In the year 2001, my dad hired an architect to design our new house. In one of those early meetings with the family, the architect mentioned the word master bedroom (I know recent developments in DEO that call for getting rid of the word master but the term master bedroom was so common during those times). They knew that unlike in some of their other projects with other clients, this master room was not meant for the my dad—it was meant for where the family spends time together.
We as a family had our own understanding of the terminology of our past homes and we were carrying the same notion while designing this new house, and our defaults were specific. The architect sensed it and they could plan it accordingly—it was not a template—neither for designing and nor the vocabulary.
Their intelligence was not only in the architecture, or budgeting, or in designing the models. It was in quickly adapting to our terminology, our beliefs, and our defaults as a family. It is as much about the confidence as it is their specialized skills.
A couple of weeks ago, I saw an architecture consultancy firm using AI to gather a client’s requirements for designing a new home. I am not sure how AI will learn about the family, the legacy, the occupants’ vocabulary and their sensibilities when they move around in the house.
A family has some food habits, a navigation systems how people move around, their preferences to move in specific situations, their pets, Netflix moments, entertaining guests for a dinner, or for hosting school going kids’ parties. Most of these are undocumented practices—it is the job of architects to sense these and discuss, and keep some fine margins for the flexibility in how those margins change with time.
AI has the capability to make positive impact on the architecture practices—designing, rendering, and of course workflow efficiency and speed.
Terminology and shared language are as important in the architecture agency work, as it is in any other industry and category. However, many agencies often either miss out on this opportunity to build the right learning models, or they find it too late and by then they have done more damage to at least some of their customers.
