Drawing QC as a Connected System

Drawing quality control has always been one of the most experience-dependent tasks in architecture. When I was younger, a senior architect or designer would review the full drawing set — not just for design intent, but for consistency, tagging, spelling, drafting standards, and constructability. They were checking the entire system, not just the drawings.

That process is becoming harder to maintain. Project timelines compress. Teams grow. The drawing set still gets reviewed, but the focus narrows to design. The smaller checks — annotation consistency, tagging logic, spelling errors, drafting standards — fall through. Not because people don't care, but because there isn't enough time to catch everything manually across a full set.

This is the kind of problem that systems solve better than individuals.

A rubric, not a replacement

We've been building a system-based approach to drawing QC. The idea is straightforward: take the same questions a senior reviewer would ask and run them consistently across every sheet in the set.

We use Autodesk Construction Cloud drawings and apply a structured rubric using AI vision. The system evaluates whether drawings are clean, aligned, and compliant with company standards. It checks for missing tags, inconsistent naming, drafting errors, and general coordination gaps.

The important distinction: this is not replacing the experienced reviewer. It's giving them a structured starting point. The system catches the repetitive, pattern-based issues so the reviewer can focus on judgment calls — design coordination, constructability, things that require real expertise.

From one-time check to feedback loop

The part that makes this a system rather than a tool is the feedback loop.

Each review is recorded as a structured dataset. When drawings are updated, the system runs again to validate that the flagged issues were actually resolved. Over time, this creates a clear record of what was found, what was fixed, and what still needs attention.

That's the difference between a QC check and a QC system. A check happens once. A system tracks state over time. It becomes trackable, repeatable, and measurable — and critically, it builds institutional knowledge rather than depending on it.

What this means for practice

This shifts drawing QC from a manual, experience-based task into a connected system of review and validation. It doesn't replace the role of experienced designers — it supports them. It ensures that important details aren't missed as projects move faster.

Good drawings aren't just about design. They're about clarity, consistency, and constructability.

The goal isn't just better models. It's better buildings.