Scenario-insights endpoint — LIGHT-tier AI narrative via the org's provider

Shipped

July 9, 2026 at 4:23 PM UTC

Author

Kamo

Commit

ce62abe

POST **************** takes a factual (estimated) scenario summary, calls MlosAiClient.text(orgId, prompt, LIGHT) through the org's configured AI provider, returns {narrative} (null when no provider — the Scenario Workspace keeps its deterministic insights). Server owns the prompt; AI only augments, never decides. First MLOS consumer of the tiered AI client.