Refined taxonomy and routing for grouped knowledge

• Day: 2026-05-25
• Time: 11:20 to 11:30
• Project: Dev
• Workspace: WP 2: Operational
• Status: Completed
• Priority: MEDIUM
• Assignee: Matías Nehuen Iglesias
• Tags: Schema-Design, Taxonomy, Routing, Knowledge-Organization, Prompt-Engineering, Data-Governance

Description

Session Goal

Refine the extraction taxonomy and downstream routing rules for an AI-based knowledge organization pipeline, with emphasis on case grouping, publication lanes, and sensitivity handling.

Key Activities

• Reviewed multiple reflection-style notes about schema and enum tuning across organizational, political, public-sector, and workflow concept families.
• Evaluated where current enums were too coarse, especially around internal organizational dynamics versus formal institutional cases.
• Proposed introducing a new organizational_case group kind to prevent overmerging and better separate internal group dynamics from formal cases.
• Clarified boundaries for collaboration_example, technical_asset, political strategy, economic policy, and public-sector/state-capacity concepts.
• Considered adding post-consolidation metadata such as claim_posture, case_reality_status, move_role, and review/merge flags to enrich grouped atoms without rerunning extraction.
• Repeatedly emphasized a deterministic inventory/classification layer before any further AI pass, so routing decisions are stable and reusable.

Achievements

• Established a clearer downstream architecture for grouped atoms and claims.
• Identified that collection-level buckets are mostly stable, so the next bottleneck is materializing coherent inventories and routing them into the correct publication lanes.
• Confirmed the need to preserve private-casebook boundaries while separating objective wiki material from sensitive or self-positioned content.
• Converged on a narrower, iterative scope: taxonomy refinement and routing tweaks rather than a full pipeline redesign.

Pending Tasks

• Implement or test the new organizational_case group kind in the schema.
• Decide whether to add claim_posture, case_reality_status, and move_role now or defer them to a later iteration.
• Define final routing rules for public-sector technology, state capacity, political rhetoric, and organizational conflict content.
• Build the deterministic inventory classifier that assigns final publication lanes and review flags after consolidation.

Evidence

• source_file=2026-05-25.sessions.jsonl, line_number=10, event_count=0, session_id=1654f415bf49a2515d541de4c81ca8bc1be99849a841201a46268d059e04f630
• event_ids: []