Kulfi gets couples playing together. What keeps them connected once they've played it all?
If you've never opened it: Kulfi is a couples-only app made in India. Two phones paired by a code, no singles, permanent dark mode. It's built for couples spending time together while apart, and most of that happens late at night.
Both Online games can be played any number of times a day, as long as you're both there. Both Offline games are the daily set: roughly 12 to 14 of them, each played once a day. That's the part that runs out fast: 15 to 20 minutes and you've been through it.
A questions space: you answer for your partner, they answer for you, and answers only unlock once you've both replied. Three are served daily, with a back-catalogue you can pull more from. When you're done with games and still missing them, you come here, and answer a few more.
I didn't start this as a case study. I started as a Kulfi user.
The games were playful and low-pressure, a real reason to open the app, and a nice way to spend time with someone without it being another call or another conversation. But the Both Offline set is finite. You play through the day's games, you answer the Pillow Talk questions, and then you hit the same quiet moment:
"I've done everything the app has. I still miss him. Now what?"
That's the moment I started designing for. Not "out of games," but out of everything the app offers, and still wanting to reach him. That's when I stopped just using Kulfi and started looking at it, as a product designer.
The Both Offline home: the daily set, mostly played through by night.
The problem wasn't really a lack of games. Kulfi already had a growing library and added new ones every week. But the daily Both Offline set exhausts fast, and once it's gone, along with the Pillow Talk questions, there isn't much left to do. (Honestly: that's an observation from my own repeated use, a hunch, not a proven user problem.)
The obvious answer was to add more games. But that only ever buys a few minutes, and it doesn't touch the actual feeling in that moment.
What if the next thing wasn't a game to play together, but something one person could make for the other? That reframe fit Kulfi more than it first seemed. Closeness across distance leans less on being online at the same time, and more on small, asynchronous gestures. Rhythm over volume, something to look forward to. It echoed Kulfi's own thinking, too: retention has been their hardest problem, and the founders call games "the wedge, not the destination."
A shared experience. Both people, online, at the same time. Wonderful, and finite.
One person making something small for the other inside the app. Not bought, not another game, built by hand and sent across. Works even when only one of you is around.
Both make the distance feel smaller, but only one of them survives an empty, one-sided late night.
I could have picked a hundred things to improve retention. Most of them were about filling the now: more games, more prompts, more to do tonight. A gift was the only direction that pointed at the future of the relationship instead of just occupying the present.
A game is fun in the moment, and it is a little one-sided: you play, you both pass the time. A gift is different. You are handing over a piece of yourself, telling the other person how you feel about them. That is more emotionally invested, more personal, more heart to heart, and it makes the bond stronger rather than just busier.
It also fits a real constraint of the people I was designing for. Long-distance couples can't easily make and post something physical, and you can't buy a real gift every day. There is no occasion, no budget, no time. A digital gift is low-cost and low-effort to send but still carries the one thing that actually matters: the thought. You made something, on purpose, for them. That is the signal, whether it arrives online or offline.
And quietly, it nudges the relationship in a healthy direction. When someone does something like this for you, you look for a way to do it back. The surprise is part of it too, that small "wait, he actually made me this?" moment. I built Pigeon Post to make that easy to give.
Pigeon Post is a small, made-by-hand gift you build inside Kulfi and send to your partner: a letter, a song, a voice note, a doodle, a photo, wrapped up and carried across by a pigeon. Here is how the idea got there, through a kill, a shift, a four-part experience, and the details that carry the feeling.
My first idea was the digital version of a keepsake: letters your partner saves and opens later, each tagged to a moment.
For a long-distance couple it felt right, but it carries a gap that even real letters have. Are you actually holding it at the moment you're sad? What stops someone opening them all at once? Does the effort land, or does the letter just get buried? A digital version doesn't fix any of that, it inherits it.
So I let the open-when idea go. A saved letter only copies something that already exists offline. I wanted the feature to be something new, made only because you made it, so I moved to building and customizing a gift instead.
The killed direction: "The Loft" of open-when letters.
So I simplified. Instead of messages tagged to hypothetical emotional moments, the interaction became one clear thing: when you're done playing, you can make something for your partner. Nothing you buy and nothing that ships. A small digital gift you build inside the app, out of your own words, voice and pictures, and send across in a tap.
And I kept it deliberately secondary. It only appears once you've exhausted the core experience, so the feature never announces "Kulfi is a gifting app now." It says something quieter: you're out of things to play, here's another way to reach them.
A concept study's deliverable is the experience itself, so here it is as four beats, each a real screen from the prototype.
You build a gift from things that feel personal: a letter, a song, a voice note, a doodle, a small memory. Each is its own object in the box, something they actually open.
Then you choose how it looks: the wrapping colour, the flowers, the way the bouquet comes together. The point was to make the gift feel chosen, not generated.
You decide the moment: now, a chosen day, or a surprise. Then a bird flies the gift across the night sky before it lands.
Your partner opens it, sees what's inside, and keeps it, landing in your normal chat, so it doesn't vanish with the session.
The heart of Create. Each one is a different piece of you, not a template, mixed and matched into one box.
One finished box, opened.
Words in your own hand. The thing you would actually write if they were next to you.
The track that says the thing you can't. It plays when they open it.
Your actual voice. A rant, a good-morning, a laugh, kept in the box.
Us, drawn badly. The imperfect sketch that only means something to you two.
A photo of a moment you want to keep, tucked in with a line about why.
I started with single flowers. Feedback from someone I showed it to pushed it further. Being able to compose a bouquet felt more personal. "Pick a flower" became "choose what you'd give them."
The sender picks the colour, based on what their partner would like, not a generic Kulfi gift, but one that looks like it was meant for a specific person.
It isn't needed for the feature to work. It's there so receiving feels like an event, because anticipation, not just the reward, is where the delight comes from.
None of it is decoration. Each detail reinforces the same thing: two people apart, and the wish to close the gap.
This was a 24-hour sprint, built by vibe-coding with Claude rather than the full Figma-first process. That flips the old order: you build first and figure the design out while building, deciding as working screens appear in front of you instead of on paper beforehand.
Solo, that meant wearing every hat a real Kulfi team splits up: the backend, the business and attention side, and the UI. You can't do all three well at once, so some things had to be compromises made on purpose. The real work was choosing what to keep, what to cut, and what to redo, and making those calls myself instead of letting the tool decide.
Deciding where not to spend effort mattered as much as anything I built.
If I took it further, I wouldn't just ask "does it work." I'd want to understand how people actually live with it:
Do couples actually make gifts for each other, and how often?
Are they intentional about it, or is it a throwaway tap? How much time do they spend making one?
Does it spark a conversation between them, or just sit there once opened?
Which pieces get used most: voice, music, letters, doodles, memories?
Does it feel like a natural part of Kulfi, or a separate product bolted on?
Does it pull them back on the nights the games and questions are already done?
Most of all, I'd test the assumption it rests on: does giving couples something to create for each other actually answer the "nothing left to do, still missing you" moment? That's what this project can't answer yet, and for a one-day exploration, that feels like the honest place to stop.
Going full-AI was chaotic. You get handed an overwhelming number of directions at once, and the early prototypes come out nothing like what's in your head. The content is off, the visuals are off, everything is off. It's easy to drown in the options.
What worked was treating AI as a thinking partner to explore with, never the decision-maker. That part stays mine.
My instinct was that the fix for people running out of things to do isn't more games, but a different reason to come back, an emotional one. To be clear, this is my reasoning, not a proven result. I had no user access and only the raw signals Claude helped me pull in a day, so I don't know what Kulfi's team has already tried. It's the hypothesis, not the verdict.
The bird, the flowers, the wrapping aren't random charm. They reinforce the emotional reality of distance. If they didn't, they wouldn't earn their place.
I came in as a user who hit a wall, and left with a working answer to it, one I'd still want to keep testing.
Disha · LinkedIn