diff --git a/connectors/src/connectors/slack/temporal/activities.ts b/connectors/src/connectors/slack/temporal/activities.ts
index 5e17df8bf8567..c0fe42d5fb210 100644
--- a/connectors/src/connectors/slack/temporal/activities.ts
+++ b/connectors/src/connectors/slack/temporal/activities.ts
@@ -59,11 +59,26 @@ const MAX_SYNC_NON_THREAD_MESSAGES = 4000;
  * Broadly, you'll encounter limits like these, applied on a
  * "per API method per app per workspace" basis.
  * Tier 1: ~1 request per minute
- * Tier 2: ~20 request per minute (conversations.history)
+ * Tier 2: ~20 request per minute (conversations.history, conversation.list)
  * Tier 3: ~50 request per minute (conversations.replies)
  */
 
-export async function getChannels(
+/**
+ *  Call cached to avoid rate limits
+ *  ON RATE LIMIT ERRORS PERTAINING TO THIS FUNCTION:
+ * - the next step will be to paginate (overkill at time of writing)
+ * - see issue https://github.com/dust-tt/tasks/issues/1655
+ * - and related PR https://github.com/dust-tt/dust/pull/8709
+ * @param connectorId
+ * @param joinedOnly
+ */
+export const getChannels = cacheWithRedis(
+  _getChannelsUncached,
+  (connectorId, joinedOnly) => `slack-channels-${connectorId}-${joinedOnly}`,
+  5 * 60 * 1000
+);
+
+async function _getChannelsUncached(
   connectorId: ModelId,
   joinedOnly: boolean
 ): Promise<Channel[]> {
@@ -73,6 +88,8 @@ export async function getChannels(
   do {
     const c: ConversationsListResponse = await client.conversations.list({
       types: "public_channel, private_channel",
+      // despite the limit being 1000, slack may return fewer channels
+      // we observed ~50 channels per call at times see https://github.com/dust-tt/tasks/issues/1655
       limit: 1000,
       cursor: nextCursor,
     });