How AI changes the design process

Managing 100 constraints at once

One of the most underappreciated capabilities of AI in architecture is constraint management. A residential project involves dozens of overlapping requirements: budget, client preferences, municipal regulations, provincial and federal codes, construction constraints, site limitations, and more.

"AI can hold about 100 constraints simultaneously: budget, client preferences, municipal, provincial and federal regulations, technical construction constraints. When I change something, it flags what's no longer within the agreed parameters."— Stéphane Turbide

In traditional practice, keeping track of all these constraints is a manual, error-prone process. An architect reviews a change, mentally checks it against the relevant rules, and sometimes misses a conflict that only surfaces weeks later (often at a much higher cost to fix).

Two concrete examples illustrate this. First, budget: if a client asks about adding a garage, AI can immediately flag the cost impact (typically $30,000 to $50,000) and show what changes downstream if the garage is added or removed. Second, site constraints: if a client gets excited about extending the façade during a design session, AI flags that the setback margin would be exceeded, even in the middle of a fast moving conversation. No one has to remember the rule. The system catches it.

Exploring more options, faster

The traditional process forces a bottleneck in exploration. A client arrives with Pinterest folders, inspiration images, a list of needs (bedrooms, bathrooms, views), and municipal regulations. Without AI, the architect takes all of that away, analyzes it, and comes back for a second meeting with a first sketch: basic block diagrams showing room placement and circulation.

With an AI floor plan generator, that entire exploration phase can happen during the first meeting.

"With AI, I can process all those constraints in real time during the first meeting and generate multiple block diagrams in minutes, rather than presenting just one option."— Stéphane Turbide

This changes more than speed. It changes the quality of the conversation. Instead of proposing one layout and hoping it lands, the architect can show several directions and explain why certain approaches won't work. That transparency matters because clients often feel their needs aren't being considered when options are rejected without explanation.

What makes a floor plan actually good

AI can generate layouts quickly, but what separates a functional plan from a great one? According to Stéphane, the answer is less technical than most people think.

"When someone walks into a home and says 'wow, it's beautiful here,' that's hard to explain, but it's almost always tied to circulation and the emotions the space creates."— Stéphane Turbide

A good plan uses the least space possible while being optimally organized: less impressive rooms tucked away efficiently, more impressive spaces front and center. In Québec specifically, people love gathering around a kitchen island, open concept living, and views. Orienting the kitchen so you can see the front door when guests arrive is a deliberate emotional design choice, not just a layout preference.

This is exactly where AI has made the most progress, and Stéphane has watched it happen firsthand.

From blocks on a screen to real architecture

Stéphane was there when Maket produced one of the first AI-generated architectural plans. His initial reaction was purely technical.

"It was impressive that it could understand constraints and generate something, but it was placing blocks just to place them, without asking whether it was beautiful or functional."— Stéphane Turbide

That was the starting point. Since then, the leap has been dramatic. AI-generated plans now look like professional architectural drawings with correct nomenclature (doors, windows, walls), and the system understands circulation and spatial emotion, not just proximity between rooms. It went from solving a puzzle to designing an experience.

The budget problem no one talks about

One of the most common issues Stéphane encounters isn't technical at all. It's the gap between what homeowners imagine and what they can afford.

"Their vision is almost always bigger than their budget, often double. They see things on Pinterest or Instagram that look great and want to reproduce them, but the more detailed you get, the more expensive it becomes."— Stéphane Turbide

AI helps here in a way that traditional processes can't. Because generating and modifying layouts is fast, the architect can walk a client through realistic alternatives in real time. "You want three bedrooms? Let's rethink what that actually means for your family's daily routine." Sometimes the best design solution isn't giving the client exactly what they asked for. It's reframing the question entirely.

This extends to the broader housing crisis. Architect fees are actually a small fraction of total project costs (typically $3,000 to $4,000 on a $600,000 build). The real cost drivers are construction materials, land scarcity, and slow permitting processes. AI's biggest potential impact isn't reducing architect fees. It's streamlining the full process so each step communicates with the next, reducing delays and waste across the board.

Where architects fit in the future

With AI handling more of the process, the natural question is what's left for architects. Stéphane's answer is more optimistic than you'd expect.

"AI will probably handle 60 to 80% of the process, which would be wonderful. The last 20%, for what is the biggest investment of most people's lives, you'd still want a human to confirm you've made the right choices."— Stéphane Turbide

But the picture is more nuanced than a simple percentage. Stéphane sees two tiers of architecture emerging.

The first tier is accessible architecture: a young couple building a home simply to have a roof over their heads. In that context, an architect doesn't add enormous value. That's where architectural technologists and AI tools come in, making decent design available without the cost of a full architectural engagement.

The second tier is what he calls "architect's homes": projects where someone genuinely wants a well-thought-out design, deeply connected to the site and their lifestyle. In that space, the architect's value is irreplaceable.

"What I'd love is for all architects to move toward that second tier, using AI to bring their fees down from $40,000 to $50,000 to $2,000 to $4,000 while still delivering that quality. AI won't replace the architect there. It'll make great architecture accessible to more people."— Stéphane Turbide

That's the future Stéphane sees, and it's already taking shape through initiatives like Maket's partnership with Lib Work in Japan. Not fewer architects, but architects working on better projects, spending less time on repetitive layout work and more on the deeply human parts of design that clients are willing to pay for.

The 80% problem most homeowners don't know about

About 80% of people who search online for "how to make a house plan" end up on plan-selling websites. They buy something inexpensive, but when they start customizing, fees skyrocket.

The money isn't saved at the first step (buying a generic plan). It's saved in the experience itself: using AI tools to explore and personalize from the start, rather than trying to retrofit a template that was never designed for your specific needs, lot, or budget.

Tools like Maket let homeowners, contractors, and architects generate personalized layouts based on real constraints, so the first design is already tailored to the project. No template. No surprise customization fees. If you're planning a home project and want to see what AI can do before committing to anything, try Maket free and generate your first layout in minutes.

Conclusion

The future of home design isn't about choosing between AI and architects. It's a new process where AI handles constraint management, rapid exploration, and early stage layouts, while architects focus on the creative, emotional, and deeply personal work that makes a home feel like more than just a building.

For homeowners and builders, this means better design at lower cost, with faster timelines and more options to explore before committing. For architects, it means the opportunity to do more meaningful work and reach more clients without sacrificing quality.

The home design process has always been slow, expensive, and opaque. That's what's changing. And the people who will shape that future are the ones already building with both: professional expertise and AI.