AI Opportunity Assessment for Architecture Practices: What It Looks At

If you've read the story of what an AI transformation looks like inside an architecture firm and you're convinced it's worth doing, the next question is practical: what does an assessment actually examine, and what do you walk away with?

Here's the honest answer, without the generic checklist.

Why a generic AI assessment misses what matters here

Most AI readiness checklists are written for back-office businesses: invoicing, scheduling, customer email. An architecture practice has all of that, plus a layer of work that generic checklists don't touch: tender documents with strict formatting rules, coordination between consultants across disciplines, and compliance questions like the EU AI Act's labeling requirements for AI-generated imagery. A checklist built for a plumbing business doesn't know any of that exists.

The workflow categories an assessment actually examines

In practice, four categories come up in almost every architecture office we've assessed:

• Documentation. Meeting minutes, site visit notes, and project correspondence, usually written from memory or rough voice notes hours after the fact.
• Tender response. First-pass drafts of tender submissions, which follow a repeatable structure but currently get rebuilt from scratch each time.
• Site condition tracking. Turning site photos and rough observations into a structured condition report, instead of someone typing it up later from memory.
• Client coordination. Status updates and follow-ups that get delayed because writing them takes longer than the update itself deserves.

None of these are unique to any one firm. What's unique is how much time each one currently costs your specific team, and that's what the session measures.

Design-driven work vs. back-office work

The hardest part of assessing a design-driven business is separating the two kinds of work happening inside it. The creative core, concept design, spatial reasoning, client vision, isn't something an assessment should touch. The surrounding administrative layer, the paperwork that supports that creative work, is where the opportunity actually lives. A good assessment draws that line clearly and never suggests automating judgment. It only targets the assembly work sitting between a decision and the finished document.

What you get at the end, sized for a 10-25 person office

At the end of a half-day session, you get a prioritized list of workflows ranked by time saved versus effort to implement, not a generic tool recommendation. For an office this size, that usually means two or three quick wins that can start within a week, and one larger project, often team-wide documentation or tender response, that needs a proper rollout plan rather than a single afternoon of setup.

Where this connects to the fuller story

If you want to see what this looks like once it's running, not just planned, read what an AI transformation actually looks like inside an architecture practice. This piece is the "what happens first" companion to that story.