Configure follow-up pings

Goal: automatically re-engage leads who haven't replied, in timed steps.

A ping configuration belongs to an agent and holds an ordered chain of steps; each step fires after a delay and sends a message.

curl -s https://api.sance.ai/v1/ping-configs/ \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{
        "agent_id": 3,
        "is_enabled": true,
        "config": {
          "steps": [
            { "step_number": 1, "after_minutes": 1440,
              "source": { "text": "Just checking in — still interested?" },
              "dialogue_init_msg_mode": "only_with_replies" },
            { "step_number": 2, "after_minutes": 4320,
              "source": { "text": "Last nudge from us — happy to help when you're ready." },
              "dialogue_init_msg_mode": "only_with_replies" }
          ]
        }
      }'

Each step's after_minutes is measured from the last activity; dialogue_init_msg_mode targets all dialogues, only_with_replies, or only_without_replies; an optional new_lead_stage_id moves the lead when the step fires. source is a message source (text, prompt_id, prepared_message_id, or a WABA template). Retrieve an agent's configs with GET /v1/ping-configs/for-agent/{agent_id}.

See: POST /v1/ping-configs/, GET /v1/ping-configs/for-agent/{agent_id}, PUT /v1/ping-configs/{id}.