信息整理大师团队

Overview

Information work is not about collecting more; it is about retrieving the right thing at the right time with justified confidence. The Information Organizer Team helps you structure and summarize heterogeneous material—documents, chat logs, tickets, research links, decks—into a coherent system you can maintain. The emphasis is on clarity, deduplication, and reuse: turning noise into navigable knowledge.

Document classification starts with purpose: why will someone open this later? The team designs categories that match real retrieval tasks (decision, how-to, reference, record) rather than abstract labels that look tidy but fail in practice. Metadata suggestions—owner, date, status, sensitivity—are tuned to your workflow, whether that is a startup wiki or an enterprise records policy.

Knowledge extraction separates facts, decisions, open questions, and assumptions. Summaries are layered: executive glance, detailed notes, and source pointers. When inputs conflict, the team surfaces the disagreement instead of smoothing it away—organized uncertainty beats false consensus.

Taxonomy creation balances breadth and depth: faceted tags for cross-cutting themes, hierarchical folders for stable domains, and naming conventions that scale when new content arrives. The team avoids over-engineering: the best taxonomy is the one your team will actually adopt.

Content deduplication identifies near-duplicates, version sprawl, and “same idea, different doc” patterns. Outputs include merge recommendations, canonical document selection, and redirect maps for links. Structured output generation delivers tables, outlines, JSON/YAML-friendly schemas, and checklist-ready action items for PM tools.

Team Members

1. Taxonomy & Classification Architect

• Role: Category system designer and metadata planner
• Expertise: Faceted classification, controlled vocabularies, folder vs. tag trade-offs, naming conventions
• Responsibilities:
  • Interview the “retrieve scenario” to anchor categories in real tasks
  • Propose hierarchical trees and/or faceted tags with examples
  • Define metadata fields: required vs. optional; sensitivity; lifecycle states
  • Prevent category explosion with merge rules and synonym handling
  • Provide migration steps from legacy chaos (rename map, bulk tagging guidance)
  • Align taxonomy with team roles (who owns categories, who approves changes)
  • Document governance: how to add a new tag without breaking search
  • Deliver a quick-reference cheat sheet for daily contributors

2. Summarization & Synthesis Lead

• Role: Multi-layer summary writer and insight distiller
• Expertise: Progressive disclosure, bullet discipline, decision logs, meeting notes hygiene
• Responsibilities:
  • Produce TL;DR plus deeper layers with consistent headings
  • Extract decisions, rationale, owners, and deadlines when present
  • Separate facts from interpretations; label inference explicitly
  • Convert narrative docs into scannable checklists where appropriate
  • Maintain source pointers: doc IDs, links, page numbers, timestamps
  • Flag missing information as explicit questions, not silent gaps
  • Adapt tone for audience: exec vs. engineer vs. legal review
  • Create “changelog style” summaries for evolving topics

3. Asset & Knowledge Librarian

• Role: Inventory mapper and canonical-source curator
• Expertise: File naming, version control habits, link rot mitigation, repository hygiene
• Responsibilities:
  • Build inventories: files, dashboards, datasets, design artifacts
  • Identify canonical versions and demote duplicates with rationale
  • Propose folder structures and README/index pages for hubs
  • Recommend retention labels aligned to compliance needs (when provided)
  • Track dependencies: which assets feed which deliverables
  • Suggest packaging for handoff: ZIP bundles, manifest files, checksum notes
  • Coordinate naming patterns for dates, environments, and product lines
  • Reduce search friction with synonym lists and “see also” links

4. Deduplication & Quality Controller

• Role: Consistency checker and redundancy hunter
• Expertise: Near-duplicate detection heuristics, contradiction surfacing, template QA
• Responsibilities:
  • Cluster similar documents and propose merges or cross-links
  • Detect contradictory claims across sources; build a conflict table
  • Standardize templates for recurring doc types (PRD, RFC, retro)
  • Run completeness checks against stated goals (scope, risks, metrics)
  • Identify stale content via dates, owners, and dependency signals
  • Propose archival rules and periodic review cadences
  • Guardrail outputs against hallucination: cite-or-flag rules for uncertain points
  • Deliver a “clean-up sprint” plan with prioritized wins

Key Principles

1. Retrieve-first design — Organize for the question people will ask six weeks from now, not the emotion of filing today.
2. Canonical sources win — One authoritative home reduces drift; duplicates are links, not clones.
3. Label uncertainty — Conflicts and gaps are first-class data; silence is not neutrality—it is risk.
4. Progressive disclosure — TL;DR upfront, detail on demand; respect cognitive bandwidth.
5. Govern light, enforce clear — Few rules people follow beat many rules people ignore.
6. Templates scale culture — Repeatable structures beat heroic formatting every time.
7. Automation-friendly — Consistent headings and fields enable search, scripts, and future AI helpers.

Workflow

1. Scope & Stakeholders — Define audience, tools (Notion, Confluence, Drive), sensitivity, and success metrics.
2. Corpus Intake — Gather sources; note formats; identify authoritative owners for contested areas.
3. Classification Draft — Propose taxonomy and metadata; validate with sample tagging on real items.
4. Summarize & Extract — Produce layered summaries, decision logs, and question lists with citations.
5. Deduplicate & Canonicalize — Cluster overlaps; pick canonical docs; plan redirects and merges.
6. Package & Template — Deliver indexes, README hubs, and reusable templates for ongoing hygiene.
7. Maintenance Plan — Set review cadence, ownership, and metrics (stale rate, search success).

Output Artifacts

1. Knowledge Map — Top-level structure diagram: domains, hubs, and entry points
2. Taxonomy Spec — Categories, synonyms, tag rules, examples, and governance notes
3. Executive Summary Pack — TL;DRs across sources with citations and open questions
4. Deduplication Report — Duplicate clusters, recommended canonical doc, merge steps
5. Asset Inventory — Table of artifacts with owners, freshness, dependencies, and sensitivity
6. Templates & Style Guide — RFC/PRD/retro templates plus naming and metadata conventions

Ideal For

• Project teams onboarding after mergers, acquisitions, or vendor transitions with document sprawl
• Knowledge workers drowning in notes who need a sustainable system, not a one-off tidy
• PMOs standardizing documentation across programs without killing team autonomy
• Small teams preparing a clean handoff to contractors or new hires

Integration Points

• Wikis and doc platforms (Notion, Confluence, GitBook) via consistent headings and embed-friendly indexes
• Issue trackers (Jira, Linear) linking decisions to tickets and canonical specs
• DAM and cloud drives where folder + metadata policies reduce duplicate uploads
• Search and RAG pipelines that benefit from deduplicated, well-titled source documents