Published September 18, 2026
Published September 18, 2026.
Research delegated with a vague prompt comes back vague. The assistant guesses what you wanted, you reject the result, and both of you repeat the cycle. A written brief prevents most of that waste and makes research one of the easiest tasks to hand off well.
research brief template for a virtual assistant
A research brief template is a short document that states the decision the research will support, the specific questions to answer, the sources you accept, the output format, and the deadline. It replaces open-ended instructions like "look into this" with a bounded assignment the assistant can complete and you can evaluate.
The template does not need to be long. Half a page is usually enough. What matters is that each section is specific enough that two different assistants would produce comparable work. That is the test of a good brief.
Build a reusable template and fill it in each time. The first brief takes a few minutes longer than a casual request. Every brief after that is faster, and the quality of the returned work rises immediately.
Start with the decision, not the topic
The most useful first line states what decision the research will inform. "Decide whether to renew this vendor" is far more useful than "research vendors." The decision tells the assistant which facts matter and which can be skipped.
State who will read the output and what they already know. An assistant briefing you does not need to explain your own business. An assistant briefing a client needs more context and a different tone. This single line changes the whole assignment.
Add the stakes. Research for a low-cost tool choice can be quick and directional. Research supporting a contract or a public claim needs verification and citations for every material fact. The stakes tell the assistant how much rigor to apply.
Write the questions precisely
List the exact questions you want answered, in priority order. Keep each question narrow enough to answer with evidence. "Is this vendor reliable" is broad; "what do current customers report about support response and billing accuracy" is answerable.
Cap the list. Three to five questions is usually enough for a single brief. A longer list either hides the real priority or signals that the task should be split into several assignments.
For each question, note what a good answer looks like. It might be a range, a yes or no with reasoning, or a comparison across options. When you define the shape of the answer, the assistant can tell when the work is done.
Name the sources you accept
Specify which sources count as evidence. For market questions, first-party reports and industry publications may be appropriate. For technical or legal questions, primary documents usually are. For anything changeable, such as prices, policies, or statistics, require a dated source and a link.
State what does not count. Anonymous forum posts, undated blog summaries, and AI-generated text without a source are common exclusions. Naming these upfront prevents the assistant from returning a confident answer built on weak evidence.
For source evaluation, a short guide helps. The Purdue OWL guide to evaluating sources covers the questions that separate a reliable source from a weak one.
Define the output format
Tell the assistant exactly what to deliver. A one-page memo, a comparison table, a slide, or a set of bullet points each require different work. Include the fields you expect. For a comparison, list the columns. For a memo, list the sections.
Set a length limit. A page limit forces prioritization and makes the output easier to review. Without one, research tends to expand to fill the available time.
Ask for a confidence note. Have the assistant mark which findings are well supported and which are uncertain. This turns an overconfident summary into something you can actually use.
Include a time box and a finish line
State the deadline and the effort limit. If the assistant is paid by the hour, a time box prevents an unbounded search. If the work is priced per task, the time box clarifies the expected scope.
Define what "enough" looks like. Research is never truly complete, so the brief should say when to stop: when the questions have adequate support, the key uncertainty is named, and further searching is unlikely to change the recommendation.
If the questions cannot be answered within the box, ask the assistant to report the gap rather than keep searching. A clear statement of missing evidence is more valuable than a padded report.
Review against the brief
When the research returns, check it against the brief rather than reacting to whether you liked it. Did it answer the listed questions? Did it use accepted sources? Did it follow the format and length? Did it flag uncertainty honestly?
| Brief section | Review question |
|---|---|
| Decision | Does the output support the named decision? |
| Questions | Was each priority question answered? |
| Sources | Are the sources acceptable and dated? |
| Format | Does the output match the requested structure? |
| Confidence | Are uncertain findings labeled clearly? |
| Finish line | Did the work stop at the agreed point? |
If a section failed, fix the brief, not just the output. Repeated failures in the same section mean the template needs a clearer instruction.
Build a library of briefs
Save each brief and the finished work together. Over time you build a set of examples that shows new assistants exactly what good output looks like. This shortens every future ramp and reduces corrections.
Tag the briefs by type, such as vendor comparisons, market scans, or background checks. When a similar need arises, start from the closest brief and adjust the questions. The template does the heavy lifting.
Review the library quarterly and retire briefs that no longer match how you work. A stale template produces stale research.
Frequently asked questions
How long should a research brief be?
Half a page to one page is usually enough. The goal is specificity, not length. If every section is clear, the brief is complete.
What if I do not know the exact questions yet?
Write one broad question and ask the assistant to propose a narrowed list. Review that list before the full search begins. This keeps the later work focused.
How do I handle a research task that keeps expanding?
Set a time box and a stopping rule in the brief. If the questions cannot be answered in the box, ask for a report on the gap and a recommendation for the next step.
Brief first, then delegate
A good brief is the difference between research you can use and research you have to redo. Pair this template with virtual assistant research tasks for task ideas, and use how to delegate effectively to set the boundaries that keep the work moving without you.