which accounts are eligible for the test? |
| Evidence | Published competitor offers plus permissioned support themes | Record source dates and sample limits |
| Stopping rule | Describe both alternatives and list remaining unknowns | Do not treat desk research as proof of demand |
The completed brief identifies what research can settle and what still requires a real customer test. A recommendation is provisional until the decision owner reviews that distinction.
05 / KEEP YOUR JUDGMENT
What to watch before you use it.
A topic without a decision
Research the market is too broad. State the action, alternatives and deadline so each finding has a purpose.
An unspecified source standard
A confident summary can mix advertising, opinion and measurements. Say which source types count for each claim.
A false stopping rule
Finding enough positive evidence is not a neutral finish line. Include contrary evidence and unanswered questions.
06 / THE FINAL PASS
A little review goes a long way.
Your checks stay on this page. They do not approve or run a task.
07 / GOOD QUESTIONS
Before you begin.
Does this template perform the research?
No. It defines the assignment and output. Research still requires accessible evidence and review; the template does not retrieve a proprietary dataset or interview customers.
Can I use it without web search?
Yes. Provide the source material and tell the assistant to use only that material. Ask it to list missing evidence instead of producing unsupported citations.
What should I remove before sharing?
Remove unnecessary personal data, private credentials and confidential records. Share only material you have permission to use in the selected service.