Does Unity need a better out of the box AI toolset?

I would say it's that it's not a common concern. Most of the people working with Unity probably aren't interested in making cover based shooters, and if you're working in 2d, line of sight is more interesting to play around with. The ones that do try it are liable to be too inexperienced to even get that far, or know what they are doing in the first place.

 

Things like cover probably don't deserve to be built in. It would be useless for many games. It exists in unreal because of its heritage as an FPS engine. And if you want to build FPS games, then you are wasting your time with Unity.

When I think AI tools I think a navmesh where I can easily mess with pretty much every weight on an agent level. One I can dynamically manipulate at runtime. And one that is not just restricted to being baked from 3D terrains.

I think of an FSM implementation that is not so heavily tied to the animation system. I think of built in support for behaviour trees.

These things could be done on a genetic basis for multiple projects. The specific stuff I can deal with.

 

Seeking Cover is just a use case for an Environmental Query System something that can probably be used in any games that have agents that need to respond to or act within an Environment 2D or 3D.

There are games that will not need this system, e.g. Quizzes and Logic puzzles but lots of games will probably find a use for this type of system/module within Unity.

 

Obviously, this is not part of the whole including such a system inside Unity idea but you may get something out of these videos:

Also this SILVER: Advanced AI asset lists "Cover-Finding (dynamic & manually placed cover areas)" as one of its features

I guess I am at the opposite end of the spectrum. All of this built-in pathfinding and other stuff I don't pay much attention to. But that's because I see it as just one more system made by someone else that I'd need to learn. If it was worth the time spent I could see it. Like the standard stuff learning the API for audio, graphics and other core stuff is necessary of course. But this higher level stuff often seems to me like it's more trouble than it's worth. On the other hand, it's enjoyable, fast & productive to build such things yourself. So I prefer just "give me solid building blocks so I can build whatever I need".

 

Got ya. A generic query system might be a useful addition. If it could be made successfully generic without losing too much power.