In an ideal world, there would be tools to deal with every special case efficiency, and reduce time spent isolating special-case problems down to zero. However, in practice, the reality of game development quickly kicks in and forces you to make pragmatic decisions. What do the options look like and how do you decide where to draw the line?
My personal experience is that tools pay off when done well, as long as you keep control of the process and can accommodate custom changes. In the case of a Behavior Tree Editor, the tool quickly pays for itself compared to writing the trees manually. You’ll find your time shifting towards maintaining the tools — but that’s generally time well invested.
This question came up during the 4h Marathon Q&A I held live last weekend. Here’s the video of the answer; it’s almost 5 minutes long and 6.4 Mb. (Note, it was DDd who asked this question again, like the last one available here… Nice work!)
If you have any further questions or points to make about this topic, they are more than welcome below in the comments, or in The Game AI Forums — as usual. (Comments you post in the forum thread will show up in the blog.)