A game the office plays and everyone else can join

Two rooms, one game

Someone in the office throws a ball at a target. Everyone else, at home or in another city, opens the app and guesses the result before it happens.

Kovendan scored 918. Can you guess Belmont's score?

Sixty seconds to commit to a number. Points are awarded on how close you were. A leaderboard settles it.

Hybrid teams no longer share a room

In hybrid workplaces, meaningful engagement is hard to create. Traditional activities struggle to include both in-office and remote employees, and one-off events give nobody a reason to come back.

The idea started inside Adappt itself. Half the company worked from home, half from the office, and nothing connected the two socially. The team's research found the problem is widespread, and published industry data says the same.

My role

I owned the visual direction and the mobile app's interface, meaning every screen an employee sees. Two other designers worked on the platform alongside me: one led the problem research, the other worked on the admin side.

It worked. It did not feel like a game.

The first version was blue and light. Everything functioned and everything was legible, and it looked like software you would be issued at work, which is the one thing it could not be.

Make it feel like a game is not a specification. What it told me was that the interface was reporting on the event instead of taking part in it.

I do not have the earlier screens to show. They were replaced rather than archived, which I would handle differently now.

A guess is not a form field

The player has to enter an arbitrary number. Not a choice from four options, because the score could be 30, or 300, or 918. Under a countdown, one-handed, while watching the game.

A number keyboard was the obvious answer. It is free, and it is right for a form. It is wrong here: it covers half the screen, and half the screen is the photograph they are predicting. It wants two hands. And typing 600 into a field feels like filing a return, not placing a bet.

I designed a dial. Rotate to a value, watch the number climb, commit. One thumb, the image still visible, and turning something to a value feels like a wager. The interaction carries the tension instead of flattening it.

I explored a linear slider first. It was more precise and less fun, and precision is not what the moment needs.

The order of the reveal is the design

Poll closed. Your number stays on screen at full size while the round resolves, because it is what you are invested in now, with a piece of trivia underneath. The alternative was a spinner, which tells you nothing and makes five seconds feel like thirty.

Round end. The correct answer lands first, the biggest thing on the screen. Then the winning guess and the room's average, because a number means nothing without a spread. Then the leaderboard, with your row picked out.

Joining cannot take longer than playing

A session runs for a few minutes. Anything that stands between a person and the game has to be shorter than the game, or there is no game.

You register once with a work email and confirm a code. Then you land on a home screen that is deliberately empty, carrying a single instruction: scan the host's code. The host puts a QR up on the screen in the room, and everyone in the building and everyone at home points a camera at the same thing.

The alternative was a room code. It would have worked, and it would have been circulated over Teams like everything else. But a code is something to read, remember and type correctly under time pressure, and every one of those is a step where a player gives up. Pointing a camera is none of them.

Every state, drawn once

The screens sit on a component library I built alongside them. Eight colour families at ten steps each, and one typeface carrying fifty-two text styles across display, title, body and caption.

Buttons come in three types at three sizes with five states each, and link and icon buttons carry the same five. Inputs run four variants through eight states, covering empty, active, filled, error, success and disabled, with separate fields for OTP entry and separate controls for checkbox, radio and toggle. Tags in seven colours at three weights. Avatars at eleven sizes, as initials or as a photograph.

There is also a focus stroke, because a live game played on a phone still has to be reachable by anyone.

Tooltips, toast alerts and the radius and spacing scales are in the library too. They are not reproduced here.

Most days, nothing is scheduled

An engagement app that only works during company events is one people delete. Quick Guess sits directly under the live game on the home screen: short general-knowledge rounds, always available.

It began as something to do while waiting for a session to start. The more useful framing turned out to be the reverse, a reason to open the app when nothing is happening at all.

Reading down the home screen is reading down a priority order: the live game and its countdown, then Quick Guess, then the social blocks.

What shipped was general knowledge, and I would argue for something else. Trivia is finite, and knowing that honey never spoils connects you to nobody you work with.

The version I proposed points the same prediction mechanic at the team. Everyone answers one question about themselves in the morning, something as small as tea or coffee. Later, everyone guesses what the room said: what share chose tea. The answer is revealed at a set time.

It uses the mechanic the product already has, it never runs out of content because the team generates it, and it is the only part of GameJamm that would work across time zones, since nobody has to be present at the same moment. One constraint I would hold to: questions stay about preferences, never about people. Tea or coffee is safe. Who is always late is a workplace problem waiting to happen.

Where it got to, and what I would change

The app was designed and handed to engineering. It has not launched, so there are no numbers. I would rather say that than quote a figure I cannot stand behind.

If it does launch, the measure worth watching is not total plays. It is whether remote employees take part at the same rate as people in the office, because that is the actual claim the product makes.

The dial needs a caption. It currently reads rotate to select your answer. A control that has to explain itself inside a sixty-second window is not finished. I would test a one-time animation on first use, then remove the caption entirely.

Content is the retention problem. Research on workplace gamification consistently finds engagement drops after roughly four weeks unless the content stays alive, and a fixed set of trivia categories is exactly the thing that empties out. The team-question format above is my answer to that, because the content is generated by the people playing. A recurring host character, the way Duolingo uses one, would give the app a voice to deliver it in. Both are ideas rather than designs, and I have not drawn either.

The deeper one. The person in the office plays and the person at home guesses. That is much closer than a video call, but it is not yet the same game. Giving the remote player something only they can do, like setting the challenge or picking who goes next, is where I would take it.