Create an executable research project plan based on the inputs below. Do not present assumptions, summaries, or AI-generated examples as findings. If a decision requires regulated, technical, legal, medical, or financial expertise, identify the appropriate qualified reviewer. PROJECT INPUTS [DECISION TO INFORM] [RESEARCH QUESTION AND SCOPE] [STAKEHOLDERS AND RESOURCES] [TIMELINE AND CONSTRAINTS] [EXISTING EVIDENCE] Begin by restating the decision in one sentence and defining what evidence would change that decision. Break the broad question into answerable subquestions. Set boundaries for geography, audience, period, and exclusions so the project does not grow without control.
Produce a project plan with: 1. Decision brief: decision owner, decision date, options being considered, and success criteria. 2. Research questions table: primary question, subquestion, why it matters, evidence needed, and acceptance threshold. 3. Method plan: recommended mix of desk research, data analysis, interviews, survey, observation, or experiment. For each method, state sample or source criteria, collection steps, bias risks, consent or privacy considerations, and expected limitation. 4. Work breakdown: numbered tasks in sequence, each with owner, input, output, dependency, estimate, and completion definition. 5. Timeline: week-by-week milestones from kickoff through synthesis, review, and decision meeting. 6. Evidence log template: source, date, method, claim supported, quality assessment, conflicts, and citation or file reference. 7. Deliverables: research brief, findings deck outline, appendix, and decision memo; specify the audience and length of each. 8. Risk register covering access, response bias, data quality, scope creep, and deadline risk, with a mitigation and owner for each. Distinguish facts already established from claims to investigate. Do not assign owners who are not listed; mark them “unassigned.” Before answering, check that every research question connects to the decision and every milestone has an accountable owner or is marked unassigned. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [DECISION TO INFORM] | State the specific decision this research must help a person or team make. | Decide whether to launch a self-serve onboarding path for the company’s analytics product in Q1. |
| [RESEARCH QUESTION AND SCOPE] | Provide the main research question, target population, geography, time period, and exclusions. | Why do US trial users at companies with 10-100 employees fail to activate in the first 14 days? Exclude enterprise and existing annual customers. |
| [STAKEHOLDERS AND RESOURCES] | List decision owner, contributors, available data, budget, tools, and access to participants or sources. | Decision owner: VP Product; owners: Priya Research, Ben Data, Omar PM; Amplitude events, support tickets, and $4,000 incentive budget available. |
| [TIMELINE AND CONSTRAINTS] | Give the final decision date, interim deadlines, and practical, privacy, or compliance constraints. | Decision meeting January 16; two weeks for fieldwork; do not contact users who opted out of research. |
| [EXISTING EVIDENCE] | List current reports, datasets, prior findings, links, and known gaps. | Activation dashboard shows 31% activation; September support analysis suggests CSV import confusion; no interviews with churned trials. |
How to use
- State the decision before stating the topic, so the plan can exclude interesting but non-decisive questions.
- Name actual owners and available data sources; leave unknown ownership unassigned rather than guessing.
- Check the acceptance thresholds and completion definitions with the decision owner at kickoff.
- Follow up with: “Convert the work breakdown into a CSV-ready task table with dates starting [START DATE] and preserve all dependencies.”
Variations
Interview research plan
Use this when interviews are the main source of evidence.
Design a qualitative interview study to inform [DECISION]. Research questions: [QUESTIONS]. Target participants: [PARTICIPANT CRITERIA]. Constraints and available access: [CONSTRAINTS]. Produce a recruitment screener, sample plan, 45-minute discussion guide, consent and privacy notes, interviewer roles, coding framework, and synthesis schedule. Distinguish exploratory questions from questions that need quantification. Do not assume participant opinions represent the full market. Check that every interview question maps to a decision-relevant research question.
Desk research plan
Use this when the project depends mainly on public or internal documents.
Create a desk-research plan for [DECISION] about [TOPIC]. Existing sources: [AVAILABLE MATERIALS]. Required scope: [SCOPE AND DEADLINE]. Produce prioritized search themes, source hierarchy, inclusion and exclusion rules, a source-evaluation rubric, claim-tracking table, synthesis outline, and task timeline with owners from [TEAM]. Prefer primary documents, original datasets, and dated sources. Do not treat search snippets as evidence. Flag claims that require expert review or cannot be validated from available sources.
Research findings outline
Use this after fieldwork and analysis are complete.
Turn these research materials into a decision-focused findings outline: [EVIDENCE AND NOTES]. The decision is [DECISION], and the audience is [AUDIENCE]. Produce a 10-slide deck outline, a one-page decision memo, and an evidence appendix structure. Separate observations, interpretations, and recommendations. Quantify only where the material supports it, include conflicting evidence, and state limitations beside key findings. Check that each recommendation traces to evidence and that no participant quote is fabricated or presented without attribution.
Tips
- Write acceptance thresholds before fieldwork, such as the number of validated interviews or data-quality checks needed to act, so conclusions are not chosen after results appear.
- Keep a claim-level evidence log rather than only a bibliography; it makes conflicting sources and unsupported statements visible during synthesis.
- Interview guides should test behavior, context, and tradeoffs before asking for preferences, which are often less predictive of action.
- Schedule synthesis before the final readout; collecting evidence until the last day leaves no time to reconcile conflicts or define implications.
FAQ
Can AI conduct the research project for me?
It can plan the work, help analyze supplied material, and draft structured outputs. It should not be treated as a source of verified findings without checking primary evidence.
How many interviews are enough?
It depends on population diversity and the decision risk. Define a sample rationale and stop condition in advance, then note what the study cannot generalize.
What makes a research project decision-ready?
It names the decision owner, the alternatives, evidence thresholds, limitations, and the date when the findings will be used.