KAIMo Integration

Status: Implemented (Beta, as of February 2026)

KAIMo is integrated in LLARS and ready for use. Admin and user panels are available.

Overview

KAIMo (AI‑assisted analysis and modeling) is a training tool for professionals dealing with child welfare risk. In LLARS it is implemented with two panels:

• KAIMO Admin Panel (for researchers): create cases, manage documents/hints, review results
• KAIMO Panel (for evaluators): work through cases, assign hints, submit assessments

Documentation

Document	Description	Status
Panel Concept	Detailed concept with admin/evaluator split	Reference
Request Assessment	Historic options assessment	Archive
Integration Concept	Technical analysis + architecture	Updated

Roles and Permissions

Role	Panel	Permissions
Researcher	Admin Panel	Create/edit cases, view results
Evaluator	User Panel	Work cases, submit assessments

Permissions

feature:kaimo:view       # See KAIMO area
feature:kaimo:edit       # Submit assessments
admin:kaimo:manage       # Manage cases (Admin)
admin:kaimo:results      # View results (Admin)

Core Features (current)

Admin Panel (Researcher)

1. Create cases - New case vignettes (draft/published)
2. Manage documents - case notes, reports, protocols
3. Define hints - hints + expected category/rating
4. Manage categories - standard categories + subcategories
5. Analyze results - aggregated outcomes and exports
6. Share/Import/Export - share cases, JSON export/import

User Panel (Evaluator)

1. Work cases - read documents, analyze hints
2. Assign hints - categories + rating (risk/resource/unclear)
3. Final verdict - overall assessment with reasoning
4. Progress tracking - status and overview per case

Implementation (as of Feb 2026)

• Backend: /api/kaimo admin + user routes, services, models, seeders
• Frontend: /kaimo hub, panel, case editor, assessment view
• Permissions: feature:kaimo:*, admin:kaimo:*

Historical Effort Estimate

Component	Effort
Database & API	8-12h
Admin Panel	16-24h
User Panel	20-30h
Assessment & Results	24-36h
Total	68-102h

AI integration (optional)

The infrastructure supports later AI integration. Estimated extra effort: 16-24h

Historical Next Steps (concept phase)

1. Concept review
2. Phase 1 implementation (DB & base API)
3. Phase 2-5 without AI integration
4. Add AI features later (optional)