AI in Architecture Practice with George Guida

AI in architecture practice is not a future theory anymore. It is already affecting concept design, visualisation, option testing, research, communication and the way practices explain value to clients.

George Guida’s Architecture Social conversation is useful because it avoids the lazy question of whether AI will replace architects. The better question is where AI removes friction, where it creates risk and what still needs architectural judgement.

Watch: George Guida on AI in architecture practice

George Guida’s episode is the best place to start because it connects AI to real architectural workflows: concept design, modelling, collaboration and how practices make decisions.

What George Guida brings to the conversation

The original episode introduced George through a route that moved from architectural training at Oxford Brookes and the Architectural Association into Foster + Partners, Harvard, computational design and machine learning. That mix matters because AI is not only a software topic. It touches design thinking, production, business models and education.

His work with xFigura points towards a more collaborative design environment, where teams can test ideas quickly without pretending the tool is the designer.

The practical AI question for practices

For practice leaders, the question is not whether every tool is perfect. It is whether your team understands where AI can help and where it should not be trusted without review.

• Can AI speed up early option testing?
• Can it help explain design intent to clients more quickly?
• Can it reduce repetitive visual or research tasks?
• Can it improve internal knowledge capture?
• Can the team still challenge the output with professional judgement?

Go deeper with Architecture Social

These related Architecture Social episodes add more context once you have the practical framework.

Listen: the full George Guida AI conversation

The audio version gives you the longer conversation around xFigura, education, experimentation and what architects should learn before AI becomes normal practice.

Watch next: generative design and AI tools

For a different angle on the same AI conversation, this Arka Works episode looks at generative design, algorithms and how architects can use computational tools without losing design judgement.

Where AI can help a practice now

• Concept imagery, mood testing and visual directions.
• Research summaries for planning, context and precedents.
• Early-stage massing or option comparison when paired with proper checks.
• Client communication, workshop preparation and briefing material.
• Internal process documentation and reusable templates.

What candidates should learn

For candidates, AI is becoming another way to show judgement. Saying you use Midjourney, ChatGPT or another tool is not enough. You need to explain what problem it solved and how you checked the result.

• Show where AI supported a design process without hiding your own thinking.
• Explain what you rejected, not only what you generated.
• Keep copyright, confidentiality and client data in mind.
• Use AI to improve process, not to disguise weak design evidence.
• Stay close to fundamentals: brief, site, users, technical constraints and communication.

How to talk about AI with clients

The client conversation matters. If a practice uses AI, it should be clear about what is exploratory, what is verified and what is not yet a design commitment. That protects trust.

• Label early AI images as exploration, not resolved design.
• Explain which decisions still need technical review.
• Keep human responsibility visible in meetings and presentations.
• Use AI to make the conversation faster, not to avoid difficult design questions.

Common mistakes

• Treating AI as a magic button rather than a tool that needs direction.
• Letting generated images outrun the brief, budget or technical reality.
• Uploading confidential client material into tools without permission.
• Using AI outputs without being able to explain the design logic.
• Ignoring the topic entirely because the hype is irritating.

Architecture Social view

Stephen’s recruiter view is that AI will sharpen the gap between candidates who can explain value and candidates who only list tools. The best people will use AI to move faster, then use judgement to decide what is actually worth keeping.

Next step

Start small. Pick one workflow, test an AI tool, document the result and decide what changed. Then use Architecture Social’s AI in architecture careers guide, live jobs and podcast archive to keep the learning practical.