Sindh IT Portal — Facilitation DeskSpecification documents
Englishاردوسنڌي
← All documents

PPP Vendor & Exit Management

The contractual and operational framework that binds the operator (MAAHIR) and host (Server4Sale) to the S&ITD's mission for the Sindh IT Portal — Facilitation Desk (SITP): service levels, performance KPIs, milestone payments, source-code escrow, IP ownership, knowledge transfer, and a tested, executable exit/handover plan that guarantees continuity of service regardless of the operator.

Field Value
Doc ID 23
Status Draft
Owner S&ITD / MAAHIR
Languages EN (master) · UR · SD
Related docs /specs/en/14-roadmap-release/, /specs/en/03-non-functional-reqs/, /specs/en/11-security-compliance/, /specs/en/17-analytics-kpis/, /specs/en/22-governance-legal/, /specs/en/15-tech-architecture/, /specs/en/24-trust-safety/

1. Purpose & How to Read This Document

This document defines every binding safeguard that makes the SITP Public-Private Partnership (PPP) safe for the Government of Sindh. It exists for one reason: the platform must never become operator-dependent. The Science & Information Technology Department (S&ITD) owns the mission, the product, and the public interest; MAAHIR builds and operates it under a contract whose terms ensure that S&ITD can always take the platform back, hand it to a successor, or run it in-house — with no loss of service, data, or knowledge.

It is written to be RFP- and tender-ready and is the authoritative source for: the S&ITD program office, the MAAHIR delivery and operations leadership, the Server4Sale (host) account team, the source-code escrow agent, internal audit, and any future successor operator or in-house team. The five non-negotiable PPP safeguards named in _context.md §6 and /specs/en/14-roadmap-release/ §2.2 are detailed here:

  1. Service-Level Agreements (SLAs) — availability, defect/severity response and resolution, support hours, and credits/penalties (§4).
  2. Source-code escrow — deposit at every release; release to S&ITD on trigger events (§7).
  3. Intellectual Property (IP) ownership — all deliverable IP vests in S&ITD (§8).
  4. Knowledge transfer (KT) — per phase and before any transition (§10).
  5. Exit / handover plan — a tested transition so continuity is never operator-dependent (§11).

These map to milestones M07–M11 in the milestone register (/specs/en/14-roadmap-release/ §7). Contract clauses in this document are numbered PPP-V-<nnn> (Vendor obligation) and are stable across the EN/UR/SD translations; they are referenced by ID from the contract, statement of work, and acceptance criteria.


2. PPP Model — Rationale & Structure

2.1 Why a PPP

A Public-Private Partnership gives S&ITD specialist speed and capability it cannot quickly build in-house — MAAHIR's product, AI, and trilingual engineering; Server4Sale's verified hosting — while retaining public ownership, accountability, and continuity through the binding safeguards in this document. The model matches the SIFC-style "single window" ambition: a private-operator cadence inside a government mandate. Pure in-house delivery would defer value by an estimated 12–18 months; pure outsourcing without safeguards would create operator lock-in and sovereignty risk (risk R6 in /specs/en/14-roadmap-release/ §12). The PPP with the safeguards below is the structure that resolves both.

2.2 Structure & ownership of the mission

S&ITD is the product owner and mission owner. MAAHIR is the operator under a build–operate–maintain contract. Server4Sale is the host under an approved sub-contract for infrastructure. The table below fixes who owns what.

Party Role Owns / provides Bound by
S&ITD Owner (product + mission) Legal mandate; per-department MoUs; data-sharing approvals (NADRA/SECP/FBR/SRB/PSEB); ADP budget line; acceptance authority. Owns all IP, brand, data. Government rules; this contract.
MAAHIR Operator (build + operate + maintain) Engineering team; product/design; QA; ops/SRE; AI engineering; trilingual content; managed services across the term. Builds under contract; retains no deliverable IP. This contract; SLAs; KPIs; escrow; exit plan.
Server4Sale Host (approved sub-contractor) Verified environment (Ubuntu 24.04.4 LTS, Node 22, MariaDB 10.11, nginx, Docker; 3.4 TB disk, 243 GB RAM); networking; backups; DR facility. Footer: "Powered by Server4Sale." Hosting sub-contract; availability; security.
Escrow agent Independent third party Holds source-code escrow deposits; releases to S&ITD only on contractually defined trigger events. Escrow agreement (§7).

2.3 What S&ITD always retains (incontestable)

Regardless of contract status, operator performance, or any dispute, S&ITD unconditionally retains: (a) the product and mission; (b) all deliverable IP (§8); (c) all production data and the database; (d) all credentials and access (revocable at will); (e) the domain, brand, and public identity; (f) the legal mandate and department MoUs; and (g) the right to invoke escrow, transition, and handover per this contract.


3. Contract Scope & Term

3.1 Scope — build, operate, maintain

The contract is a single integrated build + operate + maintain engagement covering the full V1 scope defined in _context.md §4 (modules A–Q) and delivered in phases per /specs/en/14-roadmap-release/. It is not a pure development contract followed by a separate maintenance contract; the operator is responsible for the platform's behaviour in production across the whole term, including: design, build, QA, release, managed services (L1/L2/L3), incident response, security operations, trilingual content maintenance, AI model upkeep, KB/SOP curation, analytics and reporting, and continuous improvement against the KPIs in §5.

Phase of contract Scope Aligns with Exit gate
Build (Phases 0–4) Discovery, MVP, GA, V1.1, V1.2; baselined specs; environments; escrow established; KT per phase. Milestones M01–M06 (/specs/en/14-roadmap-release/ §7). M06 (V1.2).
Operate + Maintain (sustainment) Managed services from V1.0 GA onward under the ADP-funded arrangement; patch/minor/major releases on the cadence in /specs/en/14-roadmap-release/ §13. Sustainment KPIs (§5). Contract end / termination.
Transition (exit) Tested handover to S&ITD, an in-house team, a new vendor, or a hybrid; escrow released if triggered. M10 (readiness), M11 (close-out). M11.

3.2 Term, renewal, and termination triggers

The initial term spans the build phases plus a defined operate/maintain window aligned to the ADP budget cycle. Renewal is not automatic — it follows a formal performance review (§16). Termination may be:

Termination-for-cause triggers (PPP-V-101): (1) material and persistent SLA breach unremedied after the cure period (§4.4); (2) material security breach attributable to operator negligence (§13); (3) unauthorized subcontracting of core scope (§12); (4) IP or data misuse (§8, §9); (5) conviction or debarment under anti-corruption/PEPRA rules (§14); (6) abandonment or sustained inability to perform. On any termination trigger, the transition plan (§11) and escrow release (§7) activate immediately.

PPP-V-102No termination, however caused, ends the operator's obligation to hand over all code, data, credentials, and knowledge. The exit obligations in §§7–11 survive termination and contract expiry.


4. Vendor Service-Level Agreements (SLAs)

4.1 Availability

The platform shall meet NFR-AVAIL-001 — ≥ 99.9 % monthly uptime for the portal, API, and docs site (≈ 43 min/month or ≈ 8.7 h/year error budget), excluding planned maintenance per NFR-AVAIL-002 (/specs/en/03-non-functional-reqs/ §3.2). Availability is measured by the public status-page probes (NFR-OBS-005). Disaster-recovery objectives are RPO ≤ 1 h (NFR-AVAIL-004) and RTO ≤ 4 h (NFR-AVAIL-005). Graceful degradation (NFR-AVAIL-003) keeps the core ticket journey working when an AI/OCR or integration dependency is down.

4.2 Severity matrix — defects and incidents

All defects in production software and incidents affecting service are classified by severity. The matrix fixes acknowledgement, response, and resolution targets the operator must meet; it aligns the roadmap's Sev-1 patch target (≤ 72 h, /specs/en/14-roadmap-release/ §13.2) with the contract.

Sev Definition (examples) Acknowledge Response (mitigation/ comms) Resolution (fix) Support window
P1 — Critical Production down; data loss; security breach; core file/route/resolve journey broken; ≥ 1 dept cannot work. ≤ 15 min ≤ 30 min (workaround or status comms) ≤ 72 h (hotfix); restore within RTO (4 h) 24×7
P2 — High Major feature broken for a cohort; SLA engine misfiring; an integration (NADRA/SECP/FBR/SRB/PSEB) down with no graceful fallback; severe trilingual break. ≤ 1 h ≤ 4 h ≤ 5 business days Business hours + on-call
P3 — Normal Minor feature defect; cosmetic trilingual issue; non-blocking performance regression; analytics tile stale. ≤ 4 h (next business day) ≤ 2 business days Next minor release (4–6 weeks) Business hours

Support hours (PPP-V-201). P1 is supported 24×7 with a named on-call rotation and escalation to the MAAHIR delivery lead within 30 min. P2 is supported business hours (09:00–18:00 PKT, Mon–Sat) plus on-call. P3 is supported business hours. "Business hours" exclude Sindh public holidays and use the same holiday calendar as the SLA engine (_context.md §5).

4.3 Breach, credits, and penalties

A breach is missing an acknowledge/response/resolution target, or falling below the 99.9 % availability in a calendar month. Breaches accrue service credits (a deduction against the next invoice) and, when material or persistent, escalate to termination for cause (PPP-V-101).

Breach type First occurrence Repeat within rolling 90 days Material / persistent
Availability < 99.9 % in a month Service credit = pro-rated fee × ((99.9 − actual) ÷ 100), capped per schedule. Credit escalates 2×. 3 consecutive months below target → cause for termination.
P1 resolution missed (> 72 h) Credit per schedule. Credit 2×; root-cause review mandatory. 2 missed P1s in a quarter → cause for termination.
P2 resolution missed (> 5 business days) Credit per schedule. Credit escalates. Chronic P2 backlog → performance-improvement plan (§16).
Security incident due to operator negligence (§13) No cure; credit + remediation at operator cost. Material breach → cause for termination.
Missed milestone gate payment condition (§6) Payment withheld until cured. Repeated slippage → contract review.

Cure period (PPP-V-202). Before a non-security material breach becomes cause for termination, the operator receives a written cure notice and 30 calendar days to remedy (15 days for P1 availability), with a formal remediation plan reviewed weekly. Failure to cure within the period converts the breach into cause.

4.4 Planned maintenance and communication

Planned maintenance is excluded from the 99.9 % budget only when it (a) is scheduled in a low-traffic window per NFR-AVAIL-002, (b) is notified to S&ITD ≥ 72 h in advance, and (c) is published on the public status page. Emergency maintenance during business hours counts against availability. Every change follows release governance (/specs/en/14-roadmap-release/ §13.3) with rollback tested before release.


5. Vendor Performance KPIs

The operator's operational performance is measured continuously against the analytics plane in /specs/en/17-analytics-kpis/. The KPIs below are the vendor-scorecard subset: each has a target, a measurement source (the same anl_mv_* rollups that feed dashboards, so there is one definition per metric — AP-5), a review cadence, and a penalty/remedy linkage. Targets are configurable per anl_cfg_* and not hard-coded (AP-6).

# KPI Target Source / definition Review Penalty/remedy linkage
V1 Availability (uptime) ≥ 99.9 % monthly NFR-AVAIL-001; status-page probes. Monthly §4.3 service credit; 3 consecutive misses → termination cause.
V2 MTTR — Mean Time To Resolve (incidents) P1 ≤ 72 h; breach recovery ≤ 48 h Defect resolution clock (§4.2); KPI-SLA-011 (breach recovery). Monthly Service credit per §4.3; chronic → performance-improvement plan.
V3 CSAT ≥ 4.2 (rolling 30 d) KPI-QLT-001 (post-resolution rating 1–5). Monthly Trend below target for 2 months → KT/content remediation at operator cost.
V4 Deflection rate ≥ 30 % (KB) and ≥ 45 % (chatbot) KPI-COM-012 (KB deflection); KPI-AIM-008 (chatbot deflection). Monthly Below target → KB/SOP enrichment + chatbot tuning (operator cost).
V5 AI cost efficiency Per-call cost within budget; on-prem share rising KPI-AIM-018 (per-engine cost), KPI-AIM-019 (calls/ticket), KPI-AIM-020 (on-prem/cloud share). Quarterly Overspend → engine-mix rebalancing (pluggable engines, _context.md §3) at operator cost.
V6 Security incident count 0 reportable incidents; 0 P0/P1 due to operator negligence Incident register; SEC-069 (IR plan), SEC-092 (breach notification ≤ 24 h). Monthly / per-incident Any negligence-attributable incident → credit + remediation at operator cost; material → termination cause (§4.3).
V7 SLA compliance (platform) ≥ 95 % KPI-SLA-001 (tickets meeting SLA ÷ resolved). Monthly Driven by departments, but operator must keep the engine accurate; engine defects → P2.
V8 Audit completeness = 100 % KPI-GOV-002 (state-changing actions with matching audit row). Quarterly Gap → hotfix; reflects operator control integrity.

One definition per metric. Because every vendor KPI is sourced from the same materialized views that feed the role dashboards, there is no separate "contract numbers" stream that can drift from operational reality. S&ITD and the operator always read the same numbers.


6. Milestone-Based Payments

Payment is milestone-based and gated, tied to the phase gates in /specs/en/14-roadmap-release/. No milestone payment is released until that milestone's exit criteria are met and signed off by the S&ITD product owner; each gated release also requires the concurrent PPP actions (escrow deposit M07, IP assignment note M08, KT note M09). The funding source is the dedicated ADP line (/specs/en/14-roadmap-release/ §10.2). Indicative payment structure (percentages are of the contracted build value; the operate/maintain window is fee-for-service on the cadence in §13 of the roadmap):

Milestone Gate (must be met to invoice) Concurrent PPP actions Payment basis
M01 Contract & kickoff Contract + SoW signed. Mobilization tranche.
M02 Discovery & baseline Baseline specs approved; env live; API requests lodged; escrow agent appointed & escrow agreement signed; ≥ 3 dept MoUs in draft. Escrow agreement; IP-assignment clauses confirmed. Discovery tranche.
M03 MVP pilot (soft launch) Must journey UAT green; limited cohort live; trilingual + security baseline verified. Escrow deposit (release 1); KT delivered; IP assignment note. Build tranche 1.
M04 V1.0 Public GA Go/No-Go checklist GREEN (§9.4 of roadmap); full Must complete; public dashboard live. Escrow deposit (release 2); KT delivered; IP assignment note. Build tranche 2 (largest).
M05 V1.1 Should-tier delivered; pen-test remediated; exam-gate enforced; e-sign live. Escrow deposit (release 3); KT. Build tranche 3.
M06 V1.2 Mobile apps live; Could-tier delivered. Escrow deposit (release 4); KT; sustainment transition confirmed. Build tranche 4.
Sustainment Continuous from GA; fee-for-service against KPI scorecard (§5). Per-release escrow deposit + KT note (/specs/en/14-roadmap-release/ §13.3). Recurring managed-services fee.
M10 / M11 Exit & handover Transition plan tested (dry run); takeover capability confirmed; close-out archived; escrow released if triggered. Escrow release (§7.4); final KT. Pre-negotiated transition funding.

PPP-V-301Withholding. S&ITD may withhold any milestone payment until its gate criteria and the concurrent PPP actions are evidenced. Cure-then-invoice applies; payments withheld are not penalties and are released on cure.


7. Source-Code Escrow

Source-code escrow is the structural defence against operator lock-in (risk R6). It guarantees that, on a trigger event, S&ITD receives everything needed to build, deploy, operate, and secure the platform independently of MAAHIR. The escrow is established in Phase 0 (critical-path dependency C5, /specs/en/14-roadmap-release/ §8) and the first deposit is made at M03.

7.1 Escrow lifecycle

The lifecycle below describes the continuous deposit–verify–release loop that runs across every release for the life of the contract.

stateDiagram-v2 [*] --> Established: Phase 0 — escrow agent appointed, agreement signed (C5) Established --> Deposit: At each release (M03/M04/M05/M06 + per-release) Deposit --> Verified: Escrow agent + S&ITD verify completeness & buildability Verified --> Established: Receipt recorded; next release pending Established --> Deposit: Next release Verified --> Triggered: Trigger event occurs (§7.3) Triggered --> Released: Agent releases materials to S&ITD Released --> IndependentOperation: S&ITD / successor builds & operates IndependentOperation --> [*] note right of Triggered Triggers: vendor default / bankruptcy / termination for cause / non-renewal / sustained inability to perform end note

Written description. The escrow agent is appointed and the escrow agreement signed during Phase 0 (dependency C5), before any production code exists — this is deliberate, so the mechanism is in place before it could ever be needed. At every release (M03, M04, M05, M06, and every post-GA minor/patch release per the cadence in /specs/en/14-roadmap-release/ §13), the operator deposits the release's complete buildable materials (§7.2). The agent and S&ITD verify the deposit for completeness and buildability — a deposit that cannot be built is treated as not deposited. A receipt is recorded and the system returns to the ready state, awaiting the next release. If a trigger event (§7.3) occurs, the deposit moves to Released: the agent hands the materials to S&ITD, which can then run the platform independently (in-house, via a successor, or hybrid) per the transition plan (§11). The verify step is what makes escrow meaningful — unaudited escrow is a false safeguard.

7.2 What is escrowed

Every deposit must be a complete, buildable, operable snapshot. The "can a competent team rebuild and run this?" test is the acceptance bar; a partial deposit does not satisfy the milestone gate.

Artefact class What's included Why it's required
Source code All application code (Next.js portal, NestJS API, Python AI/OCR service), in a Git repository snapshot with full history and tags matching the release. Rebuild the software.
Build & config package.json/lockfiles, Dockerfiles, CI/CD pipelines, environment templates, .env.example, build scripts, dependency manifests with pinned versions. Reproduce a build deterministically.
Database schema MariaDB migrations, schema dumps, seed/reference data, Prisma/TypeORM schema. Recreate the data layer.
AI assets Prompts, prompt templates, eval gold sets, model configs, engine-selection rules, fine-tuning artefacts if any. Operate the pluggable AI without re-deriving it.
Infrastructure-as-code Server/Server4Sale provisioning scripts, nginx configs, Docker Compose/K8s manifests, backup/restore scripts, DR runbooks. Stand up the environment on Server4Sale or elsewhere.
Documentation Architecture, API contract, data model, runbooks, SOPs, security controls, the full docs set. Understand and operate the platform.
Keys & secrets register Inventory of all secrets and their locations/rotation procedures (never the live secret values; live secrets are rotated on release — §9.3). Identify and rotate secrets after release.
Third-party licences Full dependency licence manifest, including the operator's pre-existing tooling licensed to the project (§8.2). Confirm the right to operate each component.

7.3 Release conditions (triggers)

The agent releases the escrow to S&ITD only on a contractually defined trigger, attested by S&ITD and (where disputed) confirmed by the dispute-resolution process (§15):

Trigger Effect
Vendor default unremedied after cure period (PPP-V-101, PPP-V-202) Immediate release.
Bankruptcy / insolvency / winding-up of MAAHIR Immediate release; no cure period.
Termination for cause by S&ITD Immediate release at termination.
Non-renewal at end of term Release at start of transition period (§11).
Sustained inability to perform (e.g., loss of key team, force majeure > 60 days) Release on S&ITD attestation.
Voluntary handover by mutual agreement Release per agreed transition plan.

PPP-V-401No trigger required for the deposit obligation. The operator must deposit at every release regardless of relationship health. Escrow is a standing obligation, not a remedy invoked only at exit.

7.4 Escrow agent, verification, and costs

The agent is an independent third party (not an affiliate of MAAHIR or Server4Sale), selected via a transparent process (§14). The escrow agreement specifies: (a) the verification standard (buildability test on each deposit); (b) the release mechanics and attestation; (c) the retention period (deposits retained for the contract term + the agreed liability window); and (d) cost allocation — the operator bears the cost of deposit and verification as a cost of doing business; S&ITD bears the cost of any release process. Verification evidence (build log from the agent's attempt) is shared with S&ITD at each deposit and is a precondition of the milestone payment (§6).


8. Intellectual Property (IP) Ownership

8.1 Ownership — S&ITD owns all deliverable IP

PPP-V-501All intellectual property created, authored, developed, or customized for the SITP project vests in S&ITD, the Government of Sindh, from the moment of creation. This includes all source code, schemas, prompts, designs, the Ajrak-inspired brand, the documentation set, configuration, runbooks, models, and derivative works.

IP is not transferred at exit — it vests in S&ITD continuously. The assignment is recorded per release as part of milestone M08 (/specs/en/14-roadmap-release/ §7) so there is no IP ambiguity at any point. S&ITD owns:

Asset Owner Notes
All deliverable source code S&ITD Escrowed (§7).
Database schema, data model, reference/seed data S&ITD Data ownership reinforced in §9.
AI prompts, eval sets, fine-tuning artefacts S&ITD Part of the deliverable.
Brand, name, logo, Ajrak identity, domain S&ITD Product identity is government property.
All production data and audit logs S&ITD Operator is a custodian, never an owner.
Documentation set (this docs repo) S&ITD Including UR/SD translations.
Architecture, runbooks, SOPs S&ITD Operability knowledge.

8.2 Pre-existing tooling (vendor background IP)

The operator retains ownership of pre-existing, independently-developed tools, libraries, and frameworks that pre-date and are independent of this project ("Background IP"). However, any Background IP used in the deliverable must be: (a) disclosed in the third-party licence manifest (escrowed, §7.2); (b) licensed to the project on a perpetual, royalty-free, irrevocable, worldwide basis sufficient for S&ITD to operate, maintain, and modify the platform — including after termination; and (c) free of encumbrances that would prevent S&ITD's use. If a Background-IP component cannot be so licensed, it must not be used in the deliverable. This clause survives termination (PPP-V-102).

8.3 Assignment and moral rights

Each operator staff member and subcontractor who contributes to the deliverable executes an IP-assignment / work-for-hire clause as a condition of contributing, and waives moral rights to the extent permitted by law, so that S&ITD's title is unencumbered. The operator warrants that the deliverable infringes no third-party rights and indemnifies S&ITD against any IP infringement claim arising from the operator's work.


9. Data & Access on Exit

9.1 Data is S&ITD's; the operator is a custodian

All production data — ticket data, company and representative data, department data, audit logs, AI call records, analytics aggregates, uploads, MoM — belongs to S&ITD. The operator holds and processes it only as a custodian under the contract, the privacy obligations in /specs/en/03-non-functional-reqs/ §PRIV, and the security controls in /specs/en/11-security-compliance/. Sovereign data (CNIC, NADRA payloads) is handled under the data-sharing approvals and must never leave the approved boundary.

9.2 Handover of data on exit

PPP-V-601On exit (any termination or non-renewal), the operator must hand over ALL data in usable, documented form within the transition period (§11), at no additional charge, and must retain no copy.

The data handover comprises: (a) a full, consistent database export (MariaDB dump matching the released schema); (b) all object storage (MinIO buckets — uploads, generated letters, exports); (c) all logs and audit archives within the retention window; (d) analytics aggregates (anl_mv_*, anl_cube_*, anl_pub_*); (e) the configuration tables (anl_cfg_*, feature flags, RBAC, holiday calendar); and (f) any data held in operator-controlled SaaS (Mailjet, SMS/WA providers) within the operator's access. Delivery is to a destination S&ITD controls (the Server4Sale environment or a successor).

9.3 Credentials and access — revoked, secrets rotated

On exit: (a) all operator access is revoked — Keycloak accounts, SSH keys, deployment tokens, CI/CD secrets, cloud console access, vendor portal access (Mailjet/SMS/WA/OCR/LLM), database credentials, and any break-glass accounts; (b) all secrets are rotated — the escrow's secrets register (§7.2) lists every secret; live secret values are never escrowed and must be rotated at handover so the departing operator cannot use them; (c) a joint access-revocation log is signed off by both parties; (d) the operator confirms in writing that it retains no production credentials and no copy of production data, and deletes any local/dev copies under attestation. This satisfies the assume-breach principle (SEC control family, /specs/en/11-security-compliance/ §3) — at no point after exit can a former operator reach the platform.


10. Knowledge Transfer (KT)

Knowledge transfer is a per-phase and per-event obligation (milestone M09), not a single dump at the end. The goal is that, at any point, a competent team could take over the platform using the documentation, runbooks, and recorded walkthroughs left to date. KT is a precondition of each milestone payment (§6) and is tracked to closure.

KT deliverable Form When
Architecture & decisions Living docs (this docs set) + recorded walkthroughs. Phase 0; updated per release.
Runbooks Deploy, rollback, restore, incident, DR failover, secret rotation. Phase 1 onward; reviewed annually (NFR-MAINT).
Code walkthroughs Recorded module walkthroughs (portal, API, AI service, analytics). Each phase exit.
Operations handover On-call runbook, escalation paths, vendor/provider contacts, known-issue log. From GA onward.
AI/eval handover Prompt library, eval methodology, engine-selection rationale, quality baselines. Phase 2 onward.
Admin training Super Admin console, feature flags, config tables, RBAC, analytics. GA + per major release.
Transition KT (final) Full deep-dive for S&ITD staff and/or successor/in-house team. Transition period (§11).

PPP-V-701Training of the S&ITD team or a named replacement team is a paid, scheduled deliverable, not goodwill. A minimum number of KT hours per phase and a structured transition-KT plan during the transition period are specified in the statement of work. KT attendance and material delivery are evidenced and sign-off is a precondition of the milestone payment.


11. Transition & Exit Plan

11.1 Principle — continuity is never operator-dependent

The exit/handover plan is tested and executable, not theoretical. It is rehearsed (dry-run) before the readiness review (M10) so that S&ITD has evidence — not a promise — that it can take the platform over. The plan assumes the worst case (an involuntary exit under a termination-for-cause trigger) and must still guarantee continuity.

11.2 Reconstitution options

On exit, S&ITD reconstitutes operations through one of three options (or a hybrid), chosen based on the readiness review:

Option What it means When chosen Dependencies
In-house takeover S&ITD runs the platform with its own (or newly hired) engineering/ops team. Where capability exists or is being built; maximum sovereignty. Trained team; runbooks; access to escrow + Server4Sale.
New vendor A successor operator takes over under a new PPP/contract. Where S&ITD wants private-operator cadence but not this operator. Procurement (§14); escrow + data handover; KT to successor.
Hybrid S&ITD owns and runs core ops; vendors provide specialist capacity (AI, hosting, content). Most likely steady-state evolution. Mixed staffing; clear interface boundaries.

11.3 Transition period

A defined transition period (sized in the statement of work; indicative ≥ 90 days for a full operator change, shorter for a hybrid shift) precedes close-out. During it: (a) the current operator continues to operate the platform and supports the successor/in-house team; (b) KT (§10), data handover (§9), escrow release (§7.4 if triggered), and access rotation (§9.3) occur in sequence; (c) the successor runs the platform in shadow/parallel until S&ITD accepts the cutover; (d) continuity of service is maintained throughout — there is no scheduled outage for handover. Funding for the transition period is pre-negotiated (/specs/en/14-roadmap-release/ §10.2) so it is executable even under a termination for cause.

11.4 Exit/handover transition flow

flowchart TD A([Exit decision / trigger]) --> B{Trigger type?} B -->|Non-renewal / voluntary| C[Notice given; transition period starts] B -->|Termination for cause| D[Escrow release invoked §7.3; access preserved until cutover] B -->|Bankruptcy / inability| D C --> E[Reconstitution option chosen §11.2] D --> E E --> E1[In-house takeover] E --> E2[New vendor via procurement §14] E --> E3[Hybrid] E1 --> F[Transition workstream] E2 --> F E3 --> F subgraph F[Transition workstream — transition period, ≥90 d, no outage] F1[Knowledge transfer §10] F2[Data handover §9.2] F3[Access rotation & revocation §9.3] F4[Escrow released & verified §7.4] F5[Successor shadow / parallel run] end F --> G{Readiness review M10} G -->|Not ready| F G -->|Ready| H[Cutover accepted by S&ITD] H --> I[Former operator access fully revoked] I --> J[Close-out & lessons learned M11] J --> K([S&ITD operating independently])

Written description. An exit decision or trigger (notice, termination for cause, or bankruptcy/inability) starts the transition. For a voluntary or non-renewal exit, notice opens the transition period; for an involuntary exit, the escrow release is invoked immediately while access is preserved just long enough to execute the cutover. S&ITD then chooses a reconstitution option (in-house, new vendor via the §14 procurement process, or hybrid). All three paths feed a single transition workstream running across the transition period: knowledge transfer (§10), data handover (§9.2), access rotation and revocation (§9.3), escrow release and verification (§7.4), and a successor shadow/parallel run — all while the platform stays live. A readiness review (M10) gates the cutover: if not ready, the workstream continues; if ready, S&ITD accepts the cutover, the former operator's access is fully revoked, and close-out (M11) archives lessons learned. The end state is S&ITD operating independently — the single invariant of the entire PPP.


12. Subcontracting Control

12.1 No subcontracting of core scope without approval

PPP-V-801The operator must not subcontract any core scope of the contract without S&ITD's prior written approval. Unauthorized subcontracting of core scope is a termination-for-cause trigger (PPP-V-101).

"Core scope" includes: product/engineering leadership, application development (portal/API/AI), QA, security operations, and incident response. Approval considers the subcontractor's capability, security posture, and the safeguards in this document (the subcontractor is bound to the same SLAs, security, IP, data, and exit obligations as the operator, and S&ITD's rights flow through).

12.2 Server4Sale — approved hosting sub-contractor

Server4Sale is pre-approved as the hosting sub-contractor (the "Powered by Server4Sale" arrangement), and the hosting sub-contract is a named, accepted part of the PPP. Server4Sale's obligations include the availability/security standards in §4.1 and /specs/en/11-security-compliance/, backup and DR facility support, and cooperation with the exit plan (host access rotation, environment handover). Changes to the hosting arrangement (e.g., moving hosts) require S&ITD approval and must preserve the escrow, data-portability, and continuity safeguards.

Subcontract type Approval needed? Notes
Hosting (Server4Sale) Pre-approved Named in contract; flow-through obligations.
AI/OCR cloud engines (Azure/Google/AWS) Pre-approved, pluggable Per _context.md §3; PII-redacted before cloud; self-hosted toggle available.
Delivery channels (Mailjet/SMS/WA) Pre-approved Standard providers.
Core engineering/QA/SecOps to a third party Required — prior written approval Unauthorized = cause for termination.
Any subcontract handling PII/sovereign data Required + security review Flow-through of §13 obligations.

13. Security Obligations

The operator's security obligations flow from the control catalogue in /specs/en/11-security-compliance/ and are contractually binding. The operator must implement, maintain, and evidence these controls for the life of the contract and during the transition period.

Obligation Requirement Reference
Staff vetting & background checks All operator staff with access to production or PII/sovereign data pass background checks (identity, employment, criminal where lawful) before access; re-vetting on role change. SEC SSDLC/trust; §13.2 below.
NDAs Every operator and subcontractor staff member with any access signs a binding non-disclosure agreement covering data, code, and operations; survives termination. /specs/en/11-security-compliance/.
Access controls Least-privilege, role-based, time-bound access; 2FA mandatory for operator staff (SEC-005); step-up for sensitive actions (SEC-006); just-in-time elevation; quarterly access reviews (KPI-GOV-003). /specs/en/11-security-compliance/.
Secure development OWASP ASVS Level 2 before go-live (SEC); secure-coding standard (SEC-081); security code review on every PR; annual security training (SEC-085). /specs/en/11-security-compliance/.
Breach notification Reportable incidents notified to S&ITD immediately and to CERT-PK within 24 h (SEC-092); blameless postmortem within 10 business days for P0/P1 (SEC-094). /specs/en/11-security-compliance/ §24.
Incident response Maintain the IR plan (SEC-069); 24×7 P1 on-call; forensic readiness (SEC-093); cooperate with CERT-PK advisories (SEC-068). /specs/en/11-security-compliance/.
Penetration testing & CII Independent pen-test remediated to agreed threshold before V1.1 (roadmap M05); CII registration submitted; align with Critical Information Infrastructure expectations. /specs/en/14-roadmap-release/ §5.4.
Data protection Encryption at rest/in transit; PII redaction before cloud AI (KPI-AIM-015); sovereign data within approved boundary; honour RTI statutory handling. /specs/en/03-non-functional-reqs/ §PRIV.

PPP-V-901Negligence-attributable security incidents carry no cure period. A confirmed incident caused by operator failure to maintain the controls above triggers an immediate service credit plus remediation at the operator's cost; a material such incident is cause for termination (PPP-V-101) and activates escrow release.

13.1 Breach notification timeline

A confirmed unauthorised access to or exfiltration of personal or sovereign data is a reportable incident: notify S&ITD within 1 hour of confirmation; notify CERT-PK within 24 h (SEC-092); notify affected individuals and the Sindh Information Commission where required. The operator cooperates fully with forensics and preserves evidence (SEC-093).

13.2 Personnel security

The operator maintains a current roster of all personnel with any access to the platform or its data, provided to S&ITD on request and on any change. Departing personnel access is revoked within 1 hour of departure and their secrets/devices returned; a confirmation is logged. The operator promptly removes any staff member S&ITD reasonably objects to on security/integrity grounds.


14. Conflict of Interest & Anti-Corruption

The procurement and operation of SITP align with the Sindh Public Procurement Regulatory Authority (PEPRA) framework and prevailing federal anti-corruption law. The operator and its staff must avoid any actual or apparent conflict of interest and must not offer, promise, give, or accept any improper advantage in connection with the contract.


15. Dispute Resolution

15.1 Governing law

The contract is governed by the laws of Pakistan, and specifically the law applicable to the Government of Sindh, including Sindh procurement and contract law. The courts of Karachi, Sindh have jurisdiction, subject to the arbitration clause below.

15.2 Resolution ladder

Disputes are resolved in stages to preserve the working relationship and continuity of service:

  1. Operational resolution — the program office resolves day-to-day issues within the performance-review cadence (§16).
  2. Negotiation — formal written notice; senior representatives of S&ITD and MAAHIR meet within 15 business days to resolve.
  3. Mediation — a neutral mediator (agreed or appointed) attempts resolution within 30 business days.
  4. Arbitration — unresolved disputes go to binding arbitration in Karachi under the Arbitration Act 1940 (as amended) / the Recognition and Enforcement (Arbitration Agreements and Foreign Arbitral Awards) Act 2011 framework, with a sole arbitrator (or three-member tribunal for material disputes) agreed by the parties or appointed per the rules.

PPP-V-1101Continuity during dispute. The operator must continue to operate the platform and meet SLAs during any dispute, and S&ITD must continue to pay undisputed invoices, until the dispute is resolved or the transition plan is invoked. A dispute never justifies service interruption. Escrow release and the transition plan (§11) remain available regardless of dispute status.


16. Performance Review Cadence

Performance against the SLAs (§4) and KPIs (§5) is reviewed on a fixed cadence by the joint S&ITD–MAAHIR program office, with escalation paths to S&ITD leadership.

Cadence Forum Attendees Scope Output
Weekly Operations stand-up S&ITD product owner; MAAHIR delivery + ops leads Incidents, P1/P2 status, SLA at-risk, freshness/staleness, blockers. Action log.
Monthly Performance review + Secretary S&ITD (or nominee); MAAHIR account lead Vendor scorecard (§5): availability, MTTR, CSAT, deflection, AI cost, security incidents; service-credit calculation; KT/escrow status. Scorecard sign-off; credits raised; remediation plans.
Quarterly Strategic review + SACM/Minister's office (notified); MAAHIR leadership Trend analysis; KPI targets vs actuals; risk register (roadmap §12); roadmap progress; AI spend/cost-efficiency; audit completeness; renewal-readiness signal. Quarterly report; go-forward decisions.
Annually Contract & renewal review S&ITD executive; MAAHIR executive Full-year performance; contract change/variation review; renewal/exit decision; ADP budget alignment. Renewal, restructure, or exit decision; triggers M10 if exit.
Per-release Release review Program office Release governance (/specs/en/14-roadmap-release/ §13.3): escrow deposit verified, IP assignment recorded, KT delivered. Release sign-off; milestone payment release (§6).

17. Contract Change Management

Changes to scope, schedule, cost, or any SLA/KPI target are controlled so that the safeguards in this document are never silently eroded.


18. Insurance & Liability

Item Requirement
Professional indemnity The operator maintains professional indemnity / errors-and-omissions insurance adequate to the contract value, covering IP infringement claims and professional negligence.
Cyber liability Cyber-liability insurance covering data breach and incident response costs, aligned to the data sensitivity of the platform (PII + sovereign data).
Public liability Standard public-liability cover for any on-site activity.
Proof & continuity Certificates of insurance are provided annually and on renewal; lapse of required cover is a breach.
Liability cap The operator's aggregate liability is capped per the contract for indirect losses, except for: breaches of §8 (IP), §9 (data), §13 (security), §14 (anti-corruption), fraud, wilful misconduct, and any liability that cannot be capped by law — these are uncapped or capped at a higher threshold.
Indemnities The operator indemnifies S&ITD against third-party IP infringement claims and against losses arising from the operator's security negligence or data misuse.
Survival The IP, data, security, confidentiality, and exit obligations survive termination and contract expiry (PPP-V-102).

19. Exit-Readiness Checklist

This checklist is the acceptance instrument for the readiness review (M10) and is rehearsed (dry-run) before any actual transition. A green checklist is the evidence — not the promise — that S&ITD can take the platform over. Status: required-pass; required-acceptable-with-plan.

# Check Owner Status
1 Escrow deposit for the current production release verified buildable by the agent (§7.1). S&ITD / escrow agent
2 Full, consistent database export + object storage + logs + analytics aggregates restorable to a S&ITD-controlled destination (§9.2). MAAHIR → S&ITD
3 All deliverable IP assigned and recorded (M08); Background-IP licences confirmed perpetual/usable (§8.2). S&ITD / MAAHIR
4 Documentation set, runbooks, and recorded code walkthroughs current and complete (§10). MAAHIR
5 Transition-KT delivered to S&ITD team and/or named successor; attendance & sign-off recorded. MAAHIR → S&ITD/successor
6 Reconstitution option chosen and resourced (in-house / new vendor / hybrid) (§11.2). S&ITD
7 Successor (if any) running platform in shadow/parallel; cutover plan rehearsed. MAAHIR + successor
8 Secrets register complete; live-secret rotation plan ready; access-revocation list prepared (§9.3). MAAHIR / S&ITD
9 Hosting continuity confirmed (Server4Sale handover or new host) with no scheduled outage. Server4Sale / S&ITD
10 AI/eval assets, prompt library, and engine-selection rationale handed over (§7.2, §10). MAAHIR
11 Third-party licence manifest complete; right to operate each component confirmed. MAAHIR
12 SLA/KPI performance history exported; service-credit/penalty position reconciled (§§4–5). Program office
13 Outstanding CRs, variations, and disputes documented (§§15, 17). Program office
14 DR/BCP verified on the destination (RPO ≤ 1 h, RTO ≤ 4 h) — NFR-AVAIL-004/005. MAAHIR / Server4Sale
15 Former-operator access fully revoked; no-copy attestation signed; secrets rotated (§9.3). S&ITD
16 Close-out & lessons learned archived (M11). S&ITD
17 Readiness decision formally recorded by S&ITD (M10 green). Secretary S&ITD

20. References

20.1 Internal documentation

20.2 External frameworks


End of document. Urdu and Sindhi parallel translations (ur.md, sd.md) follow this master structure exactly.