Part 12 · Phased implementation
Implementation Roadmap — 5 Phases, 4 Gates
PH-0 design through PH-4 conditional target-scale operation — with measurable exit criteria, four evidence gates, and the 300,000 question deliberately parked at Gate 3.
The phasing logic
No calendar dates are assumed — the roadmap runs in months from authorization (M0) because the research refuses to invent an implementation date before feasibility is established. What is fixed is the evidence discipline: every phase has entry and exit criteria, every expansion passes a decision gate, and the 300,000 question is parked at Gate 3 — decided on five years of evidence, not on ambition.
The five phases
| Phase | Name | Duration | Cohort/year | Scope & activities | Entry/exit criteria | Decision gate |
|---|---|---|---|---|---|---|
| PH-0 | Design and legal preparation | 12 months | 0 (design only) | Ministry of Education + UGC + Cabinet Division working group; drafting of primary legislation for the compulsory element (voluntary pilot under existing university autonomy); MOU templates; task catalogues; digital platform specification | Entry: political decision to explore; Exit: legislation drafted, platform procured, 10-15 universities and 5 upazilas signed up | Decision gate 0: Ministry of Law confirms legal route (new statute for mandate; pilots on voluntary/credit-elective basis) |
| PH-1 | Voluntary pilot | 2 academic years | 5,000-10,000 students per year | 10-15 universities (mix: public direct, private, NU colleges in 3-4 districts) + host offices in 5 upazilas; voluntary enrollment in a credit-bearing service module; all safeguards live (protection officer, grievance, data MOUs) | Entry: Gate 0 passed; Exit criteria: completion rate 85%+, no unresolved retaliation cases, duration-neutrality verified (no graduation delay vs matched peers), cost within 150% of medium estimate | Decision gate 1: independent evaluation published; expansion only on evidence |
| PH-2 | Controlled expansion | 2 academic years | 50,000 students per year | 60+ universities across 8 divisions; NU college model begins (college-based placements near home); supervisor pipeline scaled to 2,500+; platform stress-tested | Entry: Gate 1 passed; Exit criteria: completion 87%+, placement rate 95%+, unit cost declining, zero systemic data breaches | Decision gate 2: second independent evaluation; Treasury commitment for full scale |
| PH-3 | National scale-up | 3 academic years | 100,000-200,000 per year | All divisions; full sector coverage; NU colleges carry the majority of volume; supervisor corps 5,000-10,000; quarterly public reporting of participation and quality indices | Entry: Gate 2 passed; Exit criteria: sustained quality indices, fraud controls verified by external audit | Decision gate 3: parliamentary review of the first 5 years; final decision on the 300,000 question |
| PH-4 | Target-scale operation (conditional) | Open-ended | Up to 300,000 per year | Only if Gates 1-3 all passed AND the capacity model (CSV 16) is validated by actual phase 3 data: this target REQUIRES full NU college participation (finding PL-15) and 15,000 supervisors | Entry: Gate 3 passed; standing rule: scale never expands faster than the supervisor pipeline (RSK-09) | Standing gates: annual independent audit; public dashboards; sunset review every 5 years |
The four decision gates
PH-4 (target-scale operation) then runs under standing gates: annual independent audit, public dashboards, and a five-yearly sunset review. The standing rule from RSK-09 applies throughout: scale never expands faster than the supervisor pipeline.
Pilot selection criteria
PH-1 pilot participants are selected transparently: 10–15 universities mixing public direct, private top/mid-tier institutions across divisions (readiness-verified: funding, coordinators, student records); 3–4 districts and 5 upazilas with office-type diversity (health, education, administration, local government, agriculture); 5,000–10,000 students/year across discipline clusters matched to the task catalogue; all safeguards live from day one — protection officer, grievance channels, data MOUs, insurance. Selection criteria and outcomes are published: the pilot is itself a transparency instrument.