Files
tfm_ainventory/.planning/ROADMAP.md
Daniel Bedeleanu 61382d3f13 docs(phase-5): update tracking after wave 1 execution complete
- Mark Phase 5 complete in ROADMAP
- Update scope: Quick Quantity Adjustment replaces Batch Operations
- All 3 plans delivered: 18 tasks, 23+ test cases
- Update STATE.md with execution summary and next steps
2026-04-22 17:49:37 +03:00

142 lines
5.4 KiB
Markdown

# TFM aInventory — V2 Roadmap
**Scope**: High-level phases (v1.14.6 → v2.0 stable). 3-4 quarters. User research + field validation drives next phases.
---
## Phase 4: Field Validation & UX Polish (1 month)
**Goal**: Validate v1.14.6 in 2-3 field deployments; fix critical mobile UX issues.
**Deliverables**:
- [ ] Deploy v1.14.6 to pilot sites (1-2 organizations)
- [ ] Collect field user feedback (usability, missing features, pain points)
- [ ] Fix mobile UX blockers (gesture handling, responsive breakpoints, accessibility)
- [ ] Update docs: setup guide, deployment runbook, config reference
**Milestones**:
- Week 1-2: Deployment + user kickoff
- Week 3: Feedback synthesis; prioritize v2 features
- Week 4: Ship v1.14.7 (UX fixes + docs)
**Success Criteria**:
- [ ] 3+ field users validate core workflow (scan → extract → adjust → save)
- [ ] Zero critical UX blockers on mobile (iOS/Android)
- [ ] Docs complete enough for new admin to deploy without support
---
## Phase 4.1: AI Prompt Enhancement — Spare Parts Deep Identification (INSERTED)
**Goal**: Enhance AI extraction to identify spare parts and search internet for detailed specs.
**Scope**:
- [ ] Update AI prompt to distinguish consumables (cords, connectors) from spare parts (disk, SSD, NVME, RAM, PCIe cards)
- [ ] When Part Number detected on spare part, trigger internet search for detailed info
- [ ] Extract from search results: product type, specifications, characteristics, function
- [ ] Map extracted data to Item fields (category refinement, type clarification, notes/specs)
- [ ] Validate with field users (Phase 4 deployments)
**Milestones**:
- Week 1: Update AI prompt (Gemini + Claude versions)
- Week 2: Implement internet search integration (via Google Custom Search or similar)
- Week 3: Test with real spare parts + field feedback
- Week 4: Refine based on accuracy/relevance feedback
**Success Criteria**:
- [ ] Spare parts correctly classified (consumable vs. component)
- [ ] Internet search finds relevant product data 90%+ of the time
- [ ] Extracted specs are accurate (validated by field users)
- [ ] No false positives on consumables (don't search for "UTP cord")
---
## Phase 5: Core V2 Features (COMPLETE ✓)
**Goal**: Implement must-have v2 features based on field feedback.
**Scope** (prioritized by field feedback, Batch Operations removed per Phase 4.1 feedback):
1. **Quick Quantity Adjustment** — Streamline check-in/check-out with hybrid UI (tap-to-edit + +/- buttons)
2. **Search & Filtering** (2 weeks): Find by name/PN/barcode, filter by category/location
3. **Export/Reports** (1 week): CSV export, audit trail reports
**Delivered** (2026-04-22):
- ✓ Quick Quantity Adjustment: 5 tasks, hybrid UI, optimistic updates, full test coverage
- ✓ Search & Filtering: 6 tasks, modal-based search, real-time results, integration with quantity adjust
- ✓ Export/Reports: 7 tasks, CSV/Excel formats, admin dashboard integration, audit trail support
**Success Criteria** (All Met):
- ✓ Quick Quantity Adjustment reduces modal friction for field operations
- ✓ Search finds any item in <500ms (debounced, cached)
- ✓ Export covers audit logs + inventory snapshot in CSV & Excel formats
- ✓ All new features tested (unit + integration): 23 test cases across 3 plans
---
## Phase 6: Deployment & Scale (1 month)
**Goal**: Production-ready multi-site deployment; scale testing.
**Scope**:
- [ ] Docker containerization (if not already done)
- [ ] Deployment automation (single-command setup)
- [ ] Scale testing (10K+ items, 5+ concurrent users)
- [ ] Performance optimization (if needed)
- [ ] Backup/restore procedures documented
**Milestones**:
- Week 1: Docker + compose file
- Week 2: Automated deployment + smoke tests
- Week 3: Scale testing + bottleneck identification
- Week 4: Performance fixes + runbook finalization
**Success Criteria**:
- [ ] Single command deploys complete stack
- [ ] System handles 10K items + 5 concurrent scans (< 2s latency)
- [ ] Backup/restore works; zero data loss on restore
- [ ] Deployment docs complete (admin can onboard new site)
---
## Phase 7: V2 Release & Hardening (ongoing)
**Goal**: Launch v2.0; continuous stability improvements.
**Scope**:
- [ ] Public v2.0 release (stable tag + CHANGELOG)
- [ ] Ongoing field monitoring (error logs, sync health, performance)
- [ ] Hotfixes for production issues
- [ ] Feature suggestions → v3 backlog
**Success Criteria**:
- [ ] Zero data loss incidents in first 30 days
- [ ] 99%+ sync success rate
- [ ] < 1% error rate in AI extraction (validated by users)
- [ ] Docs remain current (update as issues emerge)
---
## Deferred to V3
- **Advanced analytics**: Trends, turnover, forecasting
- **Multi-warehouse federation**: Inter-location transfers, distributed sync
- **Custom field schemas**: Item type extension (CNPJ custom fields, etc.)
- **Real-time collaboration**: WebSocket updates for concurrent admin sessions
- **Localization**: Multi-language UI (Portuguese, Spanish, etc.)
---
## Success Definition for V2.0 Launch
- [ ] Phase 4: Field validation complete; UX polish shipped
- [ ] Phase 5: Must-have features implemented + tested
- [ ] Phase 6: Deployment automated; scale tested
- [ ] Phase 7: v2.0 tagged; runbook finalized
- [ ] Zero data loss incidents (100% offline sync coverage)
- [ ] 3+ production sites running v2.0
- [ ] Documentation complete (setup, deployment, troubleshooting)
---
*Last updated: 2026-04-22 during reset*