1. Interpret — In this stage, the AI reads 14-day climate forecast data (Open-Meteo, a free public API) and manually-entered bulletins from ICPAC, FEWS NET, or regional disaster-management offices. The output is a single structured hazard assessment: hazard type (drought or flood), severity level (low, watch, warning, emergency), and how confident the AI is in that assessment — with its reasoning written out explicitly.

2. Anticipate — The hazard assessment from stage one is reasoned over a pre-built table of 15 climate-to-health impact rules, drawn from sources like WHO guidance and Ethiopia's national malnutrition treatment protocol. Each rule encodes a relationship — e.g. "flood at warning level" links to "cholera outbreak risk" within a stated time window (e.g. "1 to 4 weeks"). The AI treats these rules as hard constraints: every anticipated health risk it produces must cite at least one matching rule, and its timing must fall inside that rule's stated range. Output that doesn't meet this bar is rejected outright, not softened or approximated.

3. Advise — The AI advisor (a retrieval-grounded chat assistant) answers protocol questions from health workers, disaster-management officers, or community members. It answers only from ingested source documents (WHO guidelines, Sphere standards, etc.), with an explicit citation naming the document and section for every claim. If there isn't a good enough match in the source material, it says so plainly — "not in my sources" — instead of generating a plausible-sounding but ungrounded answer.

4. Advise (Generate) — Once a region has a confirmed health situation, the AI can draft advisory text — a WhatsApp message, a radio script, or a formal memo — in both Somali and English. The AI writes only the narrative wording; it has no mechanism to produce a new risk level in this stage. Every advisory is a draft until a health officer or admin reviews and approves it.

Grounding, not freestyling: Stages 2 and 4 cannot output an unsupported claim. Anything not traceable to a vetted rule or confirmed data is rejected automatically, before a human ever sees it.

Structured validation: Every AI response is checked against a strict technical schema. Malformed or non-compliant output is retried once; if it still fails, it's routed to a human review queue rather than silently discarded or force-accepted.

Confidence + human-in-the-loop: If the AI reports low confidence, or assesses a situation as an emergency, that assessment is automatically withheld from public display until an administrator confirms it in the admin panel. Every advisory, regardless of confidence, stays a draft until a human approves it.

Full audit logging: Every AI call — the prompt it was given, the exact response, and the validation outcome — is written to an immutable log. That log is, in effect, the methodology annex: it's how a reviewer can verify the claims on this page against what the system actually did.

The system defines four roles: Admin (full access), Health Officer (can approve advisories for their assigned region), Disaster Risk Management (DRM) Officer, and Viewer (read-only).

Hazard data with low AI confidence or an emergency-level assessment is withheld from public display until an admin confirms it. AI-drafted advisories stay in draft status until a Health Officer (for their own region) or Admin (any region) approves them — both the approval timestamp and the approver are recorded.

Rainfall forecasts: Open-Meteo, a free public API, pulled automatically on a daily schedule.

Manually-entered data: ICPAC, FEWS NET, and regional disaster-management bulletins are currently entered by hand through the admin panel — a pragmatic bridge while direct automated integrations are built out.

Protocol documents: WHO guidelines, Sphere standards, and national health protocols are ingested as PDFs, chunked, and indexed so the AI advisor can retrieve and cite them.

In the interest of being straightforward: the system currently relies on a mix of automated (Open-Meteo) and manual (ICPAC/FEWS NET/DPPB) data entry — a direct automated integration with ICPAC has not yet been built.

The AI advisor's minimum-confidence threshold for answering (rather than refusing) is a starting estimate, not yet calibrated against real field queries — that calibration is itself a planned, fundable piece of work.

Authentication was only recently wired up. Role assignment (who becomes a Health Officer or Admin, and for which region) is currently done directly against the database rather than through a self-service admin UI — that's the next piece of access-control work.