• Cross-government “platform of platforms” providing: identity choice, payments, notifications, forms, cases/workflows, metrics, and a shared UI kit.
• Initial migration of 3–5 high-impact public services (licensing, benefits, permits) to prove speed & reliability.
• Standards & playbooks, not mandates; opt-in components with clear SLOs.

Delivery is slow and costly because teams rebuild the same primitives, policy and product are split, and there is little visibility of outcomes. Citizens encounter paper-era flows ported to the web.

• TTR (time-to-result): median end-to-end completion time per service ↓ 40% in 12 months.
• QoS/SLO adherence: shared components ≥ 99.9% monthly availability.
• Cost-to-serve: unit cost per transaction ↓ 30% across migrated services.
• Transparency: 100% services publish outcome dashboards & changelogs.
• User satisfaction: CSAT ≥ 80% for migrated services.

• Product trio per service: Product Manager, Tech Lead, Designer/UXR.
• Platform guilds: Identity, Payments, Notifications, Forms, Case/Workflow, Observability.
• Policy partners embedded weekly; delivery managers; content designers.
• Assurance: Security engineer, Privacy officer, SRE on-call rotation.

1. 0–6 weeks: lift a minimal “service kernel” (UI kit, registry, auth proxy, metrics). Pick the first 2 services.
2. 6–16 weeks: migrate MVP flows; publish dashboards; close the loop on support & feedback.
3. 4–9 months: expand to 5+ services; harden platform; publish playbooks; external developer sandbox.

• Service Kernel: design system + layout primitives; accessible by default.
• Identity Choice: passwordless + eID-compatible; supports “prove X, not identity”.
• Payments: card/open banking rails; refunds; split remittance; auditable receipts.
• Notifications: email/SMS/inbox with per-user prefs & GOV routing policies.
• Forms & Content: versioned schema + content model; inline validation; translations.
• Case/Workflow: event-driven state machine; role-based tasks; SLAs.
• Observability: outcomes ledger, audit trails, public dashboards.

• Citizens & businesses completing services.
• Frontline staff & case workers processing cases.
• Policy & delivery managers monitoring outcomes.
• Auditors, journalists, and the public viewing dashboards.

• Citizen Portal (task list, messages, status, appointments).
• Staff Workspace (queues, SLAs, triage, case timeline).
• Service Owner Console (metrics, config, releases, A/Bs).
• Public Dashboard (outcomes & service levels, privacy-safe).

• REST/JSON + async webhooks; OAuth2 / mTLS for staff apps; fine-grained scopes.
• Event topics: case.created, case.state_changed, payment.settled, notification.sent.
• Auditable IDs, idempotency keys, and immutable logs (append-only).

• Event-sourced core; services communicate via well-defined events and stable contracts.
• Thin platform, thick product teams: platform offers paved roads, not boulders.
• Privacy-by-design: minimum data, differential audit views, redaction tooling.
• Zero-trust boundaries; per-service data domains; shared observability mesh.

1. Pick 2 services with clear pain & measurable demand; publish baseline metrics.
2. Build the kernel + one paved road path; ship end-to-end in weeks, not months.
3. Shadow-run with frontline staff; fold feedback; remove paper detours.
4. Publish playbooks, dashboards, and a public changelog for each release.
5. Scale to additional services; formalise SLOs; open API sandbox.

• Departments & agencies owning services; Cabinet Office digital leadership.
• Treasury (funding model & outcome accounting), NAO (assurance), ICO (privacy).
• Citizen panels, civil society groups, and suppliers/partners.

• Design system; Forms service; Case/Workflow engine; Identity proxy; Payments; Notifications; Observability & outcomes ledger; API gateway.

• Seed as a Transformation Fund line item.
• Then usage-based internal pricing per transaction to sustain and signal demand.
• Outcome bounties for teams beating baseline TTR & cost-to-serve.