v1.5Sprint 1
Deadline
Jul 17, 2026
Trust Foundation
Jul 6 - Jul 17, 2026
Product roadmap
From local analytics to MEGA-ready evidence infrastructure.
The LDT roadmap sequences v1.5 through v1.9 across two-week sprint deadlines, then targets v2 migration of the full backend and frontend to the World Bank's MEGA platform in October 2026.
- Release milestones
- 6
- Delivery workstreams
- 4
- Release gates
- 8
Release horizons
Each dated version unlocks the next product capability.
The roadmap is intentionally layered: first make the current app trustworthy, then add evidence-backed drafting, scale country and registry workflows, move into scenarios and monitoring, and finally migrate the full product to MEGA.
Two-week sprint cadence
July to October 2026 delivery timeline
Deadlines land at sprint close, starting Friday, July 17, 2026. The v2 milestone targets MEGA platform migration in October 2026.
- v1.6Sprint 2
Deadline
Jul 31, 2026
Evidence to Brief
Jul 20 - Jul 31, 2026
- v1.7Sprint 3
Deadline
Aug 14, 2026
Scaling Platform
Aug 3 - Aug 14, 2026
- v1.8Sprint 4
Deadline
Aug 28, 2026
Decision Studio
Aug 17 - Aug 28, 2026
- v1.9Sprint 5
Deadline
Sep 11, 2026
Delivery Loop
Aug 31 - Sep 11, 2026
- v2Sprint 7
Deadline
Oct 31, 2026
MEGA Platform Migration
Oct 19 - Oct 31, 2026
v1.5 deadline
Jul 17, 2026
Trust Foundation
Make the current LDT reliable, consistent, and trustable, with AI brief outputs saved and replayed before heavier public investment workflows.
Sprint 1
Jul 6 - Jul 17, 2026
- Release metadata source of truth
- AI Brief Cache and Run Store
- Country Trust Card and evidence gap badges
- Command center polish, export MVP, responsive states
Release gate: All public pages, exports, and AI stages use the same release metadata, caveats, source fingerprints, and cached run records.
v1.6 deadline
Jul 31, 2026
Evidence to Brief
Convert diagnostics into evidence-backed planning briefs, opportunity families, and concept-note starters.
Sprint 2
Jul 20 - Jul 31, 2026
- AI Planning Brief Report and AI Audit Drawer
- Investment Opportunity Finder and Evidence Graph MVP
- Concept Note Starter and recommendation review flow
Release gate: Every recommendation is traceable to evidence links, caveats, and a human review status.
v1.7 deadline
Aug 14, 2026
Scaling Platform
Make country onboarding repeatable and introduce a lightweight public investment lifecycle view.
Sprint 3
Aug 3 - Aug 14, 2026
- Country Onboarding Factory and validation reports
- Document Intelligence Workbench and Trust Center
- PIM Registry Beta, OC4IDS-ready mapping, observability
Release gate: A new country can be staged from manifests, data packages, source registries, and preview releases.
v1.8 deadline
Aug 28, 2026
Decision Studio
Add transparent spatial prioritization, climate-risk screening, scenario analysis, and workshop-ready outputs.
Sprint 4
Aug 17 - Aug 28, 2026
- Geospatial Prioritization Studio and hazard screen
- Scenario Builder with weights, sensitivity, and caveats
- Counterpart Mode and controlled API/data product
Release gate: Scenario outputs show the weights, evidence, assumptions, map, table, caveats, and sensitivity.
v1.9 deadline
Sep 11, 2026
Delivery Loop
Link upstream planning to delivery monitoring, procurement transparency, asset records, and ex-post learning.
Sprint 5
Aug 31 - Sep 11, 2026
- GEMS, KoBo, or ODK-style field evidence imports
- Procurement bridge, asset registry, transparency portal
- Cost, delay, anomaly flags and evaluation learning loops
Release gate: Project delivery evidence can inform public transparency and future investment decisions without hiding the rule logic.
v2 deadline
Oct 31, 2026
MEGA Platform Migration
Migrate the entire LDT backend and frontend to the World Bank's MEGA platform while preserving evidence lineage, country routes, exports, and auditability.
Sprint 7
Oct 19 - Oct 31, 2026
- MEGA frontend shell migration and route parity
- Backend service, data, auth, and deployment migration
- Cutover rehearsal, rollback plan, and stakeholder acceptance
Release gate: MEGA production cutover is rehearsed, reversible, security-reviewed, and validated against existing country workflows.
Execution map
The roadmap is both frontend and backend work.
A credible LDT expansion needs visible product surfaces, typed data services, evidence governance, and a repeatable country operating model to advance together.
Frontend experience
The user-facing route from command center to brief, scenario studio, and transparency surfaces.
- 1Trust Cards, evidence badges, AI cache provenance, command center actions
- 2Planning Brief, Audit Drawer, Opportunity Finder, review status controls
- 3Document Workbench, Trust Center, PIM Registry, admin observability
- 4Scenario Builder, Counterpart Mode, geospatial prioritization workspace
- 5MEGA-compatible frontend shell, navigation, theming, and route parity
Backend and data model
The services, manifests, validation states, and graph links that let every visible claim be traced.
- 1Typed release metadata provider, AI run store, cache keys, caveat model
- 2Evidence graph nodes and edges, opportunity triggers, export lineage
- 3Country manifests, document states, registry schema, OC4IDS-ready fields
- 4Climate/geospatial layers, scenario packages, API/download contracts
- 5MEGA backend services, data migration, auth, deployment, and rollback plan
AI and evidence governance
The policy layer that keeps AI in a synthesis role and makes source quality visible at the point of use.
- 1No source, no answer; persist and replay every LLM-backed brief stage
- 2Prompt versions, source fingerprints, audit records, human review status
- 3Readiness gates for translations, OCR, validation, incident handling
- 4AI summarizes tradeoffs only after deterministic outputs are shown
- 5MEGA migration preserves audit records, source fingerprints, and review states
Operating model
The repeatable country delivery motion for analysts, data teams, counterparts, and future transparency users.
- 1Country trust review and release QA
- 2Analyst-reviewed opportunities and concept-note starter exports
- 3Country onboarding factory and document operations queue
- 4Workshop packs, data products, field monitoring, public learning loop
- 5MEGA cutover readiness, stakeholder signoff, support model, and training
Governance
Ambition only ships through release gates.
The roadmap treats quality, evidence, accessibility, and AI safety as part of the product architecture. These gates protect users from unsupported outputs and keep decision authority where it belongs.
AI product position
AI is an evidence-synthesis and drafting assistant. It can support analysis, but it cannot act as appraiser, budget authority, procurement authority, or final decision maker.
Human authority stays explicit
LDT can summarize, compare, draft, classify, and explain. It must not approve projects, appraise investments, commit budgets, or make procurement claims.
Deterministic logic comes first
Opportunity and scenario logic should expose triggers, weights, assumptions, caveats, and supporting evidence before AI writes a rationale.
Evidence gaps are product states
Missing plans, stale sources, blocked translations, unmatched boundaries, and sparse coverage need visible badges and exportable caveats.
Every output is inspectable
AI outputs need cache status, run ID, model, prompt version, release ID, source fingerprint, retrieval IDs, generation time, caveats, and review status.
Release gates
Definition of ready for every future roadmap slice.
These gates should be applied before the team treats a roadmap slice as ready for stakeholder use, especially when outputs feed planning or investment conversations.
- 1Type checks
- 2Unit tests for data providers and evidence gating
- 3Integration tests for release metadata consistency
- 4Accessibility checks for new UI surfaces
- 5Security review for AI and server routes
- 6Manual QA against at least one country workspace
- 7Export QA for Markdown, CSV, and PDF-ready outputs
- 8Source and citation QA for AI outputs