Scoping Research Projects
Don't ask open-ended clarifying questions and wait. Instead, elicit the minimum viable scope with a single, structured prompt that gives the user easy options to react to, then immediately produce a draft artifact (outline, tracker, or plan) they can refine. Motion beats perfect information.
Bad: "What is your research topic, and what kind of tools would help you most?"
Good:
"Let's lock down three things in one pass: (1) your topic in one sentence, (2) the type of output you're working toward (paper, decision, product, report), and (3) your rough timeframe. Even a rough guess is fine — I'll draft a starter plan from it and we'll refine. If you're not sure yet, pick the closest answer below: [A/B/C options]"
Progress:
- Step 1: Extract a working topic (even if vague)
- Step 2: Identify the end goal / deliverable
- Step 3: Identify constraints (time, scope, resources)
- Step 4: Draft a skeleton artifact immediately
- Step 5: Iterate based on user reaction, not further interrogation
Step 1 — Extract a working topic. If the user has no topic, give them 3-4 plausible directions based on context clues rather than asking "what's your topic?" blank. A menu is easier to react to than a blank page.
Step 2 — Identify the end goal. Research projects fail to organize because "research" is treated as the goal instead of a means. Ask: what decision, document, or artifact will this research produce? (e.g., literature review, grant proposal, product decision, personal understanding). This determines structure far more than topic does.
Step 3 — Identify constraints. Time horizon and depth (quick survey vs. exhaustive systematic review) change everything about structure. Get a rough number even if approximate ("a few weeks" vs "ongoing").
Step 4 — Draft a skeleton immediately. Don't wait for complete information. Produce one of:
- A project outline: research questions → sub-questions → methodology → milestones
- A source/task tracker: columns for source, claim, status, priority, next action
- A phase plan: Scoping → Literature Review → Synthesis → Output
Pick based on Step 2's answer. If goal is unclear, default to the project outline — it's the most generally useful.
Step 5 — Iterate from the artifact. Once something concrete exists, the user can react ("no, drop that section," "add a timeline") far more easily than generating structure from scratch. Treat the first draft as disposable scaffolding.
Example 1: Input: "I need to organize a complex research project" (no topic given) Output:
Let's get a skeleton in place fast. Fill in what you can:
1. Topic (one sentence, rough is fine): ___
2. End goal: [ ] Write a paper/report [ ] Make a decision [ ] Build something [ ] Just understand a topic deeply
3. Timeframe: [ ] Days [ ] Weeks [ ] Months [ ] Ongoing
I'll draft a research outline + milestone plan from whatever you give me,
even partial answers — we refine from there rather than start blank.
Example 2: Input: "Topic: effects of microplastics on soil fungi. Goal: write a lit review. Timeframe: 6 weeks." Output:
RESEARCH PROJECT PLAN: Microplastics & Soil Fungi (Lit Review, 6 wks)
Phase 1 (Wk 1): Scoping
- Define inclusion criteria (species, plastic types, timeframe of studies)
- Identify 3-5 seed papers via Google Scholar/Web of Science
Phase 2 (Wk 2-3): Source Collection
- Target 25-40 sources; track in source tracker (see attached)
- Snowball from seed paper citations
Phase 3 (Wk 4): Synthesis
- Cluster findings by mechanism (physical, chemical, microbial community shift)
- Identify consensus vs. contested findings
Phase 4 (Wk 5): Drafting
- Outline → draft sections by cluster
Phase 5 (Wk 6): Revision & gap-check
- Identify open questions for future work
- Propose, don't interrogate. Offer concrete options/menus instead of open questions — people react faster than they generate.
- Produce an artifact on the first turn whenever possible, even a rough one. A draft plan, tracker, or outline is more useful than another question.
- Default to the most generally useful structure (outline with phases/milestones) when the end goal is unclear.
- Separate "what to research" from "how to manage the research." Topic scoping and workflow/tracking are different problems — solve both but don't conflate them.
- Timebox phases explicitly. Open-ended "research" phases are where projects stall; always attach a rough deadline or volume target (e.g., "25 sources by week 2").
- Asking multiple open-ended questions in a row before producing anything tangible — causes stalling and abandonment.
- Treating "gather sources" as the whole project instead of a phase within a larger plan with a defined output.
- Waiting for complete/perfect information before drafting a structure — draft early, refine often.
- Assuming the user knows what deliverable format they want (PDF plan vs. spreadsheet tracker) — suggest a default and let them redirect.