Case studies

From problem
to plan.

How we think through a product, from the first question to the plan. These are concept studies, not client work.

SaaS concept

Flowdesk

The challenge

Small teams keep their work in chat threads, spreadsheets and half-used task tools. Nobody can see what is moving and what is stuck without asking around.

How we'd approach it

  • Start from one question: what does a team lead need to see first thing in the morning?
  • Keep one source of truth for tasks, then let every view read from it.
  • Design the empty state first, because that is what every new team sees.

What we'd build

  • A dashboard that answers questions
  • Views for how people work
  • Quiet by default
Read the concept
Mobile app concept

Habitly

The challenge

Most habit trackers punish you for missing a day. Lose the streak once and the app starts to feel like a judgement, so people stop opening it.

How we'd approach it

  • Make logging a habit take a single tap from the home screen.
  • Reward coming back, not only perfect streaks.
  • Work offline first, because habits do not wait for a signal.

What we'd build

  • One-tap check-ins
  • Kind progress
  • Gentle reminders
Read the concept
Web app concept

StudioSite

The challenge

Marketing teams wait on developers for every small page change, while most website builders trade speed and quality for that independence.

How we'd approach it

  • Give people well-made blocks instead of a blank canvas and endless options.
  • Make brand rules part of the tool, so every new page still looks right.
  • Treat page speed as a feature that nobody can accidentally switch off.

What we'd build

  • Blocks, not pixels
  • Safe to publish
  • Fast by default
Read the concept
AI tool concept

BriefAI

The challenge

A good project starts with a good brief, but writing one takes time most people do not have. Vague briefs turn into vague estimates and long email threads.

How we'd approach it

  • Ask a few sharp questions instead of offering one big blank box.
  • Show what the AI assumed, so people can correct it instead of trusting it.
  • Keep the output as an editable document, not a chat transcript.

What we'd build

  • Guided questions
  • Visible assumptions
  • Ready to share
Read the concept
SaaS concept

ClientHub

The challenge

Most CRMs are built for sales departments. Small creative teams end up paying for features they never touch, and keep client details in their inbox anyway.

How we'd approach it

  • Build around projects and relationships, not sales pipelines.
  • Capture information from the tools people already use, with as little typing as possible.
  • Answer one question quickly: who do we need to get back to today?

What we'd build

  • One place per client
  • Follow-up reminders
  • Simple by design
Read the concept

Want this thinking
on your product?

Bring us the problem. We will come back with questions and a first plan.

Talk It Through