diff --git a/contrib/containerized-pbench/pbench_demo b/contrib/containerized-pbench/pbench_demo
index 015acf7a53..fb29eb5bd1 100755
--- a/contrib/containerized-pbench/pbench_demo
+++ b/contrib/containerized-pbench/pbench_demo
@@ -1,34 +1,52 @@
-#! /bin/bash -xe
+#! /bin/bash -e
 #
 # This script provides a demonstration of the contrib/containerized-pbench/pbench
-# wrapper script.
+# wrapper script, using it to run the FIO workload.  It is not really intended
+# to be a tool unto itself -- it is intended to document the workflow which a
+# user might follow using the containerized Agent.
+#
+# Inputs to this script are provided via environment variables:
+#   PBENCH_AGENT_SERVER_LOC:  the Pbench Server host address, including schema
+#                                 and port (e.g., https://example.com:8443)
+#   PBENCH_API_KEY:  the Pbench Server API Key token value, which is obtained
+#                                 from the Pbench Dashboard profile page
 #
 
 #+
-# Set up a few things to make life simpler.  Typically, these would already be
-# set in the users environment (e.g., the `pbench` command alias would be done
-# by the user's login script; we wouldn't need the `shopt` command if these
-# commands were being run interactively; and, we only need `PB_AGENT_IMAGE_NAME`
-# here because we're not using the default image).
+# Set up a couple of things to make life simpler.  Typically, these would
+# already be set in the users environment (e.g., the `pbench` command alias
+# would be done by the user's login script, and we wouldn't need the `shopt`
+# command if these commands were being run interactively).
 #-
 shopt -s expand_aliases
 alias pbench="$(git rev-parse --show-toplevel)"/contrib/containerized-pbench/pbench
 
-FIOTEST=${PWD}/fiotest
-export PB_AGENT_PODMAN_OPTIONS="--pull newer -v ${FIOTEST}:/fiotest:z"
-export PB_AGENT_IMAGE_NAME=quay.io/pbench/pbench-agent-all-fedora-38:main
-
-mkdir -p "${FIOTEST}"
-
-# Before we run the demo, we need a Pbench API key. The key can be an
-# environment variable of PBENCH_API_KEY or need to be supplied to this
-# script as an argument.
-api_key=${1:-${PBENCH_API_KEY}}
-if [[ -z "${api_key}" ]]; then
-  echo "Pbench API key must be provided, either on the command line or via the PBENCH_API_KEY environment variable"
+#+
+# In order to upload the results of the demo to the Pbench Server, we need a
+# Pbench API key; and, to simplify things, we specify the Pbench Server on the
+# pbench-results-push command line instead of modifying the Agent configuration
+# file.  Make sure that the environment variables for them are set.
+#-
+if [[ -z "${PBENCH_API_KEY}" || ! (${PBENCH_AGENT_SERVER_LOC} =~ ^https?://[a-z0-9.-]+(:[0-9]+)?$) ]]; then
+  echo "Environment variables PBENCH_API_KEY and PBENCH_AGENT_SERVER_LOC must be properly defined"
   exit 2
 fi
 
+#+
+# Echo the rest of the commands as though the user had typed them.
+#-
+set -x
+
+#+
+# The workload that we're going to run requires a directory to run in, so
+# create that directory if it doesn't exist and then arrange for it to be made
+# available at a particular path inside the container where the workload will
+# be run.
+#-
+FIOTEST=${PWD}/fiotest
+mkdir -p "${FIOTEST}"
+export PB_AGENT_PODMAN_OPTIONS="--pull newer -v ${FIOTEST}:/fiotest:z"
+
 #+
 # Run the demo!
 #-
@@ -38,4 +56,4 @@ pbench pbench-user-benchmark --config example-workload -- \
     fio --directory=/fiotest --name fio_test_file --direct=1 --rw=randread \
         --bs=16k --size=100M --numjobs=8 --time_based --runtime=5s \
         --group_reporting --norandommap
-pbench pbench-results-move --token=${api_key}
+pbench pbench-results-move --server ${PBENCH_AGENT_SERVER_LOC} --token=${PBENCH_API_KEY}