diff --git a/lumen/ai/coordinator.py b/lumen/ai/coordinator.py
index 409242b85..b98a84f6b 100644
--- a/lumen/ai/coordinator.py
+++ b/lumen/ai/coordinator.py
@@ -521,14 +521,10 @@ async def _lookup_schemas(
         cache: dict[str, dict] | None = None
     ) -> str:
         cache = cache or {}
-        to_query, queries = [], []
         for table in requested:
             if table in provided or table in cache:
                 continue
-            to_query.append(table)
-            queries.append(get_schema(tables[table], table, limit=3))
-        for table, schema in zip(to_query, await asyncio.gather(*queries)):
-            cache[table] = schema
+            cache[table] = await get_schema(tables[table], table, limit=3)
         schema_info = ''
         for table in requested:
             if table in provided:
diff --git a/lumen/ai/models.py b/lumen/ai/models.py
index 06f307d95..6a7686a2e 100644
--- a/lumen/ai/models.py
+++ b/lumen/ai/models.py
@@ -116,7 +116,7 @@ def make_plan_models(agent_names: list[str], tables: list[str]):
         extras['tables'] = (
             list[Literal[tuple(tables)]],
             FieldInfo(
-                description="A list of tables you want to inspect before coming up with a plan."
+                description="A list of tables to load into memory before coming up with a plan. NOTE: Simple queries asking to list the tables/datasets do not require loading the tables."
             )
         )
     reasoning = create_model(