diff --git a/integration/common/basesuites.go b/integration/common/basesuites.go
index 00e6b478..717c3049 100644
--- a/integration/common/basesuites.go
+++ b/integration/common/basesuites.go
@@ -69,9 +69,10 @@ func (s *BaseSuite) TestSimpleBuild() {
 			"--builder", s.Name,
 		)
 	}
+	imageName := "dummy.acme.com/" + s.Name + "replaceme:latest"
 	args = append(
 		args,
-		"--tag", s.Name+"replaceme:latest",
+		"--tag", imageName,
 		dir,
 	)
 	err = RunBuild(args)
@@ -81,6 +82,9 @@ func (s *BaseSuite) TestSimpleBuild() {
 	} else {
 		require.NoError(s.T(), err, "build failed")
 	}
+
+	err = RunSimpleBuildImageAsPod(context.Background(), s.Name+"-testbuiltimage", imageName, s.Namespace, s.ClientSet)
+	require.NoError(s.T(), err, "failed to start pod with image")
 }
 
 func isRootlessCreate(flags []string) bool {
diff --git a/integration/common/kubeclient.go b/integration/common/kubeclient.go
index 3a252359..d8cfc318 100644
--- a/integration/common/kubeclient.go
+++ b/integration/common/kubeclient.go
@@ -3,8 +3,17 @@
 package common
 
 import (
+	"context"
+	"fmt"
+	"time"
+
+	"github.com/sirupsen/logrus"
 	"k8s.io/cli-runtime/pkg/genericclioptions"
 	"k8s.io/client-go/kubernetes"
+
+	corev1 "k8s.io/api/core/v1"
+	v1 "k8s.io/api/core/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 )
 
 // GetKubeClientset retrieves the clientset and namespace
@@ -22,3 +31,96 @@ func GetKubeClientset() (*kubernetes.Clientset, string, error) {
 	clientset, err := kubernetes.NewForConfig(restClientConfig)
 	return clientset, ns, err
 }
+
+func RunSimpleBuildImageAsPod(ctx context.Context, name, imageName, namespace string, clientset *kubernetes.Clientset) error {
+	podClient := clientset.CoreV1().Pods(namespace)
+	eventClient := clientset.CoreV1().Events(namespace)
+	logrus.Infof("starting pod %s for image: %s", name, imageName)
+	// Start the pod
+	pod, err := podClient.Create(ctx,
+		&v1.Pod{
+			ObjectMeta: metav1.ObjectMeta{
+				Name: name,
+			},
+
+			Spec: corev1.PodSpec{
+				Containers: []corev1.Container{
+					{
+						Name:            name,
+						Image:           imageName,
+						Command:         []string{"sleep", "60"},
+						ImagePullPolicy: v1.PullNever,
+					},
+				},
+			},
+		},
+		metav1.CreateOptions{},
+	)
+	if err != nil {
+		return err
+	}
+
+	defer func() {
+		err := podClient.Delete(ctx, pod.Name, metav1.DeleteOptions{})
+		logrus.Warnf("failed to clean up pod %s: %s", pod.Name, err)
+	}()
+
+	logrus.Infof("waiting for pod to start...")
+	// Wait for it to get started, and make sure it isn't complaining about image not being found
+	// TODO - multi-node test clusters will need some refinement here if we wind up not scaling the builder up in some scenarios
+	var refUID *string
+	var refKind *string
+	reportedEvents := map[string]interface{}{}
+
+	// TODO - DRY this out with pkg/driver/kubernetes/driver.go:wait(...)
+	for try := 0; try < 100; try++ {
+
+		stringRefUID := string(pod.GetUID())
+		if len(stringRefUID) > 0 {
+			refUID = &stringRefUID
+		}
+		stringRefKind := pod.Kind
+		if len(stringRefKind) > 0 {
+			refKind = &stringRefKind
+		}
+		selector := eventClient.GetFieldSelector(&pod.Name, &pod.Namespace, refKind, refUID)
+		options := metav1.ListOptions{FieldSelector: selector.String()}
+		events, err2 := eventClient.List(ctx, options)
+		if err2 != nil {
+			return err2
+		}
+
+		for _, event := range events.Items {
+			if event.InvolvedObject.UID != pod.ObjectMeta.UID {
+				continue
+			}
+			msg := fmt.Sprintf("%s:%s:%s:%s\n",
+				event.Type,
+				pod.Name,
+				event.Reason,
+				event.Message,
+			)
+			if _, alreadyProcessed := reportedEvents[msg]; alreadyProcessed {
+				continue
+			}
+			reportedEvents[msg] = struct{}{}
+			logrus.Info(msg)
+
+			if event.Reason == "ErrImageNeverPull" {
+				// Fail fast, it will never converge
+				return fmt.Errorf(msg)
+			}
+		}
+
+		<-time.After(time.Duration(100+try*20) * time.Millisecond)
+		pod, err = podClient.Get(ctx, pod.Name, metav1.GetOptions{})
+		if err != nil {
+			return err
+		}
+		logrus.Infof("Pod Phase: %s", pod.Status.Phase)
+		if pod.Status.Phase == v1.PodRunning || pod.Status.Phase == v1.PodSucceeded {
+			return nil
+		}
+	}
+	return fmt.Errorf("pod never started")
+}
diff --git a/integration/suites/rootless_test.go b/integration/suites/rootless_test.go
index ccbf3c36..b4c9412c 100644
--- a/integration/suites/rootless_test.go
+++ b/integration/suites/rootless_test.go
@@ -22,6 +22,8 @@ func TestRootlessSuite(t *testing.T) {
 	})
 }
 
-// func (s *rootlessSuite) TestSimpleBuild() {
-// 	s.T().Skip("Rootless doesn't support loading to the runtime")
-// }
+func (s *rootlessSuite) TestSimpleBuild() {
+	// This test in the Base Suite attempts to run a pod, so we need to skip it
+	// Other tests will exercise the builder without running a pod
+	s.T().Skip("Rootless doesn't support loading to the runtime")
+}
diff --git a/pkg/driver/kubernetes/driver.go b/pkg/driver/kubernetes/driver.go
index 794df07b..e76f8d58 100644
--- a/pkg/driver/kubernetes/driver.go
+++ b/pkg/driver/kubernetes/driver.go
@@ -272,6 +272,7 @@ func (d *Driver) wait(ctx context.Context, sub progress.SubLogger) error {
 
 						sub.Log(1, []byte(fmt.Sprintf("WARN: initial attempt to deploy configured for the %s runtime failed, retrying with %s\n", attemptedRuntime, runtime)))
 						d.InitConfig.DriverOpts["runtime"] = runtime
+						d.InitConfig.DriverOpts["worker"] = "auto"
 						err = d.initDriverFromConfig() // This will toggle userSpecifiedRuntime to true to prevent cycles
 						if err != nil {
 							return err