2025-12-10
I’ve been trying to make sense of why design technology inside AE (Architecture and Engineering) firms feels the way it does. Not broken, but kind of… stitched together. A script here, a checklist there, a dashboard someone made two years ago that barely connects to anything else. We keep layering tools on top of tools, and somehow expect that to solve deeper problems.
What I keep coming back to is this:
AE is trying to build modern systems with a mindset that wasn’t designed for modern systems.
And I don’t mean that in a dramatic way..maybe a little. It’s just how the industry evolved. We train architects and engineers to think in terms of deliverables, not tech systems. So when firms try to adopt technology, they naturally approach it the same way: project by project, tool by tool, fix by fix.
But underlying real systems don’t work like that.
Something I’ve been thinking about a lot is that the design-tech architecture inside firms shouldn’t be built only by people who grew up in the industry. They should be involved, obviously. But you also need people who think in terms of modularity, platforms, ecosystems, and long-term maintainability.
Without that balance, the tech environment ends up exactly how it is today:
reactionary, fragmented, and inconsistent.
AE doesn’t have technology systems.
It has a lot of tools, but not much in the way of an underlying system. Those tools are often not connected, not coherent, and not designed to really grow.
That’s why workflows feel brittle. It’s not that the tech can’t support something better but it’s that the architecture behind the tech doesn’t really exist.
When I’ve built AE-type agents and modular architectures (primarily around programming languages), one thing I noticed showing up:
If you design something modular, you don’t have to predict everything.
You build what you know, and the structure supports what you don’t know yet.
Anyone can extend it.
Anyone can plug into it.
Change becomes normal instead of painful.
But that requires a certain culture: trust, empowerment, willingness to let people adapt systems to their own workflows. A lot of firms aren’t there yet.
Most firms try to keep knowledge alive by recording videos, writing documentation, or building searchable libraries. Those are fine, but they’re static. As soon as a process changes, the “library” just gets bigger, not better.
Agents force you to codify the process itself like the logic, dependencies, priorities, and context that make the work what it is. Once captured, it can evolve. It can adapt. It can tell you what matters and what doesn’t. It becomes a living thing.
That’s the part that feels modern to me but also more aligned with how people actually work.
Not endless how-to recordings,
but systems that actually reflect how the work happens.
This certainly requires a structured approach, and I think it’s a good recent example of what Calm Technology is aiming for. Technology that works even when it fails. Technology that stays in the background and only comes to the foreground when needed. Technology that adapts to the user, not the other way around.
I am not saying AI agents are the answer to everything. But they do represent a different way of thinking about how tools can work. Not as a static thing, but as a adaptable system that evolves with the work.
AE loves the idea of standards. We have so many “standards” that it basically means we have none. Every new rule creates new friction. Every “fix” creates a new layer of complexity.
But what if technology behaved more like the brain? Bear with me here. I get that might sound out there.
LLMs and agent workflows are the closest we’ve seen to tech that adapts with us:
I am not saying no structure but one where structure emerges instead of being imposed.
It’s a weird idea, but I think it matters. Again LLMs and agents aren’t perfect, but they hint at a different way of thinking about technology in AE. Not as rigid systems we have to bend around, but as fluid systems that adapt to how we work. I think with the right underlying structure and mindset, AE could start building tech environments that feel less chaotic and more like a natural extension of the work itself.
Design tech should feel like a natural extension of how people think: fluid, adaptive, responsive. Not something rigid we have to bend around.
I don’t think AE is there today. But I don’t think this is futuristic. We’re already seeing glimpses of it in LLMs, agents, adaptive interfaces, modular libraries, and how people are experimenting in tech.
One framework I keep coming back to is a principle from Calm Technology:
Technology should work even when it fails.
That’s the kind of environment AE needs.
Not perfect. Not brittle. Not over-engineered.
Just systems that support the work, evolve with the work, and reduce friction instead of creating it.
Maybe none of this solves everything. But it feels like the direction worth exploring.
At least, this is where my head’s been at lately.