Severance terms, notification method, and a 0–100 treatment score from public details. Tap any card for the receipts.
Sources. Layoff events scraped from public reporting (LayoffHedge, press releases). Each event re-read by a language model that extracts cited reason, likely reason, severance terms, notification method, and a treatment score (0–100) from public details.
Treatment score. A blended rubric: how people were notified, severance weeks, healthcare continuation, notice period, tone of the CEO letter. Higher = more humane. Score is mechanical — same inputs always produce the same number.