An AI NPC That Answers Your Question About What Your Zoi Wants

What if there were an AI NPC you could ask questions about your Zois?

You could ask things like:

  • “Does this Zoi seem interested in marriage?”
  • “What kind of relationship would these two Zois have?”
  • “Who does my Zoi seem to like more?”
  • “What kind of job would suit this Zoi?”
  • “What might happen if I choose this person instead of that one?”

The NPC would look at things like personality, relationships, preferences, recent experiences, and current circumstances, then give the player their own interpretation.

Purrsona, Psycat’s junior assistant

The NPC could be Purrsona, a cat who occasionally roams around the world of inZOI.

Lore-wise, Purrsona could be Psycat’s junior assistant at the AR Company.

Since Purrsona is still a junior assistant, their analysis does not necessarily represent what a Zoi is truly thinking. They are simply observing the available information and making an educated guess.

This means Zois always have chance of acting differently from Purrsona’s interpretation.

Purrsona can occasionally be wrong without it feeling strange.

“I’m still new at this job, meow…”

The important part: Purrsona does not reveal a Zoi’s true feelings

I think this distinction is important.

For example:

Hailey(Player Zoi) — Neat Freak, Devoted
Jason(NPC Zoi 1) — Slacker, Wandering Heart
David(NPC Zoi 2) — Workaholic, Cheerful

If we ask a Zoi directly:

“Do you want to marry Jason?”

and the Zoi clearly says:

“No.”

then choosing Jason anyway may feel strangely uncomfortable, because the game has explicitly told us what the Zoi wants.

But Purrsona is only an outside observer.

Instead of saying:

“Hailey wants to marry David.”

Purrsona might say:

“Life with Jason could have more unpredictable and dramatic episodes. Hailey seems to take relationships very seriously, while Jason may have a harder time settling down.”

“Life with David might be calmer and more stable, although his tendency to focus heavily on work could sometimes create distance between them.”

Neither choice is presented as better.

One relationship might create more drama and unexpected events, while another might create a more peaceful and stable life.

The player chooses which kind of story sounds more interesting.

An option to hide exact traits

Some players may also prefer not to know another Zoi’s exact traits.

In that case, Purrsona could use those traits internally without directly revealing their names.

Instead of saying:

“Jason has the Wandering Heart trait.”

Purrsona could say:

“Jason seems to enjoy meeting new people and may not be the type to settle down easily. Since Hailey seems to take relationships very seriously, their expectations about romance might clash.”

This would let Purrsona give useful advice without turning into a system information screen or spoiling the fun of gradually learning another Zoi’s personality.

Why I thought of this

inZOI already has many parameters that can matter when thinking about relationships between characters: traits, preferences, relationships, current emotions, past experiences, and potentially many other factors.

As these systems become more complex, it can also become difficult for the player to mentally combine all of that information.

You might find yourself thinking:

“This Zoi has this personality, that Zoi behaves like this, their relationship is like that, they recently had this experience… so what would actually happen if I put them together?”

Purrsona could help with that mental calculation.

Rather than simply displaying more numbers or parameters, AI could translate those systems into an easy-to-understand interpretation of what kind of story or situation might emerge.

There is also another reason I thought of this.

Sometimes players also want to hand part of a decision over to someone else.

You might have two possible partners, careers, or story directions and simply think:

“I honestly don’t know which one I want for my Zoi. What do you think?”

One way to create more unexpected situations would be to make Zois increasingly autonomous and let background AI constantly make these kinds of decisions.

But I imagine that could become computationally expensive if many Zois are constantly evaluating relationships, preferences, and possible decisions.

There is also a gameplay trade-off.

If Zoi autonomy becomes too strong, players may feel like they are losing control over their own stories.

But if autonomy is too limited, events can become predictable or repetitive.

So I thought an in-game AI advisor could be an interesting middle ground.

Instead of AI constantly making decisions in the background, the player could ask for advice only when they actually want it.

Purrsona would interpret the current situation, explain what kinds of outcomes or stories seem likely, and leave the final choice entirely to the player.

Technic-wise, since this npc will require quite heavy computing as well, we could make a system where we can ask this cat only once in a game day and receive the answer the next day

or we could visit a entirely separate, blank space secluded from Zoi world (to reduce computing) when asking questions.

It’s my first time writing here, so I’m a lot nervous, I’m sorry if there’s a similar topic already, but I’m open to suggestions and discussions!

1 Like

Welcome to the forum​:four_leaf_clover:

I think the presentation is cute, but I am still unsure what gameplay gap this would fill.

From my understanding, Purrsona would take information the game already has—traits, relationships and recent events—and generate a possible interpretation of it. That sounds less like uncovering an actual hidden simulation and more like creating headcanon for the player. For example, saying that a Zoi “might fear commitment” does not mean the game has simulated that feeling. It is simply one plausible story generated from their data.

inZOI already has a fortune teller, and I think that would be a much more natural way to offer this kind of optional guidance. The fortune teller could provide vague relationship readings, warnings or possible futures based loosely on the Zois involved:

“One heart seeks stability, while the other still looks toward the horizon.”

Because it is fortune-telling, the reading could remain symbolic, uncertain or even wrong without presenting itself as the true explanation of a Zoi’s personality. It would give players inspiration while still leaving room for their own interpretation.

I also do not think a separate AI space is necessary. Moving the Zoi into an empty room may reduce what needs to be rendered, but it would not reduce the processing needed to analyse the data and generate the response.

So personally, I would prefer an expansion of the existing fortune teller with carefully written, data-aware readings rather than another generative-AI feature. Otherwise, this mostly feels like an official headcanon generator telling me how I could interpret my own Zois.

I am also not sure why combining this information would be especially difficult for the player. At the moment, Zois do not have such an enormous number of deeply interconnected psychological systems that we need an AI to interpret them for us. Traits, preferences, emotions and relationships are generally visible and understandable.

The question “What would happen if I put these two together?” also cannot really be answered in advance, because future interactions, player choices and random events have not happened yet. The system could only generate one possible story based on their current data—it would not be revealing an actual predicted outcome.

I think clearer relationship summaries or more noticeable effects from traits would be more useful than an AI generating interpretations the player could already form themselves.

2 Likes

Thank you for such a thoughtful reply! It made me realize that I probably explained the feature before properly explaining why I wanted it in the first place.

So here was my original thought process:

  1. It can be tiring for me to make every single decision for my Zois.

  2. I imagine Zoi autonomy will take time to develop and polish, especially if it eventually covers more meaningful life decisions.

  3. So I thought: what if I could simply ask my Zoi? Even about trivial things, like what they want to wear tomorrow, where they want to go, or who they would rather date.

  4. But then I realized that creates another problem. If I don’t like their answer, I am suddenly arguing with my own Zoi—or I may feel guilty for ignoring something they explicitly told me they wanted.

  5. So I arrived at the idea of a third-party NPC who could give me an outside opinion about what might suit the Zoi or what they seem inclined toward, without presenting it as the Zoi’s actual will.

I also think the fortune-telling idea is great, but this is why I ended up imagining Purrsona instead.

You probably wouldn’t visit a fortune teller to ask what your family should have for dinner this weekend—but you might still have trouble deciding the menu and ask someone nearby, “What do you think?”

That’s basically how I imagine Purrsona:

a slightly more informed Magic 8 Ball. A tool that helps you decide.

What this could add to gameplay is a way to ask the game for a second opinion while still keeping full control over your Zois.

You wouldn’t have to wait for a future autonomy system where Zois independently quit their jobs, choose their own vacation destinations, or make other major decisions on their own. A simple way to enrich your story.

If you feel indecisive, you could simply ask your ‘rather untrusty’ AI NPC advisor:

“What do you think would suit this Zoi?”

Purrsona gives you a suggestion based on what they can observe, and then you still make the final decision.

Thank you for clarifying. I understand now that Purrsona would not represent the Zoi’s actual wishes, but would instead offer an optional outside suggestion so the player can ignore it without feeling as though they are going against their Zoi.

However, that makes me question even more why this would need to be a separate generative-AI feature. For everyday examples such as food, clothing or outings, the Zoi’s autonomy could already express a choice directly. If a hungry Zoi autonomously chooses and cooks a particular meal, isn’t that already the clearest answer to what they wanted to eat at that moment?

Simpler context-sensitive options such as “Suggest an outfit,” “Suggest a meal,” “Suggest an activity,” or “Surprise me” could also help with decision fatigue without requiring an AI character to interpret the household.

I also find the proposed interaction rather cumbersome for such ordinary questions. Entering a separate enclosed area—possibly through a loading transition—and using the single question allowed that day just to ask what a Zoi might want for dinner feels like a long detour. Limiting it to one question per day may reduce how often it runs, but it would not make the feature simpler to develop.

The argument that this would avoid waiting for improved autonomy also seems a little backwards to me. Developing a separate AI adviser would still require resources, integration with save data, UI work, localisation and testing, while the Zoi themselves would remain unable to make those decisions. I would prefer gradual improvements to their actual autonomy, so they can express their preferences through gameplay instead of a cat-shaped chatbot imagining choices on their behalf.

The character is still a cute idea, but after this clarification, the feature feels less like discovering what a Zoi wants and more like an elaborate suggestion or headcanon generator.

1 Like