Concept to Clickable: How We Started Building Moola with No-Code

When we set out to build Moola Money — a financial wellness platform to help millennials make better money decisions — we had more ambition than engineering capacity. Like most early-stage startups, we wanted to validate our idea fast with customers, test real user flows, and show investors a working product without assembling a full dev team.

We leaned on AI and no-code tools to go from concept to clickable at speed and then brought in people at the right moments to turn prototypes into a product. Here’s how that journey unfolded across different dimensions of Moola’s journey.

Lovable: Prototyping the Financial Questionnaire

Our first milestone was creating Moola’s psychographic assessment, a guided flow that assesses users’ risk appetite, patience and financial confidence. Using Lovable, we designed an interactive prototype in hours. It was intuitive and beautiful, but there was one problem: we forgot to connect it to a database.

That oversight led to one of those classic startup woes. A test user with a product background pointed this out, and as the survey was already with other test users, we quickly vibe coded a Supabase table within 4 hours so that user inputs could actually be saved. It eventually worked, but it was a clear reminder that no-code tools could still overlook key architectural details.

After we had built our three separate customer input components in no-code, we brought in Marajul, our developer, to clean up the code and integrate these disparate components into a modern tech stack. He connected it with the other input modules and made sure everything ran smoothly under one roof. AI had helped us get the idea off the ground, and human hands gave it durability.

Replit: Turning Financial Models into Code

Before Moola had a product, we had a spreadsheet. Over the summer, our MBA intern Dylan, an ex-EY consultant, built ten end-to-end financial models for our alpha test users. Each model mapped income, savings, and goals to personalized recommendations, and the process of modelling was done over weeks per customer, over email back-and-forths to clarify details and in a GSheet with output charts generated.

When we tried to convert those sheets into real product logic, we got stuck. The formulas were complex, and the translation to code wasn’t straightforward. Then we discovered Replit, another vibe coding environment that made it possible to prototype backend logic and write test scripts. Replit even uses multiple agents with different roles - software engineer, data architect - to build and test code autonomously.

With Replit, we could take Dylan’s GSheets, extract the logic, and build executable models that mimicked the same financial reasoning. We could also create an isolated testing environment based on our existing code exported from GitHub, so that the code would fit better into our base. It became the bridge between Excel-based consultancy and scalable technology.

We also brought in Fil, our head of product with expertise in building fintech products for banks during his time at McKinsey, to lead the build of the financial algorithm into our codebase. His combination of deep financial understanding and product expertise was invaluable to bridging the technical and non-technical product development aspects.

Learning to use Replit crystallized something important for us: AI and no-code tools can speed up iteration, but the team still needed domain expertise to give the algorithm meaning.

Midjourney: Finding the Visual Language

While the product logic took shape, we turned to Midjourney for branding and stock image generation. We wanted Moola to feel approachable but trustworthy, the kind of financial platform that felt like a friend, not a bank. Midjourney helped us experiment with colours and general visual direction (orange and navy blue emerged early as our signature tones) and gave us a quick visuals for our website and pitch decks. We used AI-generated stock-like images to communicate the brand in our early pitch decks and website images.

As the need for content, website and product visuals increased, we brought in Heidi, our brand & marketing lead. She translated our early Midjourney explorations into a cohesive visual system, refining typography, composition, and tone, and created systems for prompts to ensure creative assets supported marketing efforts. The key to finding assets that aligned with our vision and built trust was ensuring our creative brief and prompts were informed by strategy, taste, precise art direction and continuous refinement. With this strategic framework locked in, we turned to tools such as Midjourney, ChatGPT-4o, and Claude.ai, Canva AI, Nano Banana, Relume.io, Figma Design/Make, Notion AI to create workflows, mixing in photography and stock imagery to rapidly explore and validate and create assets. The tools allow us to create efficiently and provide valuable content with a lean team – AI-enabled but with a human defining strategy and staying in the loop. We curate imagery incorporating authentic human elements for pitch decks and landing page prototypes and marketing collateral, allowing us to test our brand with users before committing to shoots or design sprints. We are constantly testing new tools to find the best fit for our needs.

AI creative tools are powerful accelerators, guided by human insight. Our competitive research and user interviews created the strategic foundation and AI helped us visualise and validate it at speed.

UXMagic.ai: From Demo Day to Design Reality

At 500 Global’s Demo Day in Silicon Valley, I met the founding team behind UXMagic.ai, an AI tool that generates interactive prototypes directly from prompts. Within minutes of seeing it, they solved a UX problem I’d been wrestling with for weeks.

Back at Moola, we used UXMagic.ai to whip up new screens and flows, especially helpful for those of us without Figma or UX backgrounds. The tool could even export directly to Figma through a simple copy-and-paste function, saving days of design work.

Still, we found limitations: while the mockups were fast and functional, editing them inside Figma was still cumbersome and required a specialized skillset. It was often easier to treat the AI-generated prototypes as inspiration and rebuild them from scratch in our stack.

The key takeaway: AI can now help to rapidly visualize ideas, but true usability still comes from iteration and user testing.

When We Brought People Into the Mix

Each tool helped us leap forward, but Moola only came to life once we blended automation with human collaboration:

• Marajul (Developer): integrated our no-code prototypes and experiments in Lovable, Replit and UXMagic into our own stack and codebase.

• Dylan (MBA Intern): translated manual spreadsheet modeling into reusable financial logic, transitioning from GSheets to Replit.

• Fil (Head of Product): bridged the technical and non-technical aspects of product development, supplemented with his own financial expertise, to deliver the financial algorithm.

• Heidi (Brand & Marketing Lead): refined the mid-journey aesthetic into a cohesive and human visual identity and used creative production tools to streamline our content workflows.

  
  

Together, they bridged the gaps between quick experiments and sustainable systems. What started as a collection of AI prototypes is evolving into a functioning product built in months, not years.

Lessons Learned

Here’s what we learned from going from concept to clickable:

• AI is a great accelerator, not a replacement. It removes friction but still needs human intent.

• No-code doesn’t mean no logic. You still have to think like a systems builder and bring in the domain expertise (in our case, financial modelling).

• Hire for leverage, not volume. We brought people in to improve the quality of our product and brand.

• Show, don’t tell. A clickable prototype inspires more confidence than a deck.

• Build to learn. Our biggest priority during build has always been to stay as close to our customer as possible, and to test with them and learn from their feedback.

  
  

If you’re a founder experimenting with AI and no-code, we’d love to hear your story. You can learn more about Moola Money at www.moola-money.com.