Redesigning a learning platform into a modular workspace
OVERVIEW
12 months, 180°
When I joined the company, Naviri was entirely focused on e-learning, trying to teach founders how to build a business. However, this direction wasn't truly what they wanted to build—nor what the market needed.
After a year of iterative research and design, we transformed it into a modular workspace platform that helps users think and work more effectively. The results? More freedom, more flexibility, and most importantly: more value.
PROBLEM(S)
Not knowing what we (and others) want
The original vision didn't translate into a product users wanted. The initial features were built based on assumptions rather than actual needs, which resulted in low engagement and unclear value proposition.
My contribution consisted in transforming the approach by establishing a user-centered design process, conducting interviews, and validating every decision before building.
PROCESS
Iterative prototypes, iterative prototypes, iterative prototypes...
Rapid iteration was key to finding product-market fit. Every week was made of user interviews, prototype testing, and validation cycles.
Over time, we slowly moved away from structured learning to reach a flexible modular workspace.
1. First Prototype: "Project Map"
The first innovative prototype after my arrival organized business development into a digital canvas with 7 structured learning phases. This brought a lot more clarity to users, who quickly understood what the tool was capable of.
However, the problem lied elsewhere: it still wasn't what the market truly needed. After organizing a startup bootcamp and testing with dozens of users, we realized it was time for a change: we needed to move away from learning asap.
2. Building the Data Model
After weeks of ideation, because we wanted to harness the power of AI and to apply it to business development, we came up with a solution to the main weakness of other tools: the lack of context.
To build the most accurate and reliable AI, we designed a custom data model built around small, structured company data, which connect to each other to form a living knowledge graph of the user's business.
3. Second Prototype: Driven by Data
Because of how much data-driven the product had become, it was important to update the visuals for the next prototype. On top of that, since we really believed in the power of the knowledge graph, I kept the canvas in this iteration.
We also added extra features such as pitch deck export, data room view and AI-driven analysis. However, after a second bootcamp with founders, we discovered a major pitfall in our product: it was way too restrictive.
4. Interviews and Flows
At that point, months had already passed, and we were pressed for time. Instead of building a complex prototype again, I took more time to engage in user interviews, and thoroughly re-built the main user flow.
What came out was critical for how the platform should work: the data model was still a great idea, but users needed freedom, not a treadmill asking them to fill data. In a way, we were still an e-learning product.
5. Rebranding and Web Presence
I then realized the branding felt off, and that we never took enough time to fix it. Because the next step would be to try and convince investors and early adopters, now was a great time to take a week to rebrand Naviri.
This massive overhaul included brand guidelines and a whole website redesign. The reasoning was to reflect how safe, reassuring and modular Naviri really was: all great qualities for your guide in the arduous startup journey.
6. Final Prototype: AI-backed Freedom
Times were frustrating. We knew how efficient our data model was, and how accurate the AI would be... if only users took time to fill in their data. But we were all aware of how no one in their right mind would do that. Also, not everyone had the same questions or needs, and we weren't covering every startup stage.
It turns out the solution was right in front of our eyes: whenever the user asks something, AI can break it down into simple, small tasks and fill the relevant company data itself—no matter the question. After designing an advisor view to sell to incubators, we finally reached our goal: guided freedom.
SOLUTION
Creating a proper startup OS
The final product had to reflect how variable and flexible a business is. It had to adapt to any kind of founder, at any stage of their business journey. That is why it uses AI to answer any question but still bases it on its own data model.
While company data became more and more hidden in the backend, the AI analysis were moved to be a centerpiece of the interface, and a goal for users to strive towards: making their business as good as possible. All, thanks to Naviri.
RESULTS
A platform that suits users' real needs
While development is still in progress, interviews have proven how eager users are to test the real product. Personally, I can't blame them: a startup OS analyzing your current company and helping you improve its future version is invaluable.
Although my time at Naviri has come to an end, I wish them the best for what's to come, and I am really proud of how much I contributed to evolving the product's direction. Because of my responsibilities, I also learned a ton in a single year.
TAKEAWAYS
What I learned
Talk to users early and often
Building based on assumptions wastes time and resources. Constant user research and validation prevents from investing in the wrong features and helps find product-market fit faster.
Don't be afraid to pivot
Attachment to ideas is a product killer. If data proves it, completely changing direction, even after significant investment, is critical for success. As Shane Parrish claims in Clear Thinking, "outcome over ego".
Take time and think about everything
A real product has so many screens and edge cases to account for: error states, empty (or overwhelming) content & unexpected user flows to name a few. But it is my duty to cover all of them.
