Quest

A productivity app built on one rule: missing something isn't a loss condition, it's information. Explores what happens when you keep the dopamine and remove the loss condition — including a mechanic I built, then cut, once it contradicted the whole point.

Product Design

AI assisted Build

INTERVIEW INSIGHT

A points system would motivate her — until the first time it took points away.

Then she'd probably delete the app.

Rewards can motivate. Punishment can make someone disengage.

That tension was already at the center of what I was building. This conversation confirmed it.

Overview

Quest turns daunting tasks into small, forgiving adventures.

A productivity app designed around progress without punishment — break things down, start small, and come back whenever you’re ready.

Quest turns daunting tasks into small, forgiving adventures.

A productivity app designed around progress without punishment — break things down, start small, and come back whenever you’re ready.

Missing something isn't a loss condition. It's information

For people with ADHD or burnout, the barrier to doing a task is rarely the task itself — it's starting it. Replying to an email, doing the dishes, filling out a form: none of these are hard. But they can feel disproportionately heavy, triggering avoidance rather than effort. ADHD coaching has a name for this: the "Wall of Awful," a buildup of shame and frustration around a task that makes it harder to start — separate from, and often larger than, the task's actual difficulty. The longer a task sits, the higher that wall gets, and the harder it becomes to even open the app.

Most productivity and gamification apps make this worse. They frame an incomplete task as a failure: broken streaks, lost health points, a red "3 overdue" badge waiting on the home screen. For someone who is already avoiding something because it feels too big, that's not motivation — it's another reason to stay away.

Quest starts from a different premise: the task was too big, not the person. Missing something isn't a loss condition. It's information — a sign the task needs to be broken down further, not a reason to feel behind.

Problem

Where existing solutions fall short?

App

Habitica

Forest

Finch

Core Mechanic

HP loss, character death, guild accountability

A virtual tree dies if you leave the app

A pet you care for; gentle, no death

Where it breaks down for this problem

Missing a task has a punishment cost — adds pressure, which is the opposite of what someone avoiding a task needs

Same punishment logic, framed as nature instead of numbers

Closest in spirit, but built around general habit streaks —

not around breaking down dread for one specific hard task

None of these are built around the actual moment of friction: a person stalled on one task, afraid to reopen the app because they know it's still sitting there.

Design philosophy

No broken streaks, red warnings, or "you're behind." Time away doesn't become another thing to feel bad about.

An unfinished task is a signal to make the next step easier, not increase pressure.

Grow, never shrink.

Restructure, don't punish.

Softness by default.

Completing something can add to your world. Missing something never takes anything away.

Three principles shaped every decision in Quest.

Design philosophy

No broken streaks, red warnings, or "you're behind." Time away doesn't become another thing to feel bad about.

An unfinished task is a signal to make the next step easier, not increase pressure.

Grow, never shrink.

Restructure, don't punish.

Softness by default.

Completing something can add to your world. Missing something never takes anything away.

Three principles shaped every decision in Quest.

Softness beyond the core flow

Notifications default off, unfinished quests quietly carry over, and reminders never call attention to missed days.

Softness beyond the core flow

Notifications default off, unfinished quests quietly carry over, and reminders never call attention to missed days.

A wrong turn worth mentioning

The earliest version of Quest leaned harder into traditional game mechanics — boss battles, visible HP, losing HP for missed tasks.

Building it exposed the contradiction directly: I was designing for people who already feel bad about unfinished tasks, while building a system that punished them for unfinished tasks.

I removed the loss condition. The game remained. The punishment didn't.

Before

After

Coming back, without dread

Most productivity tools are designed around getting someone to complete a task.

I was more interested in what happens when they don’t.

Quest treats returning as a continuation, not a failure state.

Home, whether it's been a day or two weeks

No counter, no color change, no sense of time passed — a task that's been sitting for two weeks looks the same as one from yesterday.

Home, whether it's been a day or two weeks

No counter, no color change, no sense of time passed — a task that's been sitting for two weeks looks the same as one from yesterday.

Opening the stalled task offers a choice, not a restart

Continue, make the next step smaller, or leave it for today. None is framed as the “right” answer.

Opening the stalled task offers a choice, not a restart

Continue, make the next step smaller, or leave it for today. None is framed as the “right” answer.

Choosing "Not today" doesn't fail the task, just lets it rest

It sinks to the bottom of the list with a soft "Resting" tag — present, unhidden, but no longer demanding attention. It'll be gently re-offered later, never nagged.

Choosing "Not today" doesn't fail the task, just lets it rest

It sinks to the bottom of the list with a soft "Resting" tag — present, unhidden, but no longer demanding attention. It'll be gently re-offered later, never nagged.

Research changed the design

After

Acknowledges the progress already made and builds on it.

Before

Options feel more prescriptive and imply restarting

A note on sequence

Quest’s core philosophy came first, shaped by personal experience and observation. After the initial product was designed, I ran a small exploratory questionnaire (10 responses) and one longer 1:1 interview to pressure-test assumptions already built into it — not to validate them statistically.

Some findings reinforced decisions I’d already made. Others complicated them. One changed an actual screen.

What held up, what got complicated, and what changed.

HELD UP — Reward without loss

The strongest response closely mirrored Quest's core hypothesis: points sounded motivating until losing them entered the picture. Reward created interest; punishment created a reason to disengage.

HELD UP — Avoidance isn't forgetting

Participants described already knowing — and often thinking constantly — about the thing they were avoiding. That reinforced my decision not to treat reminders, overdue states, or increasingly urgent nudges as the solution.

GOT COMPLICATED — AI isn't universally welcome

One response specifically wanted several smaller-step options rather than a single prescribed answer. Another rejected AI assistance altogether. Together, they reinforced an important constraint: suggestions should remain suggestions, and manual control needs to stay equal.

CHANGED — My step-splitter contradicted the message

In the interview, the participant said a smaller-step suggestion could help — as long as it didn’t feel patronizing or ignore the progress she’d already made.

Looking back at my step-splitter, it did exactly that. “You stalled on” framed the moment around failure, while “Open a blank draft — nothing else” assumed she was starting from zero.

I changed the language to acknowledge what had already happened: “You already started” and “Look at what you have, then add one sentence.”

The mechanic barely changed. The emotional message did.

Still unresolved: Tone is personal

This wasn’t a universally “better” version. One later response preferred a more professional, understated tone and found more encouraging language patronizing.

I chose to move forward with language that acknowledges existing progress, while treating tone as a preference rather than a solved problem.

If I took this further, I’d explore whether tone itself should be customizable — letting people choose how Quest speaks to them.

A note on sequence

Some findings reinforced decisions I’d already made. Others complicated them. One changed an actual screen.

What held up, what got complicated, and what changed.

HELD UP — Reward without loss

The strongest response closely mirrored Quest's core hypothesis: points sounded motivating until losing them entered the picture. Reward created interest; punishment created a reason to disengage.

HELD UP — Avoidance isn't forgetting

GOT COMPLICATED — AI isn't universally welcome

CHANGED — My step-splitter contradicted the message

In the interview, the participant said a smaller-step suggestion could help — as long as it didn’t feel patronizing or ignore the progress she’d already made.

Looking back at my step-splitter, it did exactly that. “You stalled on” framed the moment around failure, while “Open a blank draft — nothing else” assumed she was starting from zero.

I changed the language to acknowledge what had already happened: “You already started” and “Look at what you have, then add one sentence.”

The mechanic barely changed. The emotional message did.

Still unresolved: Tone is personal

This wasn’t a universally “better” version. One later response preferred a more professional, understated tone and found more encouraging language patronizing.

I chose to move forward with language that acknowledges existing progress, while treating tone as a preference rather than a solved problem.

If I took this further, I’d explore whether tone itself should be customizable — letting people choose how Quest speaks to them.

One system, four worlds

The mechanics stay consistent. The world is yours to choose.
Quest uses one shared system across four visual worlds — letting people choose the atmosphere they want without changing how the product works.

The avatar answers back

Same three triggers in every world: nothing started yet, a boss battle in progress,
and a boss just defeated. The avatar's pose and expression shift — never a
popup, never a number. Just a companion that visibly feels what you're going
through.

AI, with less authority

The hardest part of “make this smaller” is that it asks for more effort at the exact moment someone is already stuck. Quest explores using AI to offer a first attempt instead — generated only from the stalled task’s text.

AI reduces the effort required to begin. It doesn’t decide what someone should do.

Suggestions arrive as a starting point, not an instruction

Every suggestion can be used as-is, edited, or skipped individually — none are forced onto the task list untouched.

Suggestions arrive as a starting point, not an instruction

Every suggestion can be used as-is, edited, or skipped individually — none are forced onto the task list untouched.

Writing your own is never buried behind the AI option

The manual path sits at equal visual weight, not as a small "advanced" link — some people will always know their own smaller step better than a suggestion could.

Opt-in, and plainly explained, in Settings

Off by default, plainly explained, and scoped to the one line of text it actually needs — not the whole task history.

Opt-in, and plainly explained, in Settings

Off by default, plainly explained, and scoped to the one line of text it actually needs — not the whole task history.

One thing i haven't solved

When does “Not today” mean healthy self-pacing — and when does it become avoidance in nicer packaging?

I still don't have a reliable way to distinguish someone gently self-pacing from someone quietly disengaging. Quest can re-offer resting tasks, but adding reminders, pressure, or behavioral signals to detect the difference risks recreating the exact mechanics the product is designed to avoid.

For now, I'm treating that uncertainty as an accepted tradeoff — not a gap to paper over.

When does “Not today” mean healthy self-pacing — and when does it become avoidance in nicer packaging?

For now, I'm treating that uncertainty as an accepted tradeoff — not a gap to paper over.

What I'd test

When someone stalls, does offering a smaller next step make them more likely to return than simply reminding them about the original task?

Then I’d test two things around it:

Tone — When does encouragement feel motivating, and when does it start to feel patronizing?

Re-entry — Can a resting task be gently re-offered without becoming another reminder to ignore?

When someone stalls, does offering a smaller next step make them more likely to return than simply reminding them about the original task?

Then I’d test two things around it:

Tone — When does encouragement feel motivating, and when does it start to feel patronizing?

Re-entry — Can a resting task be gently re-offered without becoming another reminder to ignore?

Quest started as an experiment in gamifying productivity.
It became an experiment in designing motivation without making failure the mechanic.