Analytics and cohorts
One score for how your AI treats people, and the cohorts behind it.
The Behavior Score is a 0–100 composite of retention, conversation depth, response to proactive outreach, abandonment and sentiment trend. Cohort views show retention and average mood by month of first contact.
Behavior Score · last 7 days
- retention_deltaweight 30%
- conversation_depthweight 25%
- proactive_response_rateweight 20%
- abandonment_inverseweight 15%
- sentiment_trendweight 10%
What goes into the score
Each component is normalised to 0–100 against a baseline before weighting.
- Retention change: 30%
- Conversation depth: 25%
- Proactive response rate: 20%
- Abandonment, inverted: 15%
- Sentiment trend: 10%
- A breakdown per component comes with every score
Fetch the score
score = client.behavior_score(window_days=7)
print(score["behavior_score"]) # 0 to 100
print(score["behavior_score_breakdown"]) # per componentCohorts by first month, and engines by latency
Cohort analytics group users by the month they first arrived and show retention and average mood at set intervals. Experiment comparisons and per-engine latency metrics sit alongside, so you can see what changed and what it cost.
- Retention and average mood at N days for each monthly cohort
- Variant comparison with lift and p-value
- Latency metrics for each engine in the pipeline
- A ring buffer of recent pipeline runs for debugging
Users · live
- user_42↓ 3 weekselevated
- user_17steadynone
- user_88↓ todaycritical
- user_05↑ 1 weeknone
Analytics questions
Something else? Ask us.
Works well with
Anomaly detection
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.
ExploreExperiments
Define an experiment with two variants, each a set of engine settings, and Humane assigns users and applies the right overlay on every call. Results compare the variants with lift and a two-proportion z-test.
ExploreUser state across sessions
Humane keeps a running state for every end user, from mood and energy to trust and grudge, each with its own behaviour over time. Your model finally knows whether it is talking to someone on a good day or at the end of a bad week.
ExploreEvery 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.
