Assuming you are doing things right, after you get things automated and into a somewhat steady state, which means that all the really boring testing is out of the way, you get to do exploratory testing. I really enjoy doing exploratory testing because I get to sit with the developer, pick their brain on how their widget works, talk through scenarios of what to test and what to look for, and come up with a rough game plan on how I will test their widget. With that playbook in hand I then proceed to take their widget and waterboard it until I find defects. I frequently come up with new and interesting things to do to their widget that no one has thought of before and as a result usually find some interesting defects. After exploratory testing is over it is time to automate the regression tests, verify bug fixes, and move on to the next widget. Suprisingly there is never a shortage of things to test.