A big part of my rpg sessions is spent doing social scenes. The players are talking with a noble to try to understand the job offer he is presenting. The players are befriending the elves to see if they can learn more about the magical youth living in their community. The players are negotiating with the prince to see if they can get a higher price for the assassination job he has offered them. These sorts of scenes take up a lot of play time. In the kinds of games I run, they often take up more time than combat.
These social scenes take up a lot of time because they are important. It can be important that information is given to the players so they know more about the plot or gain the next clue in their investigation. It can also be important for world-building so that the players can pick up on details like accents and mannerisms for NPCs or cultures. It is also the best opportunity that players have to show everyone at the table what makes their characters tick. We can learn about the personalities and priorities of the PCs.
They're also fun! It can be fun to put on funny accents and attempt to embody and flex some amateur dramatics with my friends.
So if interaction scenes with NPCs are such an important part of my games, shouldn't I be drawn to play an rpg system that has a well thought out and detailed social encounter system? Maybe. I have played several systems that have tried to crack this nut, with mixed results...
D&D5e : A Skill Check
This is the most basic system. The GM pretends to be the NPC and the player takes the role of their character. Everyone talks while pretending to be their character. They lay down their points of negotiation. Then the GM asks for a skill check against a defined target number. It either works or it doesn't.
Chronicles of Darkness : Doors
Basically, the player declares their objective and the GM announces a difficulty. For example, the difficulty is described as "3 Doors" or "4 Doors". This is the number of meetings or interactions with the NPC you will need to have. You can try to blackmail the target, which can bypass 1 door without rolling dice. You need to wait a period of time (e.g. 1 day) in between attempts to talk. Essentially you are wearing them down, and applying leverage to get the NPC to agree with you. If you fail too many times, the NPC tells you to bugger off and won't listen to your arguments any more.
Legend of the Five Rings 5e : Social Techniques
Social characters have defined abilities. They are purchased in the same way as a samurai would buy certain sword moves or other combat abilities. Typically, when a social ability is successfully activated against an NPC, then they must agree to your point, or else they take damage to one of their social scores (like Honour). The NPC has a choice. They can choose to suffer the consequences and go against you. Or they can comply. It is possible to buy defensive social abilities that block or increase the difficulty of such moves being used against you.
The One Ring 2e : Council Scenes
The party stands before Elrond. They need to convince him to loan them some fast elvish horses. The goal is declared and one party member is chosen as the spokesperson. Other members of the party make dice checks in support, then the spokesperson makes a roll to clinch the negotiation.
Here are my thoughts on how these systems have played at the table.
Firstly, adding a game system to the interaction makes it structured. Usually, a big part of a social scene is informal conversation. The players are speaking with the NPC, asking questions and answering. They are beating around the bush, feeling out the NPC, and receiving exposition before forming a plan of action in their own minds. In my opinion, this part of the interaction shouldn't be gamified.
However, once the chatting part of the scene is over, there comes a time when the players need to focus a bit and the scene needs to be wrapped up. Otherwise the pacing of your game can drag a bit as conversation begins to go in circles. This is the point where it is helpful to have a social subsystem in your game. I'd ask the players to specify the main points of their argument, or to specify what they want to achieve. Then we can roll some dice to determine how the scene unfolds.
For me, there is no point in rolling if there is no dramatic risk. What will the players risk by attempting this negotiation? What happens if they fail the dice roll? Will they accept a higher price in their bargaining? Will they lose reputation or status? Agree to this first, before reaching for the dice.
There is a problem, because I sometimes want to be able to employ these subsystems *against* the players. Players don't like that. Not one bit.
In D&D5e, suppose a player wants to try to negotiate with the shopkeeper to reduce the price of the armour from 100gp. They should be ready to accept the possible outcome that the shopkeeper rolls better than them, that increases the price to 125gp.
In Vampire, suppose a rival is trying to force a player to surrender territory. The rival phones them up and begins laying out terms. Should the GM tell the player the rival has 3 Doors to overcome over the next few days, and if they succeed then they have won the territory?
Why is it perceived as unfair for NPCs to use social abilities against PCs, but not the other way around? Players don't mind losing hit points (or even dying!) in a combat encounter as much as they loathe losing face or assets from a social combat.
At least in L5R, the player has a choice. They can accept the Honour loss to oppose the slimy courtier in front of their daimyo, or they can keep the Honour and go along with them. Those Honour points are hard won! They're much harder to recover than hit points lost in a combat encounter would be - but at least the player has a choice. Choice is important. Social moves are not mind control.
In Summary, a good method for me is to open the scene with some informal chatting and back-and-forth in character conversation. Then, when that wraps up, to summarise what is hoped to be gained, what threats or pressure can be brought to bear. Also you should specify what the players are risking here - what could they lose if things go badly. Decide how many dice checks will be needed, and have a way to count progress (e.g. Ironsworn tracker or a Clock). Now ask the players if they want to proceed. This is their choice. If yes, then call for a few dice rolls to determine the outcome and costs to the players.
This is the kind of system I would use.