diff --git a/.jules/bolt.md b/.jules/bolt.md
new file mode 100644
index 00000000..d5d04628
--- /dev/null
+++ b/.jules/bolt.md
@@ -0,0 +1,3 @@
+## 2024-03-24 - Parallelize sequential I/O queries in map loops
+**Learning:** Sequential `for...of` loops performing I/O queries (like DB reads) for each iteration cause O(N) latency, creating significant bottlenecks as data grows.
+**Action:** Replace sequential loops with `Promise.all(array.map(async item => { ... }))` to run independent queries concurrently, reducing latency to O(1) bounded by the slowest query. Always ensure queries are truly independent and DB connection limits allow the concurrent load.
diff --git a/apps/api/src/application/usecase-invocations.ts b/apps/api/src/application/usecase-invocations.ts
index 8625c08b..e85821c7 100644
--- a/apps/api/src/application/usecase-invocations.ts
+++ b/apps/api/src/application/usecase-invocations.ts
@@ -46,9 +46,21 @@ export async function invocationGraph(
   projectId: string
 ): Promise<InvocationGraph> {
   const graph: InvocationGraph = new Map();
-  for (const usecase of await deps.useCaseStore.listUseCases(projectId)) {
-    graph.set(usecase.key, await invocationEdgesFor(deps, usecase));
+  const usecases = await deps.useCaseStore.listUseCases(projectId);
+
+  // ⚡ Bolt: Execute invocation edge queries concurrently instead of sequentially
+  // Reduces latency from O(N) to O(1) in the number of use cases
+  const results = await Promise.all(
+    usecases.map(async (usecase) => {
+      const edges = await invocationEdgesFor(deps, usecase);
+      return { key: usecase.key, edges };
+    })
+  );
+
+  for (const { key, edges } of results) {
+    graph.set(key, edges);
   }
+
   return graph;
 }