Understand usersObserve

Anomaly detection

Catch the slow slide that no single message shows.

A scheduled sweep compares the last day with a 30-day baseline for each user and each tenant. When a metric moves two or more standard deviations, Humane records an anomaly, rates its severity and fires anomaly.detected.

Anomalies · user_42

Sentiment · daily

Warning

sentiment_drop · z = -3.4 · dashed: baseline band

info · 2–3σwarning · 3–4σcritical · 4σ+
Illustration of an anomaly against a user's own baseline
How it works

Each user judged against their own history

Some users are always quiet and others always upbeat, so a global threshold misses the change that matters. Humane computes a z-score for the last 24 hours against the last 30 days. Users without enough history fall back to a neutral reference, so a real collapse in sentiment still registers.

  • Four metrics: sentiment drop, engagement drop, latency spike and error-rate spike
  • Severity: info at 2–3σ, warning at 3–4σ, critical at 4σ and above
  • Repeats of the same anomaly are suppressed within a cooldown
  • Critical anomalies also email an operator

Journey · user_42

  1. 09:12message

    “honestly I don’t know why I keep trying”

  2. 09:12safety.flagged

    risk elevated · flags: self_harm

  3. 09:13takeover.started

    Operator joined the conversation

  4. 09:41takeover.released

    Handed back to the AI

  5. Next dayproactive.mood_drop

    Suggested: an empathetic check-in

Illustration of a user's journey timeline
Review

List, review and resolve

python
for a in client.list_anomalies(since_hours=168):
    print(a["severity"], a["kind"], a["summary"])

client.resolve_anomaly("anom_01HQ...", note="Checked in with the user")

Fits how your team works

  • Anomaly settings tunable per tenant
  • anomaly.detected webhook for your on-call tools
  • Resolve with a note, so the record shows what was done
  • Old anomalies pruned on a schedule

Anomaly questions

Something else? Ask us.

Every AI should know who it's talking to.

The safety and human-oversight layer for emotional AI: user state across sessions, risk alerts, safety gates and human takeover, with every decision explained.