Arqgen · 2024 — present
Land Feasibility Platform built from the ground up.
A B2B SaaS used by some of Brazil's largest real estate companies. From three early clients in proof-of-concept to a product teams describe as essential to how they work.
Role
Product DesignerIndustry
Real Estate · B2B SaaSScope
0 → 1, ongoingWorking with
Engineering, PM, CSChapter 01
Opening
A product built from the ground up. I was there for all of it.
When Arqgen was joined in March 2024, there was no product to design for — not really. Three early clients in a proof-of-concept stage, a handful of journey mapping sessions already done, and a SaaS shell with basic components. The core problem the company was trying to solve was genuinely hard: land feasibility analysis in Brazil requires navigating over 5,000 different municipal zoning codes. Every new client meant a different city, different legislation, different setup. It was slow, manual, and impossible to scale.
The job wasn't just to design screens. It was to help figure out what the product needed to be.
Chapter 02
Making the invisible legible.
Two months in, the first significant contribution was structuring the logic of urban parameters — the zoning restrictions that define what can or can't be built on a given plot of land.
At the time, every new client required a long manual setup process tailored to their specific municipality. Weeks went into studying different zoning codes, interviewing clients about their workflows, and looking for the pattern underneath the variation. The goal was a generic, flexible structure that could accommodate any legislation — without requiring custom configuration from scratch each time.
That work eventually became the foundation for something bigger: clients can now build their own parameter templates autonomously, with calculation formats they define themselves. What used to be a bottleneck requiring product team intervention became a self-serve feature. A direct line from early research to a capability that changed how the product scales.
Chapter 03
Knowing who you're designing for — and who you're designing around.
The platform was originally built for technical users: architects and engineers running precise calculations. That's still the core. But over time, it became clear that the product existed inside a larger ecosystem of people with very different relationships to it.
Non-technical users needed results without friction. Managers and directors didn't touch the platform directly — but they were the ones evaluating whether it was worth the investment, based on what their teams produced with it.
Getting to that understanding wasn't a one-time research sprint. It was an ongoing relationship — biweekly calls with enterprise clients, sitting in on conversations not required to attend, listening carefully to what CS was hearing that never made it into a ticket.
The result: not just knowing the interface. Knowing the people using it, and the people it has to impress.
Persona 01
Architects & engineers
Persona 02
Non-technical operators
Persona 03
Managers & directors
Chapter 04
The decision that opened the platform to a new kind of user.
For most of the product's early life, importing a land file required a CAD format — standard for technical users, but a hard barrier for anyone else. KMZ is the format Google Earth exports. No architecture background required.
Implemented in December 2025, KMZ now accounts for 46% of all land imports. Adopted fast, used consistently. The platform moved from output to input. That's a different kind of trust.
Chapter 05
Process as cadence
When a dedicated PM joined, the way design work happened changed fundamentally.
When a dedicated PM joined, the way design work happened changed fundamentally. What had been reactive became a rhythm: discovery, research, direction, execution, refinement, validation. Not a rigid sequence — a cadence.
AI became part of that cadence too. Research synthesis happens faster, UI exploration covers more ground, and rough interactive prototypes built with AI tools have replaced a lot of the back-and-forth with engineering. The idea gets validated earlier, with less ambiguity on both sides.
The structure brought something more valuable than speed: it brought clarity about what not to build. A company restructure reduced engineering capacity, which forced more deliberation — and led directly to one of the more satisfying outcomes of the year: a full discovery process that concluded with a decision not to build a feature at all. The research showed it wouldn't solve the actual user problem. That's not a failure. That's the process working.
The difference shows in how the work lands. One client reported reducing a process that previously took three days down to ten minutes. That kind of outcome doesn't come from a single feature — it comes from a product designed, over time, to genuinely fit how people work.
Chapter 06
Where the product is today.
The platform now supports around 30 project configurations, slope and terrain analysis, detailed area reporting, and PDF export. Users can create parameter templates, manage team access, and share projects — autonomously, without involving the product team.
+95%
Monthly sessions, Oct → Apr
295 → 575
+340%
Solutions generated / month
209 → 932
46%
Of land imports use KMZ
Adopted in <5 months
4% / 32%
KMZ vs CAD error rate
More accessible, more reliable
What the numbers don't capture is what shows up in conversations. Across recent client calls, a pattern has emerged: teams describing the platform not as a tool they use occasionally, but as something that's become part of how they work. That kind of feedback is hard to manufacture.
Chapter 07
What's still unfinished.
The interface isn't as mature as the product deserves — a more considered redesign has been designed and is waiting for the right moment. A component library exists in a half-finished state that quietly demands attention.
The next frontier is ambitious: in-platform interactivity, frictionless onboarding, self-serve access — and a version of the platform that covers the full workflow of a non-technical user, end to end, without friction or workarounds.
The product has grown. So has the role. There's still a lot left to build.
* No confidential information disclosed. All details from public marketing materials and social media.
Next project · 02
Finapp — Financial Control App