diff --git a/README.md b/README.md
index 72ca71e..225b9e9 100644
--- a/README.md
+++ b/README.md
@@ -103,6 +103,13 @@ with eval_data as (
 select * from {{ dbt_ml.detect_anomalies(ref('model'), 'eval_data', threshold) }}
 ```
 
+If using a forecasting model, you can use the forecast macro in the same way. Here we are forecasting 30 units ahead with 80% confidence.
+```sql
+# forecast_model.sql
+
+select * from {{ dbt_ml.forecast(ref('model'), 30, 0.8) }}
+```
+
 ### Tuning hyperparameters
 BigQuery ML supports tuning model hyperparameters<sup>[2]</sup>, as does `dbt_ml`. In order to specify which hyperparameters to tune, and which parameterspace to use, one can use the `dbt_ml.hparam_candidates` and `dbt_ml.hparam_range` macros that map to the corresponding BigQuery ML methods.
 
diff --git a/macros/forecast.sql b/macros/forecast.sql
new file mode 100644
index 0000000..9a03598
--- /dev/null
+++ b/macros/forecast.sql
@@ -0,0 +1,10 @@
+{% macro forecast(relation, horizon = 30, confidence_level = 0.8) %}
+    ml.forecast(
+      model {{ relation }},
+      struct(
+        {{ horizon }} AS horizon,
+        {{ confidence_level }} AS confidence_level
+      )
+    )
+{% endmacro %}
+