A county health clinic requires kiosks that still guide residents when the WAN drops. How should that shape Claude integration?

County health clinic kiosks sit on variable WAN links and must still guide residents when connectivity drops. Designing offline-tolerant kiosk behavior into the Claude integration architecture is correct because infrastructure realities are requirements that shape solution design: cached clinic hours, appointment instructions, escalation paths, and graceful degradation must be planned up front, not bolted on after outages. Conceptually this is civic continuity—health guidance cannot freeze solely because the cloud path is unavailable, even when live Claude inference is the preferred mode.

Prioritizing real-time Claude API responses for kiosk navigation guidance fails by concept: it assumes the WAN is always up and leaves residents stranded when connectivity drops. Configuring centralized logging of Claude interactions for kiosk analytics dashboards fails as a resilience strategy—analytics do not help residents when the WAN is down, and logging alone provides no local fallback guidance. Enabling prompt caching to accelerate Claude-powered kiosk health responses fails because caching speeds up API calls but does not keep the kiosk useful when the WAN is unavailable; residents still need local fallback content, not faster cloud calls.

Exam caveat: infrastructure-requirement questions reward baking constraints into architecture, not assuming perfect networks. Do not pick centralized-logging distractors that ignore offline kiosk needs. Preserve clinic-kiosk municipal framing. Operational check: define offline UX for hours/directions/urgent contacts, cache approved content on the kiosk, detect WAN loss, avoid treating centralized logging as a substitute for offline capability, and test kiosk flows with the WAN intentionally disabled.