Three ways to redraw the same app
Rekhon · design proposals · none adopted
The same 25 screens, drawn three times, each version arguing with the ones before it. No application code was changed to make any of them.
Everything on the documents page describes the app as it is. These do the opposite. They are three answers to one question — what should this app feel like to use? — and each is drawn against the ones before it, screen by screen, so the comparison is about shape rather than taste.
They use the same colours, down to the hex, on purpose. If two versions look different it is because they ask something different of you, not because one got a nicer palette.
None of them has been adopted, and none has changed a line of application code. Adopting any would need a written decision about the navigation and a re-recording of every screenshot the build checks. They exist to be argued with.
4 documents · 149 pages
Idea one · one answer per screen
PDF34 pages4.7 MB
All 25 screens redrawn so that each one leads with a single big figure and answers exactly one question.
Why I wrote it, and what is insideClose
- Why I wrote it
- The app had grown screens that tried to show everything at once. I wanted to see what happened if a screen was only allowed to say one thing.
- Like this
- Tidying a bedroom by giving every toy one box, then putting the box you actually want on top.
- What’s inside
- 25 screens redrawn, with the reason written beside every decision. It answers “how am I doing?” Not one line of app code was changed to produce it.
In the repo: docs/design/AI_Personal_CFO_redesign.pdf
Idea two · one queue of decisions
PDF45 pages7.9 MB
The same app drawn as a single queue of things that need you today, instead of a set of dashboards to read.
Why I wrote it, and what is insideClose
- Why I wrote it
- Nearly every money app hands you pages to look at. I wanted to try one that hands you a list to clear, and find out which I would rather open on a Tuesday.
- Like this
- Swapping a noticeboard for a to-do list you can tick off.
- What’s inside
- 31 screens, answering “what needs me?” It copies idea one’s colours byte for byte, so putting the two side by side compares their shape and not their paint. Argued against idea one on every screen.
In the repo: docs/design/AI_Personal_CFO_v3.pdf
Idea three · the instrument panel
PDF47 pages10.6 MB
The same app again, drawn the way the money apps Indians already open every day are drawn: dense, tabular, full of charts.
Why I wrote it, and what is insideClose
- Why I wrote it
- Kite, Groww and Coin have already taught this audience what a money app looks like. A design that ignores that vocabulary is asking its users to learn the job twice.
- Like this
- Writing in the handwriting everybody at school already reads.
- What’s inside
- 34 screens: all 25 the app ships, plus nine that those apps have and this one is missing. It answers “let me look.” Its closing section argues when its own approach would be the wrong one, and names the hybrid worth considering instead.
In the repo: docs/design/AI_Personal_CFO_v4.pdf
The map of idea three
PDF23 pages3.3 MB
The zoo map again, drawn for the third idea — a version of the app that does not exist yet.
Why I wrote it, and what is insideClose
- Why I wrote it
- A pile of mockups can look finished and still not join up anywhere. Asking the proposal the same question I had asked the real app was the cheapest way to find the holes.
- Like this
- Checking a new bus timetable actually gets you home before you buy the ticket.
- What’s inside
- Five tabs, 54 arrows and ten journeys — and which of those arrows a mockup actually draws, as opposed to which are still only lines on a diagram.
In the repo: docs/design/AI_Personal_CFO_v4_flow.pdf
What would have to happen to adopt one
More than redrawing. Every screen in the app is checked against a committed picture on each build, so changing the navigation means a written decision recording why, and every one of those pictures re-recorded in light, dark and at double-size text. That cost is the reason three proposals exist and none has been adopted: it is cheaper to argue on paper than to find out afterwards.