DIREKT Phase 1B Heuristic Review

Review type: structured expert review
Date: 2026-07-14
Scope: static interactive prototype
Evidence class: SYNTHETIC-TEST
Limitation: this is not representative Zambia participant research

1. Review method

The prototype was reviewed against:

  • the primary design.md principles;
  • the Phase 1A secondary-research baseline;
  • the screen inventory;
  • accessibility and low-connectivity requirements;
  • trust, privacy and verification-state specifications;
  • common usability heuristics: system status, real-world language, user control, consistency, error prevention, recognition, efficiency, recovery and help.

Each finding is classified as:

  • PASS — design requirement is represented sufficiently for Phase 2;
  • CORRECTED — issue found and addressed in the prototype or specification;
  • ACCEPTED LIMITATION — requires native, backend or real-user testing later;
  • BLOCKER — must be resolved before Phase 2.

2. Trust comprehension

Test Result Evidence
Generic verification badge avoided PASS Separate claim cards across customer profiles
Claim scope is visible PASS Each detail dialog includes What this means
Limitations are visible PASS Each detail dialog includes What DIREKT did not check
Dates and expiry are visible PASS Trust detail dialogs and provider timeline
Pending is not presented as approved PASS BrightFix qualification and scheduled field visit
Expired state degrades public meaning PASS BrightFix identity example
Registration is not competence PASS Business registration limitation copy
Experience is not formal qualification PASS Kafue Road Auto Care example
Field visit is distinct from location evidence PASS Mwamba Water Works example
Customer understands wording in practice ACCEPTED LIMITATION Requires real-user testing in controlled pilot

3. Location and operating model

Test Result Evidence
Fixed premises explained PASS ZedSpark and BrightFix profiles
Mobile provider privacy explained PASS Mwamba profile and operating-model screen
Hybrid operation explained PASS Kafue Road Auto Care
Map has list equivalent PASS CU-005 and CU-006
Location permission is optional PASS Discover copy and denied-state simulator
Manual area and landmark represented PASS Discover and enquiry screens
Plus Code concept represented PASS Denied-state and area-selection copy
Exact private location protected PASS Profile and provider operating-model copy
Real map coverage and geocoding accuracy ACCEPTED LIMITATION Requires provider selection and field test

4. Customer journey

Test Result Evidence
Value proposition is concise PASS SH-002 welcome
Category discovery is understandable PASS CU-001 category cards
Results allow provider-model comparison PASS CU-005 cards
Profile prioritizes trust over promotion PASS CU-008 hierarchy
Enquiry purpose is explained PASS CU-011
Contact disclosure requires consent PASS CU-012
External actions are safe in prototype PASS No real phone or WhatsApp link
Review eligibility concept represented PASS Profile review section
Account, real authentication and support completion ACCEPTED LIMITATION Phase 2+ implementation

5. Provider journey

Test Result Evidence
Onboarding is resumable PASS PR-002 checklist and draft copy
Informal provider has honest pathway PASS PR-003
Evidence request is claim-specific PASS PR-010
Optional business evidence is not disguised as required PASS PR-010
Interrupted upload is recoverable PASS PR-011
Rejection reason is actionable PASS PR-014
Rejected evidence remains private PASS PR-014 copy
Timeline makes lifecycle visible PASS PR-013
Real upload, image quality and accessibility ACCEPTED LIMITATION Phase 2 native implementation

6. Operations journey

Test Result Evidence
Queue is check-specific PASS OP-002
Payment tier is excluded from priority PASS OP-002 commercial-separation notice
Evidence view is visibly restricted PASS OP-004
Synthetic evidence cannot be mistaken for real PASS Explicit fictional labels and redaction
Reviewer outcomes are structured PASS OP-005
Action-required guidance is supported PASS Reason code and provider-facing note
Audit events are append-only in concept PASS OP-016
Real role authorization and evidence access ACCEPTED LIMITATION Phase 2–4 implementation/testing

7. State and recovery review

State Result Notes
Loading PASS Skeleton content rather than unexplained spinner
Slow network PASS Text-first explanation
Offline PASS Cached-state limitation and no trust upgrade
No results PASS Widen area/remove filters; no unreviewed filler
Location denied PASS Manual area fallback
Recoverable error PASS Last safe data and explicit retry
Upload interruption PASS Draft safety, retry and acknowledgment language
Session expiry ACCEPTED LIMITATION Requires authentication implementation

8. Accessibility review

Requirement Result Notes
Semantic landmarks/headings PASS Header, aside, main, nav and sections
Keyboard navigation PASS BY INSPECTION Native buttons, links, inputs and dialogs
Visible focus PASS Global :focus-visible treatment
Colour-independent status PASS Text and icon accompany colour
Minimum target intent PASS BY DESIGN Main controls use approximately 42–50 px targets
Map list equivalent PASS Dedicated result list
Reduced motion PASS Media query present
Screen-reader labels PASS BY INSPECTION Controls and map pins include labels
TalkBack and 200% Android font scale ACCEPTED LIMITATION Native-device testing required

9. Privacy and public-repository review

Check Result
Real identities absent PASS
Real documents absent PASS
Real phone numbers absent PASS
Precise private coordinates absent PASS
Secrets or credentials absent PASS
Third-party tracking absent PASS
External data submission absent PASS
Local review notes disclosed PASS
Prototype limitation banner present PASS

10. Corrections made during review

  1. Pages generator updated to copy the static prototype into the published site.
  2. Landing page generator updated with a direct prototype launch link.
  3. Screen inventory updated to distinguish prototype coverage from implementation.
  4. Fieldwork requirements moved from Phase 1A blocker to later validation gates.
  5. Informal-provider pathway given the same profile structure while missing claims remain explicit.
  6. Fixed, mobile and hybrid examples added.
  7. No-results copy explicitly prevents filling results with unreviewed profiles.
  8. Operations queue explicitly excludes subscription status from prioritization.
  9. External call/WhatsApp actions converted to non-functional safety simulations.
  10. Feedback notes restricted to local browser storage.

11. Accepted limitations entering Phase 2

  • no representative participant usability study;
  • no native Android rendering or TalkBack run;
  • no real network, offline database or upload implementation;
  • no authentication or authorization;
  • no API, database, regulator, map, payment or notification integration;
  • no real evidence storage;
  • no measured provider/customer willingness;
  • no field-operation cost evidence.

These limitations are already assigned to later phases and do not justify keeping technical scaffolding blocked.

12. Review conclusion

Result: PASS WITH ACCEPTED LIMITATIONS

No interaction-design blocker was found that should prevent Phase 2 technical scaffolding. Native implementation must preserve the check-specific trust language, location privacy, offline semantics, synthetic-test discipline and commercial/trust separation represented here.