diff --git a/README.md b/README.md
index 12e672c..7642366 100644
--- a/README.md
+++ b/README.md
@@ -2,6 +2,17 @@
 
 Predictive Models in Production
 
+## Workflow
+
+This is the regular workflow once you've got the whole system set up.
+
+1. Edit code and commit to Github.
+1. Build the Docker image locally.
+1. Push the Docker image to the AWS ECR repository (command below).
+1. Manually upload the latest model to AWS S3 s3://<your-S3-bucket>/models/staging/YYYYMMDD/, where YYYYMMDD is today.
+1. Deploy the AWS ElasticBeanstalk API with `eb deploy`. This should pick up the S3 model you just uploaded.
+1. Run the AWS Batch training job. This will make a new model (again) and restart the EB application (again).
+
 ## Local dev
 
 ### Pro tips
@@ -30,13 +41,13 @@ Then open [http://localhost:8888](http://localhost:8888) to run Jupyter.
 If you need to enter into the container's shell, do this.
 
 ```bash
-ENVIRONMENT=dev BUCKET=fwhigh-predictive-models bash scripts/run_training_container.sh -
+ENVIRONMENT=dev BUCKET=$BUCKET bash scripts/run_training_container.sh -
 ```
 
 ### Train a model programmatically
 
 ```bash
-ENVIRONMENT=dev BUCKET=fwhigh-predictive-models bash scripts/run_training_container.sh scripts/train.sh
+ENVIRONMENT=dev BUCKET=$BUCKET bash scripts/run_training_container.sh scripts/train.sh
 ```
 
 ### Pushing the new Docker image to production for the training and Flask API services
@@ -64,7 +75,7 @@ ENVIRONMENT=dev bash scripts/run_api_container.sh "python -m pmip.routes"
 Run the Flask API locally.
 
 ```bash
-ENVIRONMENT=dev BUCKET=fwhigh-predictive-models bash scripts/run_api_container.sh
+ENVIRONMENT=dev BUCKET=$BUCKET bash scripts/run_api_container.sh
 ```
 
 Drop into the Flask API container.
@@ -76,7 +87,7 @@ ENVIRONMENT=dev bash scripts/run_api_container.sh -
 ### Push the Lambda API to Lambda
 
 ```bash
-BUCKET=fwhigh-predictive-models serverless deploy --region $([ -z "$AWS_DEFAULT_REGION" ] && aws configure get region || echo "$AWS_DEFAULT_REGION")
+BUCKET=$BUCKET serverless deploy --region $([ -z "$AWS_DEFAULT_REGION" ] && aws configure get region || echo "$AWS_DEFAULT_REGION")
 ``` 
 
 Run it
diff --git a/pmip/routes.py b/pmip/routes.py
index 4d33130..b07226a 100644
--- a/pmip/routes.py
+++ b/pmip/routes.py
@@ -23,7 +23,7 @@ def possible_types(value):
     )
 elif os.getenv('ENVIRONMENT', '') in ['staging', 'prod']:
     latest_model_id = get_latest_s3_dateint(
-        datadir='models',
+        datadir=f'models/{os.getenv("ENVIRONMENT")}',
         bucket=os.getenv('BUCKET')
     )
     model = load_from_s3_and_unpickle(
diff --git a/scripts/get_training_data.sh b/scripts/get_training_data.sh
index e577a92..8ae6335 100644
--- a/scripts/get_training_data.sh
+++ b/scripts/get_training_data.sh
@@ -8,4 +8,4 @@ echo Writing it to $DIR
 
 aws s3 cp --recursive $S3_DIR/ $DIR/
 cd $DIR
-unzip -f *.zip
+unzip *.zip