Semantic Layer for AI & BI

A robust semantic layer that works with your stack

Models and metrics live in version-controlled YAML. Your coding agents work with Supersimple's CLI and agent skills to build, validate, and test changes.

View docs ↗

SYNCS WITH YOUR EXISTING DEFINITIONS

dbt

Apache Ossie

Databricks Metric Views

Snowflake Semantic Views

delivery.yaml

model:
  model_id: delivery
  name: Delivery
  table: operations.deliveries
  primary_key: [delivery_id]

  relations:
    organization:
      type: hasOne
      model_id: organization
      join_key: organization_id

metric:
  metric_id: on_time_delivery_rate
  name: On-time delivery rate
  model_id: delivery
  aggregation:
    type: ratio
    numerator: on_time_delivered_orders
    denominator: delivered_orders

model:
  model_id: delivery
  name: Delivery
  table: operations.deliveries
  primary_key: [delivery_id]

  relations:
    organization:
      type: hasOne
      model_id: organization
      join_key: organization_id

metric:
  metric_id: on_time_delivery_rate
  name: On-time delivery rate
  model_id: delivery
  aggregation:
    type: ratio
    numerator: on_time_delivered_orders
    denominator: delivered_orders

Delivery

Priority (on Customer) is true

Metric: On-time delivery rate by Order Created Week

89.7%

On-time delivery

-0.55%

vs last week

Inbox for improving your data models and AI context.

Supersimple continuously analyzes AI agent traces against your data models and provided context.

The Inbox proactively surfaces problems with missing data, ambiguous definitions, conflicting information, missing context, and unclear business terminology.

Inbox

Items that need your attention or action

Clarify how on-time delivery is defined

Users are asking about on-time delivery, but the model does not define the threshold or timezone to use.

Add cost allocation data for gross margin

The AI can calculate revenue by product line, but the required cost allocation source is not available.

Define what makes an operation “priority”

Several questions use the term priority operation. The AI currently has to infer its meaning from context.

Governed semantic layer for both AI and humans.