Appearance
Timeline and Dependencies
This document maps dependencies between epics and tasks across all Intelligence Architecture projects. Use this to identify the critical path, parallel work opportunities, and blocking relationships.
Epic-to-Epic Dependencies
Critical Path (must be built in order)
Platform Core (Shared Types)
│
├──► DMC Vertical (Post-Spike)
│ │
│ ├──► DMC Data Enrichment
│ ├──► DMC Data Quality
│ └──► DMC Advanced Extraction
│
├──► Dining Vertical
├──► Experiences Vertical
├──► Transportation Vertical
├──► Hotels Vertical
├──► Flights Vertical
├──► Spa & Wellness Vertical
└──► Event Tickets Vertical
│
└──► Cross-Vertical Search
│
├──► Entity Resolution
└──► Itinerary ConstructionDependency Matrix
| Epic | Depends On | Blocked By | Enables |
|---|---|---|---|
| DMC Vertical | |||
| DMC Data Enrichment | Platform Core (shared types) | — | DMC Data Quality |
| DMC Data Quality | DMC Data Enrichment | — | DMC Advanced Extraction |
| DMC Advanced Extraction | DMC Data Quality | — | Production readiness |
| Dining Vertical | |||
| Dining Data Acquisition | Platform Core | — | Dining Schema |
| Dining Schema and Storage | Dining Data Acquisition | — | Dining Agent Integration |
| Dining Integration with Agent | Dining Schema | — | Agent system |
| Experiences Vertical | |||
| Experiences Data Acquisition | Platform Core | — | Experiences Schema |
| Experiences Schema and Storage | Experiences Data Acquisition | — | Booking Integration |
| Experiences Booking Integration | Experiences Schema | — | Agent system |
| Transportation Vertical | |||
| Transportation Data Acquisition | Platform Core | OAuth requirements | Transportation Schema |
| Transportation Schema and Storage | Transportation Data Acquisition | — | Agent system |
| Hotels Vertical | |||
| Hotels Data Acquisition | Platform Core | — | Hotels Schema |
| Hotels Schema and Storage | Hotels Data Acquisition | — | Real-Time Integration |
| Hotels Real-Time Integration | Hotels Schema | — | Agent system |
| Flights Vertical | |||
| Flights Data Acquisition | Platform Core | GDS relationships | Flights Schema |
| Flights Schema and Storage | Flights Data Acquisition | — | Agent system |
| Spa & Wellness Vertical | |||
| Spa & Wellness Data Acquisition | Platform Core | — | Schema |
| Spa & Wellness Schema and Storage | Data Acquisition | — | Agent system |
| Event Tickets Vertical | |||
| Event Tickets Data Acquisition | Platform Core | — | Schema |
| Event Tickets Schema and Storage | Data Acquisition | — | Agent system |
| Platform Core | |||
| Unified Data Models | — | — | Cross-vertical search |
| Cross-Vertical Search | Unified Data Models, 2+ verticals | — | Entity Resolution |
| Entity Resolution | Cross-Vertical Search | — | Itinerary Construction |
| Itinerary Construction | Entity Resolution | — | Agent system |
| Platform Infrastructure | All verticals | — | Production readiness |
| Platform Operations | |||
| Crawler Framework Enhancements | — | — | All verticals |
| Data Quality Monitoring | All verticals | — | Production readiness |
| Observability Pipeline | — | — | Alerting |
| Production Deployment | All above | — | Go-live |
Task-Level Dependencies Within Epics
DMC Vertical (Post-Spike)
| Task | Depends On | Notes |
|---|---|---|
| Implement DMC Finder crawler | Platform Core | 10,000+ DMCs, 120+ countries |
| Build Traveller Made enrichment | DMC Finder | |
| Implement Virtuoso advisor network | Traveller Made | 20,000+ luxury advisors |
| Build cross-source deduplication | All sources | |
| Implement email validation (MX check) | Extraction | |
| Build phone normalization (E.164) | Extraction | |
| Implement confidence scoring | Extraction | |
| Build department-specific email extraction | Extraction | reservations@, sales@, booking@ |
| Implement JS-rendered contact page handling | Extraction | |
| Build PDF brochure parsing | Extraction | |
| Implement cookie consent handling | Extraction | |
| Docker containerization | All above | |
| Implement incremental crawl support | Crawler framework | |
| Build rate limit enforcement per source | Crawler framework | |
| Implement error alerting | Observability |
Dining Vertical
| Task | Depends On | Notes |
|---|---|---|
| Build Resy crawler with rate limiting | Platform Core | |
| Build OpenTable crawler | Platform Core | Parallel with Resy |
| Build Yelp crawler | Platform Core | Parallel with Resy |
| Build Tock crawler | Platform Core | Parallel with Resy |
| Design restaurant data model | — | Can start early |
| Implement SQLite tables for dining | Data model | |
| Build FTS5 search on restaurant fields | SQLite tables | |
| Implement restaurant-specific MCP tools | FTS5, Data model | search_restaurant, get_restaurant_details, list_cuisines, find_by_city, catalog_stats |
| Connect Dining Research Agent to MCP server | MCP tools | |
| Implement real-time availability checks | Partner APIs | |
| Build time-slot display in Advisor Console | Availability |
Experiences Vertical
| Task | Depends On | Notes |
|---|---|---|
| Build Viator crawler | Platform Core | |
| Build GetYourGuide crawler | Platform Core | Parallel with Viator |
| Build Klook crawler | Platform Core | Parallel with Viator |
| Handle partner API authentication | All crawlers | |
| Design experience data model | — | Can start early |
| Implement SQLite tables and FTS5 | Data model | |
| Build experiences-specific MCP tools | FTS5, Data model | |
| Implement availability checking | Partner APIs | |
| Build booking flow integration | Availability | Hand-off to Atlantes |
Transportation Vertical
| Task | Depends On | Notes |
|---|---|---|
| Build Uber crawler/API integration | Platform Core | OAuth required |
| Build Lyft crawler/API integration | Platform Core | Parallel with Uber |
| Build Blacklane crawler/API integration | Platform Core | Parallel with Uber |
| Build car rental API integrations | Platform Core | Parallel with Uber |
| Handle OAuth requirements | All crawlers | |
| Design transportation data model | — | Can start early |
| Implement SQLite tables and FTS5 | Data model | |
| Build transportation-specific MCP tools | FTS5, Data model |
Hotels Vertical
| Task | Depends On | Notes |
|---|---|---|
| Build Booking.com crawler | Platform Core | |
| Build Expedia crawler | Platform Core | Parallel with Booking.com |
| Build Virtuoso integration | Platform Core | Luxury-specific |
| Design hotel data model | — | Can start early |
| Implement SQLite tables and FTS5 | Data model | |
| Build hotel-specific MCP tools | FTS5, Data model | |
| Implement real-time availability checking | Partner APIs | |
| Build pricing API integration | Availability | |
| Implement room type and amenity filtering | Pricing |
Flights Vertical
| Task | Depends On | Notes |
|---|---|---|
| Build Amadeus API integration | Platform Core | |
| Build Skyscanner crawler | Platform Core | Parallel with Amadeus |
| Build Google Flights crawler | Platform Core | Parallel with Amadeus |
| Handle GDS-dependent partner relationships | All crawlers | |
| Design flight data model | — | Can start early |
| Implement SQLite tables and FTS5 | Data model | |
| Build flight-specific MCP tools | FTS5, Data model |
Platform Core
| Task | Depends On | Notes |
|---|---|---|
| Define shared base types (Location, Contact) | — | Start here |
| Implement cross-vertical schema consistency | Base types | |
| Build schema versioning and migration system | Schema consistency | |
| Implement search_destination tool | 2+ verticals | |
| Build ranking and filtering across verticals | search_destination | |
| Implement cross-vertical result merging | Ranking | |
| Build EntityResolver | Cross-vertical search | |
| Implement deduplication across verticals | EntityResolver | |
| Implement suggest_itinerary tool | Entity Resolution | |
| Build itinerary ranking and recommendation | Itinerary tool | |
| Implement Docker + CI/CD pipeline | — | Can start early |
| Build monitoring and alerting | Observability | |
| Implement caching layer (Redis) | — | Can start early |
| Build user authentication (JWT) | — | Can start early |
| Implement notification system | — | Can start early |
Platform Operations
| Task | Depends On | Notes |
|---|---|---|
| Implement incremental crawling | Crawler framework | |
| Build change detection for directory pages | Incremental crawl | |
| Implement on-demand re-crawl | Incremental crawl | |
| Implement per-vertical coverage metrics | All verticals | |
| Build quality regression alerts | Coverage metrics | |
| Implement freshness tracking | All verticals | |
| Build stale record alerting | Freshness tracking | |
| Implement Prometheus metrics endpoint | — | Can start early |
| Build structured JSON logging | — | Can start early |
| Implement per-record tracing | — | Can start early |
| Build source health tracking | Structured logging | |
| Build k8s CronJob scheduling | Container image | |
| Implement SQLite backup to object storage | Persistent volume | |
| Build MCP server HTTP/SSE transport | MCP server | |
| Implement health and readiness probes | MCP server | |
| Build CI/CD pipeline | All above | Last task |
Parallel Work Opportunities
Wave 1 (can start immediately)
- Platform Core (shared types, schema)
- Crawler Framework Enhancements (incremental crawl)
- Observability Pipeline (Prometheus, logging)
- DMC Vertical (post-spike work)
Wave 2 (after Platform Core)
- Dining Vertical (acquisition + schema)
- Experiences Vertical (acquisition + schema)
- Hotels Vertical (acquisition + schema)
- Flights Vertical (acquisition + schema)
- Transportation Vertical (acquisition + schema)
- Spa & Wellness Vertical (acquisition + schema)
- Event Tickets Vertical (acquisition + schema)
Wave 3 (after Wave 2 verticals)
- DMC Data Enrichment + Quality + Advanced Extraction
- MCP Server Production Transport
- Data Quality Monitoring
Wave 4 (after multiple verticals exist)
- Cross-Vertical Search
- Entity Resolution
Wave 5 (after Entity Resolution)
- Itinerary Construction
Wave 6 (after all above)
- Platform Infrastructure (CI/CD, auth, caching)
- Production Deployment
Critical Path Summary
The minimum viable path to a working Intelligence Platform:
1. Platform Core (shared types, Location, Contact)
2. DMC Vertical (post-spike: enrichment, quality, extraction)
3. Dining Vertical (acquisition, schema, MCP tools)
4. MCP Server Production Transport (HTTP/SSE)
5. Crawler Framework Enhancements (incremental crawl)
6. Observability Pipeline (metrics, logging, alerts)
7. Data Quality Monitoring
8. Production DeploymentEstimated critical path: ~10-12 weeks if fully staffed and unblocked.
Risk Areas
| Risk | Impact | Mitigation |
|---|---|---|
| Partner API access delays | High | Start with public sources, escalate API access |
| JS-rendered pages block extraction | Medium | Use Playwright for specific sources, or accept lower coverage |
| Data quality regressions | Medium | Implement quality monitoring early, alert on regressions |
| SQLite scaling limits | Low | Monitor growth, plan Postgres migration at 10x |
| Cross-vertical search complexity | Medium | Start with simple ranking, iterate based on agent feedback |
| OAuth requirements for Transportation | High | Start OAuth flows early, escalate partner conversations |