소개
커리어 탐색을 위해 딥리서치를 활용해보았습니다.
진행 방법
지피티와의 일반 대화를 통해 향후 하고 싶은 일이 무엇인지 설명하고 저와 비슷한 생각을 가지고 있는 사람이나 회사, 서비스가 있는지 알아봄으로써 목표를 구체화하고자 하였습니다.
[DeepResearchPlan v1 — Screening Map for Special Populations (KOR 60 / Global 40; 2019–2025)]
Title
Screening Map of Practical Solutions for Aging & Neurodiverse Populations (Low–Mid–High Tech; Korea 60% / Global 40%)
Background (for context)
User: biomechanics/movement/cognition MS student; human-centered, field-based; wants small-scale, high-impact solutions and hybrid career (problem solver + maker + solution designer).
Use-case environments: home, community, school, workplace (daily living), not hospital-centric.
Research Objectives
1. Build a screening-level survey of real, market-facing solutions across Low/Mid/High tech for: older adults, MCI, ASD, ADHD.
2. Identify 30–50 reference entities (KOR+Global), with URLs and business model snapshots; mark a “Top 10 closest match” set.
3. Map low-cost/high-impact opportunity zones suited for a student-to-early-career creator.
4. Extract career path patterns, skills to build, and concrete next-step projects.
Scope & Constraints
Timeframe: last 5 years (2019–2025).
Geography mix: Korea ~60%, Global ~40%.
Tech ratio target: Low 40 / Mid 40 / High 20.
Evidence threshold: commercialized or actively commercializing; pilots with paying customers or credible B2B/B2G traction preferred.
Include both physical tools and digital (apps, platforms, DTx, wearables, AI); prioritize field-tested, user-involved design cases.
Language: no restriction; include KR/EN sources.
Sensitive issues (regulatory/ethics/accessibility) only as brief flags where relevant.
Methodology (how to search & judge)
Sources: company sites, app stores, accelerator/healthtech directories, assistive tech catalogs, university labs/tech transfer pages, rehab conferences (programs/papers), maker/assistive communities.
Verification: confirm active website/social update within 12–24 months; note regulatory status (MFDS/FDA/CE) if claimed; avoid defunct or purely conceptual.
De-duplication: cluster by company/product; keep the most representative entry.
Scoring (0–5 each; compute weighted score):
• Fit-to-user-philosophy (human-centered, field-based) – 25%
• Impact on target population/problem clarity – 25%
• Evidence & traction (pilots/sales/approvals) – 20%
• Feasibility for student entry (cost, buildability, openness) – 20%
• Korea relevance (for KOR quota) – 10%
Output style: table-first, bullets minimal, 1–2 sentence descriptions, working URLs only.
Deliverables (tables first, brief prose only where needed)
1. Low-tech Solutions (40%)
Columns:
| Category(센서리/균형/인지-운동/일상ADL/치매친화) | Target(User) | Problem | Solution/Product | Region | Price band | Evidence(usage/traction) | Biz model(B2C/B2B/B2G/Rx/OTC) | Design philosophy & co-design notes | Strengths | Limits | URL |
2. Mid-tech Solutions (40%)
Columns:
| Category(스마트폰/간단센서/저가웨어러블/커뮤니티헬스) | Target | Problem | Product/App | Core tech | Regulatory/claims | Biz model + GTM | Data/privacy notes | Strengths | Risks | URL |
3. High-tech Solutions (20%)
Columns:
| Category(AI gait/balance, exosuit/robotics, advanced rehab, DTx, neurotech) | Target | Clinical/Use setting | Product | Evidence(regulatory/clinical) | Price/CapEx/Opex | Biz model | Student entry points (open SDK/clinical collaborators/similar DIY) | URL |
4. Reference Companies & Labs (30–50 entities; KOR 60% target)
Columns:
| Name | Country | Tech level(L/M/H) | Population | What they do (1–2 lines) | Why it matches user’s philosophy | Traction signal | URL | Top-10? (★) |
5. Creator / Maker Examples
Columns:
| Individual/Team/Channel | Focus (assistive/maker/therapist-inventor) | Representative builds | Audience served | Openness(OSS/files) | Contact/URL | Why it’s a model for user |
6. Market Landscape & Opportunity Map
Produce a 2x2 matrix image or ASCII: (Cost/Complexity) × (Impact/Unmet need).
Bullet list: saturated vs. underserved niches; aging-in-place; community mobility; low-friction screening/monitoring; adherence/behavioral layer; caregiver tools.
Columns (summary table):
| Problem area | Current solutions snapshot | Gaps/Barriers | Low-cost high-impact ideas (student-buildable) | KR-specific angle | Notes/URL |
7. Career Path Insights for the User
Output 3–5 archetypes with “why fit” and “how to start in 6–12 months.”
Columns:
| Archetype | Core activities | Toolchain/skills | Typical employers/partners | Starter portfolio project | Long-term trajectory(5–10y) |
8. Actionable Next Steps
Columns:
| Skill to build | How to learn fast | 1–2 weekend prototypes | KR labs/companies to contact | Global references | Success metric(what to validate) |
Inclusion/Exclusion Rules
Include only entities with shipping product, pilots, or active deployments; exclude pure concepts/indefinite “coming soon”.
Favor user-tested/field-evaluated tools; mark co-design with special populations where found.
Ensure at least 60% of entries are Korea-based or KR-active; still include global exemplars for contrast.
Partition by User’s Populations & Settings
Tag each entry by {Older adult, MCI/dementia, ASD, ADHD} and {Home, Community, School, Workplace}. Prefer entries usable outside hospitals.
Final Packaging
One master spreadsheet (tabs for sections 1–8) + a 6–8 page overview with: Top-10 matches, opportunity 2x2, and career archetypes.
Use consistent tags: [pop] [setting] [tech] [bizmodel] [evidence] to enable filtering.
Quality/Style Notes
KR/EN mixed output ok; keep KR first for Korea entities. Use concise phrases, avoid marketing fluff.
Every row MUST include a working URL; if multiple, choose the clearest product page.결과와 배운 점
생각이 구체화되지 않은 상태에서는 gpts를 활용한다고 해도 한계가 있다고 느꼈습니다. 특히 한번에 많은 양을 조사하고자 할 때 정보 탐색이 제대로 되지 않는 경우가 있었고, 적당한 분량을 나누어 질문하는 것이 중요한 것 같습니다.
도움 받은 글
딥리서치 프롬프트 작성 도우미
https://chatgpt.com/g/g-67e2161dd7b88191bfc4e6b56b66ce90-dibriseoci-peurompeuteu-jagseong-doumi