Journeys & pipelines

Both add structure beyond a single agent on a single call, but they solve different problems:

• Journey — a multi-touch sequence per contact, over time: call, wait, SMS, call again.
• Pipeline — multiple agents on one call, handing off mid-conversation.

Journeys (multi-touch per contact)

A journey is configured on a campaign — it sequences steps that run per contact as calls complete. Step types are call, wait, and sms.

{
  "max_touches": 4,
  "exit_on_success": true,
  "steps": [
    { "type": "call", "label": "Initial contact" },
    { "type": "wait", "hours": 24, "label": "Wait 24h" },
    { "type": "sms",  "body": "Hi {name}, we tried calling. Reply to schedule.", "label": "SMS follow-up" },
    { "type": "wait", "hours": 48, "label": "Wait 48h" },
    { "type": "call", "label": "Final attempt" }
  ]
}

When a call ends, the next step is computed automatically: a wait schedules the next touch with a delay; exit_on_success ends the journey on a successful outcome; reaching max_touches ends it too.

configure_journey(campaign_id="CAMPAIGN_ID", steps=[...], max_touches=4, exit_on_success=true)
get_journey(campaign_id="CAMPAIGN_ID")

Pipelines (multi-agent on a single call)

A pipeline chains multiple agents within one call — one agent can hand off to another mid-conversation (for example, a triage agent transfers to a sales agent). Build them from templates such as sales pipeline, inbound routing, and appointment setting.

create_pipeline(name="Triage → Sales", ...)
create_smart_pipeline(...)        # cost-weighted build with a preview
list_pipelines()
get_pipeline(pipeline_id="...")
activate_pipeline(pipeline_id="...")
delete_pipeline(pipeline_id="...")

create_smart_pipeline produces a cost-weighted preview before you commit, then you activate_pipeline to make it live.

Next

• Agents — the agents a journey or pipeline orchestrates.
• Campaigns — where journeys are configured.
• Calls & transcripts — read the outcomes.