Agree with Marsha, in my experience timing issues due to a slower test environment are often responsible for these types of failures. If you are using scripts look into the WaitWindow and WaitProperty methods for delaying execution until a required window or control is available. With Keyword tests and scripts increasing the Auto-wait timeout value on the Playback page of the project properties might help too. That is more of a "shotgun approach" instead of a targeted approach with the Wait methods.
One more thing, consider logging a screen capture at the step right before the problem/error. I find those really helpful for debugging problems in test environments that don't reproduce on my local machine. Seeing exactly what's going on when it fails can lead you quickly to the problem and solution.
Post Screenshot Operation | TestComplete Documentation (smartbear.com)
