Read-only sandbox demonstration guide
How Health Bridge Works
From coverage and care to an accountable human decision — a Uganda-focused foundation with clear boundaries.
INSURANCE AND PEOPLEOne ecosystem. Shared progress.
The connected foundation
One claim. Six accountable stages.
This public guide contains no tenant records. The authenticated sandbox uses saved synthetic records and membership-based access; the separate offline demo uses seeded browser-local records. Their capabilities are not interchangeable.
- 01 · Saved sandbox configuration
Membership & eligibility
An insurer configures fictional policies, members and coverage. A provider verifies policy dates, premium status and available benefits. Public role selection is not registration and never grants access.
- 02 · Prior authorisation: offline only
Provider encounter & prior authorisation
The provider records the encounter. Prior authorisation can be explored in the offline demo; a connected authenticated request-and-decision experience remains a gap.
- 03 · Persisted synthetic claim and documents
Itemized claim & evidence
Submit services and quantities against insurer tariffs. In the authenticated sandbox, private supporting documents and insurer-required checklists feed the readiness checks.
- 04 · Rules-based support, not autonomous AI
Review, orchestration & human decision
Rules-based checks consolidate sources, timestamps, financial and documentation flags, reasons and unknowns. An assessor recommends; a designated independent insurer decision-maker decides when configured. The existing two-step path remains the default.
- 05 · Human decisions; status-only payments
Appeal & simulated settlement
A hospital or patient can appeal a rejected claim. An independent reviewer can uphold or return it to review, not approve or pay directly. Approved claims move through labeled simulated settlement states; no funds move.
- 06 · Permission-scoped visibility
Status, regulatory & audit oversight
In-app updates expose permitted status. Patients see their own records; observers receive only explicitly authorised aggregates. Protected decision history and hash-chained audit metadata support accountable oversight.
Presenter’s path
A repeatable fictional scenario.
Sarah Nakato is already seeded. Reset the offline scenario, then follow the provider → assessor → insurer → patient → observer prompts. The UGX 85,000 lab claim demonstrates the earlier two-step review and oversight path. The role selector also opens the broker perspective; platform administration stays in the authorised workspace.
Use the offline member and prior-authorisation views for a separate illustration. The guided scenario uses document-name placeholders, not uploads. Reset affects only fictional browser-local data, never authenticated claims.
Evidence, not a launch claim
Tested logic. Remaining pilot gates.
Rolled-back database tests cover single-decision and fallback paths, evidence checks, appeals, orchestration, simulated settlement and denied role/tenant access. They do not substitute for a complete real-account rehearsal or independent security review.
Real insurer, assessor, hospital and patient browser journeys remain unverified. Autonomous AI, external notifications, payment integrations and Uganda-specific legal, clinical and regulatory validation are not delivered. Mandatory privileged MFA, rate limits and malware scanning remain pilot gaps.
Final decisions and oversight stay with authorised people. Public role selection never grants access.