I'm waiting for a time when the friends come over again. It's been a while. I've been playing RTDs where teams are a thing, so I understand how the team playing worrks. I tried running a DnD game with my brother as a character, and improving as I went. He didn't seemto like improv too much for some reason. Doesn't every plot start as an idea? wouldn't improv be a logical first step?
A completely improv setting may feel like there's no structure or point to anything. A good approach is to drop some plot hooks, see what gets the players attention then go with that. You can also create a bunch of small events/encounters/stories, and keep them around for when they might be useful. They don't have to be tied to any specific location, the player's won't even know. This way you have a bunch of interesting stuff for the players to see whichever direction they go in, and it won't feel completely random and made-up. So just write up a big bunch of NPCs, and think of interesting ways the players could meet them, along with a bunch of varied encounter/stories. Save them up until you need them. If they don't fit, put them aside in the stack instead of trying to shoe-horn them in. Maybe the players went through one of your dungeon's but didn't visit a bunch of rooms. You can just recycle the bits they didn't visit for a later story. Keeping room/encounter etc info on cards rather than writing in a book is useful for this reason.
For a really well-fleshed out campaign I get the feeling that the characters should have a mix of short, medium, and long-term goals that they're working to. Short-term is the immediate stuff, like the current encounter or task. Medium term is stuff like clearing the current dungeon, completing the next leg of a quest or travelling to the next city they need to go to.
Long-term is the defining stuff of the campaign. It should be deeper than "level up". For example, say that the main character's father was an inveterate gambler and racked up huge debts, and there are gangsters coming after the main character to collect. Additionally, throw in a detail such as the father owned a tavern, and that tavern is now occupied by the gangsters. Now, you've got a good plot hook for a small campaign. The player needs to work out how he's either going to raise the money or gain enough strength to get the collector's off his back and/or regain his father's tavern, which itself would be a useful location for further plots. It doesn't need to be something like this, but having something that's bigger than a single quest, yet smaller and more realistic than "save the world" is better than nothing. First level characters are kids. Them heading off to save the world, gaining levels as they go makes little sense and isn't a realistic motivator. The long-term goals need to be within realistic reach of the character's based on their current level. Once they're on the way to solving the original "long term" goals you drop hints (subtly) about bigger plots. For example there could be hints that the gangsters are working with the corrupt local baron's men, which could lead into a bigger plot about dealing with that corrupt lord.