Hotel in Trentino
Testing the balance between 3D, prompting and references to create consistent architectural visuals with AI.
- Year
- 2027
- Client
- intovisual internal
- Location
- Trentino, Italy
- Services
- Modelling, AI rendering, video
More and more architectural images are being created directly with AI. We wanted to see what that actually means in practice.
So we tried it ourselves, pushing an AI workflow to create a complete set of architectural visuals while keeping one thing under control: consistency.
Creating one convincing image is one thing. Creating a whole project that feels like the same place is another. We found that the key is not simply in the prompt, but in finding the right balance between the amount of detail built into the 3D scene, the information given through prompting, and the use of carefully selected reference images.
Too much detail can limit the flexibility of the AI, while too little can make it lose the architectural character. The same goes for prompting and references: each one can guide the image in a different way, but the real challenge is finding the right combination.
This project was an opportunity to test that balance, understand where each element has the most impact, and develop a workflow that allows us to use AI while still keeping control over the final visual language.
Let’s talk
Have a project to tell?
Tell us where you are in the process and what you need to present. Together, we’ll find the most effective visual solution.
Jonas Grümann