Skip to main content
The plugin ships with a visual configuration panel — the model maker — that allows the organisation’s admin to shape their own Kai Affective Intelligence experience without writing code. This is the B2B product’s key differentiator: Kai’s intelligence is configurable, not one-size-fits-all.

​
Configuration options

Filter by team, department, age range, tenure, or custom tag. Define which users contribute to the aggregated signals displayed on your dashboard.
Show moods only, or include psychological states and topics. Control which of the four dashboard layers are active for your organisation.
Set minimum group size before data is surfaced. Default: 10. If the cohort matching your filters falls below this threshold, signals are suppressed to protect individual privacy.
Configure Kai’s language for corporate wellness, clinical, or community contexts. Available values: community, corporate, clinical.
Choose which Kai Brain skill cards are active for this cohort. Available on Signal and Brain tiers.
Set the distress delta that triggers a team notification. Default: 2.0 multiplier above baseline.
Weekly PDF digest, Slack summary, or email report on schedule.

​
Saving configuration

When an admin saves settings, the plugin validates and sends them to the API: 
ON config_saved(admin_user, settings):
  VALIDATE settings.anon_floor >= 10
  CALL POST /configure with settings
  IF response.cohort_size < 10:
    WARN: cohort too small to show data
  STORE config_id, REFRESH dashboard
If cohort_size is below your anonymisation floor after saving, the dashboard will warn that the cohort is too small to show data. Expand your filters or increase the monitored user base before signals can be surfaced.

​
API reference

Configuration is persisted via POST /configure. See the endpoint documentation for the full parameter schema and response fields.