Phase 1 – The starting point: any Street View
For this experiment I started from a real, familiar and easily recognizable urban context: a stretch of the Corso Francia service road in Turin.
I took a simple screenshot from Google Street View, as any designer would do in the preliminary study phase. No advanced settings, no processing: just a raw image, taken directly from the screen. And this is where the process began.
Step 2 – Cleaning the image
The first request to Nano Banana was to clean the image of all digital elements (icons, writing, watermarks, etc.). A necessary step to be able to work on the image as a design basis.
Prompt used:
“Remove all Google Maps overlay elements present in this image, maintaining a clean and realistic image as output”
Result:
Phase 3 – A first urban transformation
At this point I asked to transform the image into a first design scenario: the service road becomes a pedestrian area, with a new pavement and a green insertion on the left side which acts as a natural barrier. In just a few seconds, Nano Banana returned a credible and realistic image, useful for understanding an initial vision of the project.
Prompt used:
“Transform this image into an urban renewal intervention, adding a pedestrian area with gray paving and inserting perennial green flowerbeds on the left as a vegetal barrier.”
Result:
Step 4 – Add softness and detail
After obtaining the first version I opted to ask for a more fluid and organic version of the greenery, with soft lines and flowerbeds also on the right side, to accentuate the effect of the path integrated into the urban landscape.
Prompt used:
“Modify the greenery present in the image to create a softer, curvilinear and organic path, also act by adding new flowerbeds on the right side of the image”
Result:
This type of request shows how the variation of the environment and geometries can be guided quite precisely. A bit like having a very quick collaborator on the concept!
Phase 5 – Insertion of animated figures into the space
To bring the scene to life and better understand its scale and use, I asked Nano Banana to populate the image with some walking passersby, to bring more realism to the scene. This simple animation of the scene allows for better spatial reading and can also be extremely useful in the presentation phase to the public or clients.
Prompt used:
“Add some people walking in the scene to the image to animate the pedestrian space of the project.”
Result:
Phase 6 – The final technical table
Finally, I wanted to experiment with creating an architectural section drawn in an illustrated style, ready to be used in a design table. The result is surprising for its clarity and cleanliness. Once again a technical image generated from a real urban image in a few seconds.
Prompt used:
“Generate an illustrated section of the architectural project with the intervention of greenery and paving, real urban scale.”
Result:
Conclusions
Even in this test, Nano Banana demonstrates impressive flexibility: starting from a simple urban screenshot we arrived at an output that combines concept, visualization, narration and technical output.
Of course, it is not perfect in its technical output: some details are not precise (especially on buildings and elevations), and the section is more illustrative than millimetric, but if we evaluate the speed and ease of execution it is always impressive.
We’ll read about it in the next issue, with new experiments between AI and architecture. See you soon!
The weekly column “Architectural Prompting” is edited by experts Luciana Mastrolia, Giovanna Panucci and Andrea Tinazzo
>> If you are interested in these topics, also sign up to the free Linkedin Newsletter AI & Design for Technicians, we’ll talk about it here!
