Transforming AI-powered architectural design with Finch: Enhanced control, transparency and precision
As co-founder of Finch, I am deeply involved in the development of our technology and excited to share with you our latest iteration of the Finch AI-powered floor plate generator, Generate Floor Plate 2.0 — something we’ve been refining over the past year. This new release represents a significant leap forward, not just in the sophistication of the technology but also in how it empowers architects through enhanced control, transparency and precision in their design process.
When we launched the first version of our floor plate generator, Generate Building Plan 1.0, it marked an important milestone. However, the journey didn’t stop there. Our goal has always been to iterate quickly, addressing needs as they arise, improving the technology, and making sure it truly serves the architect’s vision. Hence we’ve been in close contact with our users, gathering invaluable feedback that has shaped the latest updates.
Introducing three key improvements
In this latest update, we’ve focused on three critical areas: giving users more control, making the algorithm more transparent and ensuring precise unit sizes.
1. Empowering design control
Architects want to be in the driver’s seat, guiding the design process rather than being guided by it. With this new iteration of Generate Floor Plate, we’ve significantly enhanced your ability to control the algorithm. For example, instead of specifying an exact number of stairwells or units per stairwell, you can now explore a range — say, three to five stairwells. The algorithm then finds the optimal solution within that range, giving you more flexibility and a broader search space.
This approach also extends to unit distribution per stairwell. You can now define a range of units per stairwell that aligns with your design intentions, such as avoiding corridor buildings where they’re not suitable. This level of control ensures that the algorithm’s output is closely aligned with your vision from the outset.
2. Enhancing transparency in the process
One of the biggest challenges with early AI tools was their lack of transparency. Architects would input their parameters and receive a result, but understanding how the algorithm arrived at that solution was often unclear. We’ve addressed this by making our algorithm more transparent, so you’re informed at every step of the process.
For instance, if the unit mix you’ve chosen doesn’t fit the building’s story, the algorithm will now inform you why. Perhaps the units are too large, or there are too many stairwells. This real-time feedback allows you to adjust your inputs and understand the trade-offs involved, rather than just rolling the dice and hoping for the best.
This level of transparency is crucial because it mirrors the way architects traditionally work — breaking down the problem and understanding every aspect of the design as it unfolds. By integrating this approach into our AI, we’re ensuring that you remain in control and fully aware of how your design decisions impact the final outcome.
3. Achieving precision
Precision is key in architecture, especially when it comes to meeting client briefs and adhering to building codes. In the past, our algorithm might have produced a 39 m² or 42 m² apartment when you asked for 40 m². While this might seem like a minor discrepancy, we know that accuracy matters.
With Generate Floor Plate 2.0, we’ve worked hard to ensure that the algorithm hits the exact sizes. If you ask for a 40 m² apartment, that’s what you’ll get, and any remaining square meters will be allocated to one unique unit per stairwell. This focus on precision not only better meets the needs and expectations of our users but also aligns with the practicalities of real-world projects.
Beyond these three main improvements, our algorithm has become much more sophisticated overall. It now considers factors like building depth and dual-aspect ratios, allowing for more nuanced design solutions. For example, if you’re working with a deep building and need dual-aspect apartments, the algorithm will recognize that smaller units might be challenging to fit in such a space and adjust accordingly.
This level of sophistication ensures that the algorithm’s suggestions are not just theoretically possible but practically viable, taking into account the realities of project criteria and compliance.
Mitigating risk. Embracing innovation.
We understand that there are natural concerns about using AI in design work — especially when it comes to matters of decision-making agency, transparency of process and accuracy of results. Our goal with this latest iteration is to mitigate these concerns while leveraging advancements in our AI.
At Finch, we’re committed to iterative improvement. All the updates we make are driven by two core principles: responding to feedback and continuously advancing the underlying technology. As architects ourselves, we understand the value and potential of these tools in daily practice. That’s why user-centric design is at the heart of what we do.
With our Generate Floor Plate 2.0 updates, we’re confident that you’ll find our floor plate generator to be not just a helpful tool, but an essential part of your workflow — one that keeps you informed and in control as you bring your vision to life.
Ready to experience the next generation of AI in architecture?
We’re currently onboarding architects who work with Rhino, Revit, and Grasshopper, focusing on multi-family residential projects. If you’re interested in learning more about how this technology can transform your projects, I’d love to connect.
Feel free to reach out via email at jesper@finch3d.com, or sign up for our waiting list to discover how your talents, supported by our AI, can result in an elevated work process — and truly exceptional designs.
/Jesper Wallgren
Co-founder and CPO
