Being an architect means living in constant tension. On the one hand, there is the design vision, the desire to create spaces that flexibly accommodate a wide range of human needs and integrate seamlessly into the urban fabric. On the other hand, our daily reality is burdened by rigid statutory constraints and bureaucratic struggles that often extinguish creativity. When establishing my practice, the goal was to break this cycle. Too much time was being wasted on the managing planning processes, and too little on the planning itself. 

Recently, a shift has occurred. To understand this shift, it helps to know how we operate at our studio. In STS, we design everything from private apartment renovations to complex, high-density urban renewal projects. Our core strength is navigating the friction between ambitious design, strict zoning laws , and developer requirements. We do this by embedding AI deeply into our daily workflow. Not just for rendering, but as an analytical partner to rapidly solve programmatic puzzles and test real-time building massing.

When generative tools first entered the studio, we were genuinely taken aback by its impact. Contrary to the panicked discourse that views AI as a substitute for the human creator, the experience was the exact opposite. Personally, I have  found a valuable partner - one that doesn't get tired of calculating building coverage ratios, one that allows the implications of every design decision to be seen the moment it comes to mind. 

The Ibn Gabirol Challenge 

To understand what this means in practice, consider an urban renewal project currently underway on Ibn Gabirol Street in Tel Aviv (Israel). It is a complex lot with multiple uses, a strict 10-story height limit, and a demand for a precise breakdown of usages, to guarantee that the project can succeed both architecturally and economically. For example, we needed to find the perfect unit mix considering the challenging building site and rigid regulatory restrictions. Using our generative workflow during a live meeting, we tested a denser mix of apartments and instantly proved it fit within the strict 10-story setback lines, solving a spatial puzzle in minutes that normally takes weeks of manual drafting

In the past, examining three or four alternatives for dividing the space within this statutory envelope would have required weeks of modeling and technical drafting. Today, the workflow has fundamentally changed. Spatial data and limitations  (such as that 10 story requirement) are fed directly into the computational system, allowing the technology to generate the digital scaffolding for proceeding with the design. Instead of wasting time redrawing building lines with every version change, a dialogue is managed by referring to the data. When testing how a shift from large apartments to a denser mix of smaller units will affect the building's facade, the result is visible in real-time. The transition from the dry model to an excel chart depicting the sizes of the apartments and live image happens almost instantly. 

A Return to Architectural Essence 

The true beauty of these tools is that they do not generate the architecture; they generate the infrastructure for making better architectural decisions. When AI takes on the Sisyphean task of adapting to planning laws, there is much more room to breathe and focus on what truly matters: the proportions of the balcony facing the bustling street of Ibn Gabirol, the interplay of light and shadow on the facade, and how the pedestrian experiences the ground floor. 

The profession is not facing its end, but rather a renaissance. The era where an architect had to first and foremost be a software technician is coming to a close. Today’s technology returns power to those with a design vision, a curator's eye, and a deep understanding of space. It liberates the process from technical limitations, providing a new, flexible canvas to create upon.