Modernizing an Established Property CRM with AI, One Working Prototype at a Time

The Challenge

Nekst had already won the hard part. Their property CRM was well established, with a large base of real estate professionals relying on it every day to manage transactions. That success was also the constraint.

The product needed to move forward. Screens built years ago felt dated next to modern tools. AI had opened up new ways to speed up the everyday work their customers do. But you don't get to experiment freely with a CRM that thousands of people depend on.

The real problem wasn't building. It was deciding.

• Modernizing established screens means touching workflows customers already know. Change the wrong thing and you make a working product worse.
• Layering in AI isn't a feature you bolt on. It has to fit naturally into how people already work, or it just adds noise.
• Static mockups couldn't answer the question that mattered: how will this actually feel to use?

The founder needed to see and feel the new direction, across new screens, AI-assisted flows, and redesigned classics, before committing his engineering team to building any of it.

The Approach: Prototype-First

Instead of designing screens in Figma and asking the team to imagine them working, we built a working prototype. A real, clickable version of the product where new ideas could be tried, felt, and thrown out fast.

The goal was to move through options quickly. Try a direction, click through it, react to it, and keep the ones that felt right. When you can put a real interaction in front of the founder, the conversation changes. It stops being about opinions on a picture and becomes about how the product actually behaves.

Working this way meant weak directions died in days, not sprints. Nobody spent weeks building something that only revealed itself as wrong once it shipped.

Layering in AI Where It Actually Helps

AI is easy to add and easy to add badly. The prototype let us find the moments where it genuinely earns its place in a real estate professional's day, rather than sprinkling it on for the demo.

We prototyped AI as real behavior inside the workflows Nekst customers already use:

• In the flow, not beside it. AI assistance shows up where the work happens, not as a separate chatbot bolted onto the corner of the screen.
• Tested by feel. Because these were working interactions, the founder could judge whether they helped or got in the way, which is the only judgment that matters.
• Right-sized. Some ideas felt great in the prototype and stayed. Others looked good on paper and got cut once we actually used them.

A Style Guide That Works With AI

Exploring screens was only half the job. The other half was making sure the direction could be extended long after our engagement, by both the Nekst team and the AI tools they build with.

So we established a modern style guide and component system, structured deliberately so that humans and AI read from the same source of truth. New sections can be built consistently without reverse-engineering someone else's design decisions.

This is the difference between a redesign that ages the moment we leave and a system the team keeps building on. The style guide turns a one-time refresh into a durable way of working.

How We Worked

The whole engagement was weighted toward working through uncertainty with the founder, not producing deliverables against a fixed spec.

• Explore fast. Multiple directions tried in the prototype, so choices were made on evidence.
• React together. Regular click-throughs where the founder could feel the product and steer, rather than approve flat comps.
• Converge. The strongest directions hardened into a clear product direction the whole team could see and agree on.
• Establish the system. Everything that survived got captured in the style guide so it could scale.

From Prototype to Dev Handoff

Once the direction was resolved in the prototype and captured in the style guide, the expensive part, production, could move with confidence.

The prototype wasn't a throwaway. It was the specification. The Nekst engineering team took a founder-approved direction, backed by a real component system, and implemented it knowing exactly what they were building and why. No guessing, no rework from decisions that should have been made earlier.

That's the whole point of working prototype-first. The judgment happens upstream, while it's still cheap to change your mind. By the time it reaches development, the hard calls are already made.

What Made This Work

Respect for an established product. This wasn't a blank canvas. Real customers had real habits. We modernized without breaking what already worked, because we could test changes against the current product before committing to them.

Prototypes turned opinions into decisions. A founder reacting to a working interaction gives you clearer, faster direction than a founder reviewing a static mockup. Feel is hard to argue with.

AI treated as craft, not garnish. Prototyping the AI interactions meant we could tell the difference between AI that helps and AI that's just there for the pitch. Only the useful survived.

A system built to outlast us. The AI-ready style guide means Nekst keeps moving after handoff. We didn't just hand over screens. We handed over a way to build more of them.

Product Redesign + AI / Real Estate Technology

“From start to finish, Francois and his team are able to deliver a well thought out, expertly designed and well built product in such a short amount of time. His team takes care of a lot of the little details and just knows the right way to build software. Highly recommended!”

Nekst is a clear example of prototype-first product redesign paired with an AI-ready design system: explore the options in a working prototype, let the founder feel the direction, establish a system the team can extend, then hand a resolved decision to engineering.

Have an established product that needs to modernize without betting the roadmap on a guess? Schedule a discovery call and let's work through it prototype-first.