Marco D’Ambrogio, an architect and Lead for AI and Digitalization at Sweco Architects, is a speaker at RELEASE [AEC] — the first tech event designed to help professionals stay on the cutting edge of innovation and master the tools of the future. The next edition will be held in Paris on October 20, 2026. The event is 100% free for AEC practitioners: register today!
In architecture, AI is usually framed as a shortcut — a faster way to generate images, automate repetitive tasks or streamline complex workflows. But this standard framing narrowly directs focus and understanding, obscuring some of the potentially more interesting shifts in our profession. The future of AI in our profession will not be defined by tools alone — tools are temporary. What endures is method. It is how we define value, how we organize knowledge and how we shape the process that leads to a meaningful built outcome.
When clients assess our work, they do not measure us by software subscriptions, plugins or the number of digital instruments in play; they value our deliverables, our insight and the quality of the space we create. That means we must begin with architectural intent and genuine value, then choose the most appropriate tools to support it. Starting with the tool and retrofitting the method around it is the wrong order.
A resilient methodology is never bound to a single platform. It is supported by software, and it can survive the next version or the next invention entirely. If we fail to shift our attention from tools to methodology, we risk entering what can only be called the “age of average.”
Tools Are Temporary, Method Endures
Generic AI tools can raise the floor of design. They make poor architecture less common. But because they draw from broad, distant and often uncurated datasets (potentially mixing a Midwestern office park with a Tokyo high-rise), they also lower the ceiling. Average quickly becomes acceptable, and good enough starts to feel complete.
The danger is not only mediocrity, but acceleration. When a tool can produce a rendering or a spatial layout in seconds, there is a temptation to accept the first answer rather than pause, question and refine. Speed begins to replace reflection, but in doing so, we lose the specificity that gives architecture its meaning.
Architecture is never generic. No two buildings are made in the same place, at the same time, for the same people. Architectural intelligence must therefore be adaptive, curated, and grounded in local context. To resist this age of average, we need to address a familiar paradox in the design process.
Frontloading Knowledge
At the very beginning of a project, when the greatest potential exists to shape something truly transformative, we know the least about it. By the time we understand the technical, environmental, cultural and operational realities in depth, the schedule has often narrowed, and the project has drifted toward standardized solutions.
Integrated technology should not be used to cut hours in a way that diminishes the profession — its real purpose is to frontload knowledge.
If we use technology to capture, connect and organize cross-disciplinary expertise early, we create time for reflection later. In practice, this means bringing structural, environmental and heritage expertise into the conversation on Day One, when there is still enough spatial freedom to solve problems creatively.
Collective Intelligence as a Resource
In my work leading digitalization, I see scale as a powerful creative advantage. When a practice can draw on thousands of colleagues across various markets and specialisms, it holds the power to stop knowledge from being scattered. With the right approach, it can be harnessed as a living resource. If that collective intelligence is structured early, it can shape better decisions from the start.
This is why AI should not be seen as an omnipotent designer. It is better understood as raw supercomputing power.
AI can operate within vast, multidimensional complexity. It can reveal hidden patterns, relationships and structural opportunities across historical project data that humans might miss. But that capability creates a new responsibility for architectural practices: we must curate our knowledge with far greater care.
Without structured, verified libraries, this power remains largely untapped. A tool is only as useful as the quality of the data and knowledge it can access.
Curated Libraries Instead of Raw Data
That is where the real work begins. We need internal libraries that are efficient, reliable and carefully maintained, where human experts select, test and validate the material that feeds our design ecosystems. There has never been a greater need for human authorship. The machine does not replace expertise — it depends on it.
When I speak at Release AEC this year, my challenge to the profession will be simple: we must redefine how we understand our digital assets.
Our data is not merely a collection of numbers or generic inputs. In architecture, data is not about tracking clicks or predicting the next advertisement. It is our collective expertise. It carries our values, our judgment, and our cultural, technical, and local knowledge. It is shaped by what we choose to include in the libraries we build.
We must demand more customization, more control, and more intentionality from the software developers who build the tools we rely on. That is how we protect creative agency, strengthen collective intelligence, and avoid settling for an automated future that is efficient, but average.
Marco D’Ambrogio, an architect and Lead for AI and Digitalization at Sweco Architects, is a speaker at RELEASE [AEC] — the first tech event designed to help professionals stay on the cutting edge of innovation and master the tools of the future. The next edition will be held in Paris on October 20, 2026. The event is 100% free for AEC practitioners: register today!
