I’d say that while most rulebooks say that the GM is the one that preps and handles conflicts, as any actual game is a conversation from which a fiction is created dynamycally (no matter how much you railroad), conflicts actually arise naturally from the consequences of PC actions. It’s difficult to point a moment, a combination of words or a silent agreement between players and the GM, because it’s totally different from party to party. It might even be their silence that triggers the GM into framing a conflict in a certain way, delay one already prepped in favor of a new interesting one or jump steps in their prep to get straight to the conflict.
I’d say that in more than one way, players are usually driving the story into conflicts they are interested into and crash into conflicts they want to avoid because the GM manages to bring together the requirements to trigger it. And that’s one of the reasons why there are so few books whose rules advocate for player owned threats. The main reason we jump to is of course, Czege principle, but the need for games with player owned threats would be higher if this weren’t already happening at every table.
However, comfort guides our interests and today more designers are starting to think if there could be more ways to make GM easier by having the players come up with more stuff. Definitely, they can come up with their own conflicts, they already do. It’s just that they aren’t so vocal, explicit or detailed about it.
Take for instance the usual “I don’t think this guy is telling us the truth”. There’s the seed of a conflict there that any GM can feel and escalate. Sure, you can try to stop players in their tracks, but once the seed is planted and other players show their support it becomes harder, so the GM just rolls with it and improvises a conflict, if it isn’t anything already prepped that they can use.
Now, this side has been handled by the “follow the players” and “use the consequences” advice (I can’t even remember where I read them first and how they were written but I can bet it was before AW), as well as but there’s another venue I can think from the top of my head, which is character creation.
At character creation there’s usually a lot of wasted data that by tradition designers and GMs let go untouched: player expectations. Historically it’s been wasted because players will either be too vocal and expand on what they want for their character or too polite and afraid to tell. Nonetheless they all have expectations, even if they don’t know exactly which they are. But when they take a choice for their character they are already imagining stuff they totally expect to see in the game. And if you put that in the game or make it better, they will respond to it, or dislike/despise it if you do poorly or too different in terms of the feeling they expected.
Like, if they choose to be an elf but they wanted Warcraft elves instead of Tolkien elves, or if you paint elves as forest rastafarian, some players will dig it and some don’t. Because of expectations.
So if the GM makes some simple questions about player choices for their character and uses that for the game, players get invested sooner and are more willing to contribute, at which point the GM only needs to moderate, limit their input and take notes. Totally applicable to conflicts too. Like, if they choose certain type of class feature/special power/skill, they believe it will be useful against something. You can easily imagine what from their choices, but it doesn’t hurt to ask.
And then there’s the fishing tank, which can be used at any point in the game to generate additional conflict. Just brainstorm it with the players. It already happens on some sessions when a player jokes about what could be the worst thing to happen at some point and the GM’s eyes illuminate and wham! it’s already on the table and worse than the players expected.