Why hasn’t AI transformed architecture the way it has law?

Partly it’s diffusion: a wave like this needs founders to focus hard on one industry. But the timing isn’t random. Law was disrupted first because it was the first profession whose core work matched what language models could do.

In 2023, Harvey and others saw that LLMs could already digest large volumes of text, which is legal review, and would soon draft structured, clear documents. Law was a huge industry, a natural fit, and a clear path to revenue. As agentic capabilities matured, these tools rebranded as “agentic operating systems,” expanding toward inbox management, expenses, and other admin.

An agentic OS for any profession has two layers:

1. Reducing administrative load. A proactive assistant that takes routine tasks and mental overhead off the professional. This only became buildable around the start of this year, which is why Instinct and Dots are getting attention now, and what we explored with Collie. On its own, though, it’s generic and not a strong wedge.
2. Doing meaningful parts of the actual work. This is where architecture differs. Before 2023, architects were arguably more software-savvy than lawyers. Today the reverse is true: lawyers have excellent AI-native tools, while architects still rely on traditional CAD, 3D, and rendering software built on conventional ML. That’s not because nobody tried. Until very recently, AI couldn’t do anything useful inside that stack.

The unlock is an AI that can operate architecture software directly and reliably enough not to break things, with all project knowledge under one roof. Over time, that harness could replace parts of the fragmented toolchain altogether. Arguably this became possible with Fable 5 (but at prohibitive cost) and became practical with Opus 5.5.

1. 2023Law: review and drafting
2. Early 2026Admin assistants
3. Fall 2026Architecture software

The core claim: In 2023 it was clear AI would transform legal drafting and review. Today it’s equally clear AI will transform architecture, and you couldn’t have built a meaningful AI-native stack for architects before this fall. That’s the window we’re building in.