Diagnosing 14 systemic failures in a Saudi healthcare-licensing platform — and shipping the rebuild brief.

I led an OOUX/ORCA diagnostic of Seha, the Saudi healthcare-licensing platform, and codified 14 systemic failure patterns surfaced through interviews with hospital medical directors and institutional operators. Output: an evidence-backed rebuild brief, a 6-screen WCAG-AA, RTL-first prototype, and a measurement plan aligned to ministry review.

— who

Seha · Saudi healthcare-licensing platform for institutional operators

— what

OOUX · ORCA · Stakeholder Research · WCAG AA · RTL Design · Prototyping

— result

14 systemic failure patterns documented · 6-screen rebuild prototype · WCAG AA at 4.5:1 floor

— scope

Diagnostic-to-prototype consulting engagement · stakeholder research, OOUX/ORCA rebuild, 6-screen Bet 1 prototype, WCAG AA + RTL compliance, measurement plan aligned to ministry review

Seha Bet 1 dashboard — dark sidebar, IBM Plex Mono numerals, RTL layout

— outcomes

systemic failure patterns documented

screen Bet 1 prototype shipped

4.5:1

WCAG AA floor verified across every color pair

— problem

the platform was built around the licensing process — not the people running hospitals.

Saudi institutional operators — university medical cities, hospital networks, hybrid academic-clinical institutions — depend on Seha to add services, staff, and facility types as their operations evolve. By the time the diagnostic began, those operators were stuck. Service requests sat for two months and longer with no status update. Permission matrices silently rejected legitimate combinations: a medical city could not add vaccination services because the facility type and the service type were incompatible in the underlying model. Operators couldn't add new medical staff because their roles didn't exist in the permission system at all. The platform's scope had outgrown its information model — and the cost was operational paralysis at the institutions Saudi healthcare depends on most.

— users

designing for medical directors, not licensing officers.

The platform's real users aren't licensing officers — they're medical directors, compliance leads, and facility administrators at the largest institutional operators in Saudi Arabia. Every blocked request is a clinical capability that doesn't reach patients. The stakes are operational, not procedural.

Stakeholder interviews surfaced fourteen distinct failure patterns (P1–P14) — recurring, evidence-backed, traceable to specific moments in the workflow. One representative case: the medical director of a major university medical city couldn't add vaccination services to her facility, because the platform's permission matrix didn't allow that facility type to host that service type — even though the institution had been delivering that service for years under a different code. She also couldn't add new medical staff, because the role she needed didn't exist in the permission model at all. Service requests she submitted sat in queue for two months with no status visibility and no way to escalate.

The signal in the research was sharp: the failures weren't in the UI. They were in the object model. Permission matrices were defined per facility type × service type without expressing the actual real-world relationships between facilities, services, practitioners, licenses, and authorities. The implication: this needed an OOUX rebuild, not a UI refresh.

— role

OOUX lead and prototype designer.

I led the diagnostic and the prototype build end-to-end — stakeholder interviews through ORCA mapping through high-fidelity screens. Specific ownership: stakeholder research with ten medical directors and institutional operators; ORCA process (Objects, Relationships, Calls-to-action, Attributes); Nested Object Matrix; CTA Matrix per role; wireframes; the 6-screen Bet 1 prototype; WCAG and RTL compliance review. Worked alongside platform stakeholders on ministry-alignment framing and accessibility validation. The work was reviewed by platform and ministry stakeholders.

— process

the mandate: diagnose what was broken, then propose a rebuild that survives ministry review.

Saudi healthcare governance has multiple regulatory authorities — Ministry of Health, Royal Commission, and others — with overlapping jurisdictions. The rebuild brief had to do three things at once: be defensible to operators dealing with the platform daily, survive ministry-level scrutiny on regulatory and compliance grounds, and be implementable in phases without breaking active licensing workflows already in production.

— phase one — stakeholder research

Ten interviews with medical directors, facility administrators, and licensing leads at institutional operators across Saudi Arabia. Sessions ran semi-structured: workflow walkthrough, then incident reconstruction (the last failed request, the last suspended service, the last role they couldn't add). Coded outputs into fourteen systemic failure patterns — each with a triggering condition, a downstream cost, and the operator quote that anchored it. The single most important finding came in the first three interviews: every reported "UI bug" turned out to be the surface symptom of a missing or wrongly-modeled relationship in the underlying data. That reframed the entire engagement.

— phase two — OOUX/ORCA rebuild

Object discovery against the corpus of interview transcripts: Facility, Service, Practitioner, License, Request, Authority. Mapped the Nested Object Matrix — every detail page in the platform expressed as columns, every nested object as rows, every cell as a navigation path plus a content section. Built the CTA Matrix per role: what does a Medical Director do here, what does a Compliance Officer do here, what does a Licensing Reviewer do here. Defined the KPI tree — what "good" looks like for each operator persona, traceable from session-level metrics back up to ministry-defined regulatory KPIs. The model surfaced two structural moves the original platform had missed: (a) a Facility can host multiple Service types simultaneously through a permission lattice rather than a flat matrix; (b) a Practitioner exists independently of any single Facility and can be associated with multiple institutions over a career.

— phase three — Bet 1 prototype

Six screens, prioritized by failure frequency in the research: Add Service, Add Staff, Request Status, Renew License, Notification Center, Dashboard. Lean.sa-derived visual language. IBM Plex Sans for body, IBM Plex Mono for all numerals, IDs, and stats. Radix UI component model. RTL-first using logical properties throughout (padding-inline-start, not padding-left). WCAG AA verified at the floor: primary text 17.5:1, body 9.4:1, muted 5.9:1, accent 5.8:1, and every status color pair pre-checked above the 4.5:1 line. Full keyboard navigation, skip link, ARIA roles on tables (role="row", role="columnheader"), role="progressbar" with aria-valuenow/min/max, and live region announcements on every navigation and filter change.

— evidence

ORCA map for Seha — Object/Relationship/CTA/Attribute map.

Nested Object Matrix for Seha — Nested Object Matrix.

CTA Matrix per role for Seha — CTA Matrix per role.

Add Service flow wireframe — Add Service flow — highest-stakes workflow per the research.

Bet 1 dashboard for Seha — Bet 1 dashboard — RTL, dark sidebar, IBM Plex Mono numerals.

WCAG AA contrast panel verifying color pairs — Verified WCAG AA contrast across all UI color pairs.

— bottom line

Healthcare licensing platform designed for Saudi Arabia's institutional operators · OOUX/ORCA pipeline run end-to-end from stakeholder brief through ship · defensible information architecture grounded in an object model shared by regulator and operator.

— previous

← Diagnosing 8 systemic UX failures in an FDA-regulated anemia management platform.

— next

Designing the foundation of an AI-native analytics canvas at Fusedash. →

— next step

Want a measurement plan like this for your next quarter?