Hi, i worked with construct 2 & 3 for over 10 years now, i made huge games with these engines that you don't think its even possible. (i can show you)
i still think construct 3 need some stuff like saving things to the account because some times it forget i added plugins because some plugins can't be stored in project, specially outside plugins, but overall it was a really good decision that the team made and made it based on html and web. with the rise of AI, all the new things are going in web direction now, there are AIs now that make full websites. so in the future this engine will be even more powerful, i made huge games with construct, so big thanks to the team.
also supporting JavaScript coding is really amazing now.
but these futures are missing currently:
1. save things like plugins and settings in the account or project instead of local storge.
2. doing UI is really hard in construct compared to the other engine i won't name, if you look at their UI stuff you see what i mean, they have grid systems and can quickly combine multiple things to make one prefab and use put that in the grid or scroll View and that's it. i know construct 3 has hierarchy now, but its much more difficult then it should be, having drag an drop parent and child system is much better, for example think you want to make a leaderboard with icons and names and points, its very time consuming to make it here. but other engines : create a prefab of each element put a script that handle the whole prefab and put that prefab in a vertical grid and that's it you just need to create instances of that prefab.
3. prefabs and object oriented programming : so think you want to move want to create a behavior for everything that wants to zoom in and fade in for the UI and you want to do this for all UIs. you can not use families because they are type based so in UI you have sprites and texts. you can't use a function as well, because you still need to select the time in the function and there is no overall object.
so here the problem is you need to select the object, if instead of that we had a gameObject that can set to anything, the problem can be solve. so in the function you select gameObject, and when you want to use that function then you can pass the object or family, so now we have a thing that can run on anything. gameObject can have shared values only, for example position and rotation and scale or effects or shared selected behaviors. what i mean we can have a bigger family called gameObject that only populate shared variables.
so now we can write the code once, and with that even the problem with UI will be fix because now we can write a gird system once and use that across multiple objects or even projects.
sorry for the long explanation, but after all these years that's what i miss to have in construct, compared to other engines.
but overall i think this is an amazing engine and can have really good future from here now as well. also try to implement some kind of AI generating code thing.
Thanks to the Ashly and construct team.