Cassi

Joy Sengupta

Team of 1  //  Cassi Home  //  2025

Owning a home means a hundred small decisions a year, and nobody obvious to ask about any of them.

Most home apps answer that with a dashboard. A dashboard is just the question, rearranged.

Cassi answers by doing the thing — quietly, and usually before you thought to ask.

A pre-revenue founder needed some concepts and prototype of his idea built out. In order for it to be even prototyped, the product needed to be figured out first and the idea polished. 1 year and 2 engineering teams later, I was able to design some concepts and prototype of what the product could be and how it can solve the homeowner’s problem we set out to solve.
Your home, handledSetting up a new home in CassiCassi's maintenance calendarDiagnosing a dishwasher issueAsking Cassi where the circuit breaker is

Every screen, not just the good ones.

Onboarding, uploads, the half-booked job, the fact you needed back in March. The in-between states are the product — everything else is a screenshot you cannot tap.

The splash screen
The welcome screen
The home dashboard
The operations overview
Every document, filed and searchable
An introduction to the financials view
Searching by address
The task list
An insurance summary at a glance
Operations overview, a second pass
Cassi, thinking it over
The splash screen
The welcome screen
The home dashboard
The operations overview
Every document, filed and searchable
An introduction to the financials view
Searching by address
The task list
An insurance summary at a glance
Operations overview, a second pass
Cassi, thinking it over
Welcome, screen two
The financials overview
A task not yet started, provider view
The voice assistant, mid-conversation
Welcome, screen three
Operations overview, a third pass
The task list, last summer's jobs
Financials, one step further in
Voice input, step two
Financials overview, a second look
Welcome, screen two
The financials overview
A task not yet started, provider view
The voice assistant, mid-conversation
Welcome, screen three
Operations overview, a third pass
The task list, last summer's jobs
Financials, one step further in
Voice input, step two
Financials overview, a second look

A few moments, up close.

The parts that do not fit on a phone screen: the flows, the pitch, and the states in between.

The states nobody screenshots

Uploading a document in five frames: parsing, almost-there, and the one field we could not read. Skip these and the prototype feels fake.

A season ahead

The maintenance flow hands over a seasonal checklist before the season turns, so the next right step is already waiting.

Talking to your house

Sometimes the fastest input is your voice. Describe the problem out loud, watch it listen, get an answer back.

Onboarding, mapped

The whole first run drawn end to end, from the home-intelligence score through to the first insurance upload.

Scored at a glance

Animated home-condition cards, each property scored so the comparison happens before anyone opens a spreadsheet.

The deck that carried the room

I also designed the final pitch deck of course. There were several versions of it, with quite a lot of varied pages that were used as per VC’s needs. This was designed in Figma and then later recreated in Google Slides (yes, it was a nightmare)

The questions I get asked.

I worked as a solo product designer (visual focus) and prototyper with the founder and 1 full stack engineer to test feasibility of the features.
Figma and Xcode (sometimes). Prototypes were made with Figma Make.

Well, most of our potential users were in the US and preferred meeting in person and showing us what issues they had. We also were interviewing with home maintenance companies to see what people deal with. Being remote, it was hard for me to see those, I had to empathize and rely a lot on secondary knowledge.

Also balancing the product which has the best visuals versus shipping the MVP was challenging, because the tech team wanted to cut down a lot of it, which led the company to rehire the engineering team a couple of times.

We shipped an MVP and the founder could raise $3 million first and $10 million later with that MVP, which was the whole point of this exercise.

One of our biggest mistakes was that we were trying to solve too many things at once — it would have been smarter to just focus on 1 problem that this solves and ship it.

Secondly, a lot of problems cannot be solved with tech and we have to accept it. There are regulations, people’s trust, ground level operational challenges that must be taken into account. Just an app cannot be a panacea.