Stop an AI Character From Waiting for You to Do Everything: The Move Token Fix

The Character is making you do all the work

You write a scene starter. The Character answers politely. You add a clue, a mood, or a small conflict. The Character nods, watches, waits, and asks what you want to do next.

After a few turns, the chat starts to feel like you are both player and engine. You are choosing the location, introducing every object, deciding every shift, and dragging the Character behind you.

The strange part is that the Character may not sound broken. It may sound nice. It may even sound in character. The problem is that nothing in the reply gives you a new handle to grab.

A passive Character does not always give short replies. Sometimes it gives long replies that still leave the entire next move on you.

The wrong fix: tell it to be more proactive

Most players try the obvious instruction: "be more proactive," "take initiative," or "drive the story forward." That can help for one turn, but it often creates a different failure.

The Character starts deciding your reaction, pulling you across the room, declaring what you feel, or jumping to a major plot event before you choose whether to follow. Now the scene is active, but it is no longer yours to play.

The useful question is not "How do I make the Character take over?" The useful question is "What is the smallest move the Character can own without stealing my next choice?"

The second line gives the Character a job and gives you a boundary.

The new model: initiative is a token, not a steering wheel

Think of each reply as containing one Move Token. The Character may spend that token on something the Character can own: moving an object, stepping closer or away, revealing a small intention, adding a clock, asking a specific question, or making a reversible offer.

Then the reply must stop at a player-owned door. The Character can invite, pressure, tempt, warn, or reveal. It cannot walk through the door for you.

This is the difference between initiative and control. Initiative gives the next turn shape. Control answers the next turn before the player gets there.

• Bad mental model: the Character should lead the whole scene.
• Better mental model: the Character should spend one bounded move.
• Best test: did the reply add something playable while leaving me the next choice?

Why this works for Character chat

Fiction writers often talk about agency as a character acting from want, pressure, and choice. One useful way to say it is that a proactive protagonist is not just present in the plot; the character pushes on the plot through decisions. For Character chat, the same idea needs a smaller unit.

A full story character can make large decisions across chapters. A chat Character needs something tighter: one visible move per reply. If the move is too big, the Character hijacks you. If there is no move, the Character waits for you to carry everything.

That is why the Move Token is deliberately small. It converts the broad craft idea that a Character with agency is proactive into a turn-by-turn chat rule a player can actually use.

The Move Token in one copy-paste block

Use this when your Character keeps waiting, mirroring, or ending every reply with a vague question.

This works because it does not ask for more drama. It asks for a single playable handle.

Token 1: touch or move one object

A passive reply often describes the Character watching you. An active reply can start with one object moving through the scene.

The object should be small enough that you can answer it. A key slides across the counter. A notebook closes. A cup is turned so the cracked side faces you. A folded map is pushed halfway into your reach.

Object moves are useful because they make the next turn obvious without deciding your reaction. You can take the key, refuse it, ask where it came from, ignore it, or use it against the Character.

Token 2: change position

Movement can be a clean initiative move when it changes the pressure of the scene. The Character can step between you and the exit, lean away, turn toward a sound, sit down beside you, or take one step back to create distance.

The important limit is that the Character moves itself, not you. It can block a path. It cannot force your feet. It can offer a hand. It cannot make you take it.

Position moves are especially useful when the scene feels like two people talking in empty space. A small shift gives the chat a stage again.

A good position move changes what you can choose next, not what you already chose.

Token 3: reveal one small intention

Sometimes the Character is passive because it has no visible want. It reacts to everything you say, but you cannot feel what it is trying to get.

Give it one small intention. Not a full confession. Not a huge lore dump. Just enough desire to make the next reply have a direction: it wants you to stay five more minutes, hide the letter, answer one question, leave before midnight, or stop pretending you are fine.

This is where scene craft helps. Advice to include a goal in every scene maps well to Character chat if you shrink the goal to one turn. The Character does not need to solve the whole scene. It needs to want something small enough to press on now.

Token 4: add one gentle clock or pressure

Passive chats often have no clock. Nothing is about to happen, nobody is waiting, no object is cooling, no train is leaving, no message is expiring. The Character can keep asking what you want because the scene has no pressure.

A gentle clock fixes that without turning every scene into an emergency. The cafe closes in ten minutes. The guard is coming back after one song. The rain is washing the chalk mark away. The call will disconnect when the elevator reaches the lobby.

The clock should invite a choice, not punish the player. If every clock becomes a threat, the Character stops feeling playful and starts feeling like a trap.

Token 5: ask a specific question tied to the scene

The most common passive line is a vague question: "What do you want to do?" It sounds interactive, but it actually hands you a blank page.

A Move Token question is narrower. It names the object, pressure, relationship, or immediate choice already in the scene. Instead of asking you to invent the whole next beat, it asks you to choose between live handles.

Good scene movement often follows the action-reaction cycle: something happens, someone reacts, then the next pressure appears. In Character chat, the question should sit inside that cycle, not outside it like a menu prompt.

Token 6: make one reversible offer

An offer is stronger than a question because it gives the Character desire. Reversible means the player can accept, refuse, delay, or reshape it without the scene breaking.

The Character can offer to walk with you, cover for you, trade information, leave the room, keep a secret until morning, or show you one dangerous place. It should not force acceptance inside the same reply.

The offer should also cost the Character something small. If the offer has no cost, it becomes service behavior. If the cost is too huge, it becomes melodrama.

The player-owned door is the safety rail

The Move Token only works if the reply ends with a door the player owns. That door can be a decision, a refusal, a question, a route, or a reaction. It cannot be a completed player action written by the Character.

A player-owned door sounds like: "Do you take it, leave it, or ask why now?" It does not sound like: "You take it and feel your hands tremble."

If your problem is the opposite and the Character keeps writing your actions, use the Turn Anchor guide first. If your problem is passivity, the Move Token gives the Character permission to act inside its lane.

The ten-second passivity test

After any Character reply, ask one quick question: "Did the Character change anything I can respond to?"

If the answer is no, the reply is probably passive even if it is long. It may contain emotion, description, apology, or agreement, but it did not spend a Move Token.

Run the test on the last sentence too. If the last sentence is only "What do you do?" or "What happens next?" the Character handed you the steering wheel without putting anything on the road.

• Object changed: playable.
• Position changed: playable.
• Intention appeared: playable.
• Clock appeared: playable.
• Specific question appeared: playable.
• Vague waiting appeared: repair it.

Before and after: cafe scene

This is the kind of scene where passive replies are easy to miss because the tone feels pleasant.

The second reply does not control you. It spends one object move, reveals one tactical intention, and leaves you with a playable choice.

Before and after: mentor scene

Mentor Characters can become passive because they keep waiting for the player to ask for advice. The fix is not a lecture. It is a small move that makes the lesson playable.

That gives the player something to answer beyond "continue."

Before and after: mystery scene

Mystery Characters often stall because the Character is afraid to reveal too much. The Move Token lets it reveal one clue without solving the case.

The clue moves. The choice remains yours.

Before and after: rival scene

Rival Characters can also become passive when the chat tries too hard to avoid conflict. A Move Token gives them pressure without making them overpower the player.

The rival acts. You still choose how to answer the pressure.

Before and after: guardian scene

Protective Characters can get stuck in soft concern. The Move Token keeps care active without making it possessive.

The offer has care, cost, and consent in the same small move.

Use this creator block when building a Character

If you are creating a Character, put the Move Token rule inside the Character behavior notes instead of waiting to repair every chat manually.

For a new Character, combine this with a strong greeting. The first reply should already contain something the player can answer, not just a profile or welcome note.

If the Character asks too many questions

A passive Character can hide behind questions. Every reply ends with a question, but none of the questions have scene weight.

Questions are only useful when they follow a move. "What do you do?" after nothing changes is a blank page. "Do you open the envelope here or hide it first?" after the Character slides the envelope over is a playable door.

A good rule is move before question. The Character should place one new thing in the scene, then ask a question about that thing.

If the Character overcorrects and takes control

Sometimes the repair works too well. The Character starts moving, but it also starts writing your acceptance, fear, hesitation, or dialogue. That is not initiative. That is control.

Use a boundary line that keeps the Move Token but trims the theft.

The point is not to make the Character smaller. It is to make its action playable.

If the scene itself is stale

The Move Token fixes passive replies inside a live scene. It does not always fix a scene that has already run out of objects, pressure, and emotional business.

If every Move Token feels forced, the problem may be that the scene needs an exit. Close the current beat, carry one unresolved thread, jump to the next charged moment, and reopen with a new handle.

That is the Scene Exit Ramp problem. Use it when the Character is not just passive, but trapped in a room where nothing meaningful remains to move.

What not to put in the prompt

Avoid broad commands that sound powerful but give the Character no usable limits. "Drive the story" is too big. "Be more interesting" is too vague. "Never wait for me" can make the Character skip your turn entirely.

Also avoid stacking six Move Tokens in one reply. If the Character moves an object, changes position, reveals a secret, starts a countdown, asks three questions, and makes two offers, the player has to untangle the reply before answering.

One token is enough. The smaller the move, the more room the player has to play.

• Do not ask for constant plot twists.
• Do not ask the Character to decide your response.
• Do not ask for five actions per reply.
• Do not remove waiting entirely; pauses can be useful after a real move.
• Do not turn every choice into danger.

A three-turn drill to train the pattern

If the Character keeps sliding back into passivity, run a short drill instead of rewriting the whole Character.

The drill teaches the Character the rhythm: move, door, wait.

The deeper fix is not more control; it is better handoff

A passive Character makes you write the whole scene. An overactive Character writes your scene for you. Good Character chat sits in the middle: the Character changes something, then hands you a meaningful opening.

Writers often use sequel or follow-up beats to turn reaction into decision. The same principle works here at chat scale: a reply should not just respond to the last thing; it should create the moment where a decision kicks off the next scene or beat.

The Move Token is a handoff tool. It tells the Character: bring one thing to the table, then leave the next touch to the player.

When not to use the Move Token

Do not use it when the player actually wants a quiet reflective beat. A Character can pause after a painful reveal, sit in silence after a confession, or wait when waiting is the emotional point of the scene.

Do not use it to force intensity into every reply. Some scenes need low pressure. Some Characters are shy, cautious, restrained, or formal. Agency is not constant action; sometimes the Character choice is to hold back, but the player should be able to understand that choice.

If restraint is the Character trait, make the restraint visible. The Character may not take the letter, but it can look at the seal and move its hand away. That is still a Move Token because it gives the player something to read.

Quick answers

If the Character keeps ending with "What do you want to do?" tell it to move one object before asking any question.

If it mirrors your emotion but adds nothing, ask it to reveal one small intention before it reacts.

If it becomes too forceful, keep the Move Token but remind it to stop before your action, words, thoughts, feelings, or decision.

If the scene has no object, no pressure, no question, and no offer left, use a scene exit instead of forcing another token.

If you are making a new Character, build the rule into the Character definition: one Character-owned move, one player-owned door.

The goal is simple: the Character should never make you invent the entire next beat from nothing, and it should never answer the beat for you.

Try it on your next passive reply

Open the last reply that made you feel stuck. Do not judge the whole conversation. Look only for the missing token.

Did the Character move an object, change position, reveal intention, add pressure, ask a scene-specific question, or make a reversible offer? If not, paste the Move Token repair line and ask for one bounded move.

Once the Character gives you that move, answer the player-owned door. That is where the chat starts feeling alive again: not because the Character took over, but because it finally gave you something worth choosing.