ShedBoxAI for SaaS

data_sources:
  events:
    type: csv
    path: "user_events.csv"

processing:
  contextual_filtering:
    events:
      - field: "event_type"
        condition: "in:signup,purchase,churn"
        new_name: "key_events"

  advanced_operations:
    user_metrics:
      source: "key_events"
      group_by: "user_id"
      aggregate:
        event_count: "COUNT(*)"
        first_event: "MIN(timestamp)"
        last_event: "MAX(timestamp)"
      sort: "-event_count"

output:
  type: file
  path: "user_behavior.json"
  format: json

data_sources:
  users:
    type: csv
    path: "user_activity.csv"

ai_interface:
  model:
    type: rest
    url: "https://api.anthropic.com/v1/messages"
    method: POST
    headers:
      x-api-key: "${ANTHROPIC_API_KEY}"
      Content-Type: "application/json"
    options:
      model: "claude-sonnet-4-20250514"

  prompts:
    churn_risk:
      system: "You are a customer success analyst."
      user_template: |
        Analyze these users and identify churn risk:

        {% for user in users %}
        - {{ user.name }}: Last login {{ user.last_login }}, {{ user.feature_usage }} features used, {{ user.support_tickets }} tickets
        {% endfor %}

        Categorize each user's churn risk (low/medium/high) and explain why.

output:
  type: file
  path: "churn_risk_report.md"
  format: json

data_sources:
  subscriptions:
    type: rest
    url: "https://api.stripe.com/v1/subscriptions"
    headers:
      Authorization: "Bearer ${STRIPE_SECRET_KEY}"
    response_path: "data"

processing:
  contextual_filtering:
    subscriptions:
      - field: "status"
        condition: "active"
        new_name: "active_subs"

  advanced_operations:
    mrr_by_plan:
      source: "active_subs"
      group_by: "plan.id"
      aggregate:
        subscriber_count: "COUNT(*)"
        mrr: "SUM(plan.amount)"
      sort: "-mrr"

output:
  type: file
  path: "mrr_metrics.json"
  format: json