diff --git a/data/ai/micro_ais/scenarios/recruiting.cfg b/data/ai/micro_ais/scenarios/recruiting.cfg
index 281cf5e970df..c4da46dee754 100644
--- a/data/ai/micro_ais/scenarios/recruiting.cfg
+++ b/data/ai/micro_ais/scenarios/recruiting.cfg
@@ -64,6 +64,10 @@
                 type=Swordsman,Peasant
                 probability=8
             [/probability]
+            [probability]
+            	type=Mage
+            	probability=4
+            [/probability]
 
             skip_low_gold_recruiting=yes
         [/micro_ai]
@@ -83,7 +87,7 @@
 
         {MESSAGE Langzhar "" "" _"Just watch the recruiting of both sides and see if it is what you would expect. The recruitment lists cover level 0 to level 2 units, in order to make differences more obvious."}
         # wmllint: unbalanced-on
-        {MESSAGE narrator "wesnoth-icon.png" "Notes" _"If you have not changed anything in the scenario code, Side 1 uses the Random Recruitment Micro AI, with swordsmen and peasants having been given higher probability than the other units. This is not meant as a good recruitment pattern, it simply serves as a demonstration how to use the AI.
+        {MESSAGE narrator "wesnoth-icon.png" "Notes" _"If you have not changed anything in the scenario code, Side 1 uses the Random Recruitment Micro AI, with swordsmen and peasants having been given higher probability than the other units, and mages a middling probability. This is not meant as a good recruitment pattern, it simply serves as a demonstration how to use the AI.
 
 Side 2 uses the Rush Recruitment Micro AI (which is also used in the Experimental AI).