Generative Design
A design process where software generates and evaluates many possible design solutions against a set of defined constraints and goals, such as structural efficiency, daylighting, or material cost, rather than a designer manually iterating through options one at a time, presenting a range of viable options for a human designer to review and select from.
Why it matters
Generative design is only as good as the constraints and goals actually fed into it, a tool asked to optimize purely for structural efficiency without also weighing constructability or cost can generate technically valid options a contractor would find impractical or expensive to actually build, so the human review step remains essential, not optional.
On a real project
A structural engineer uses generative design software to explore dozens of possible column layouts for a building, each optimized for material efficiency within the architect's defined constraints, then reviews the generated options against real-world constructability before selecting one to develop further.
Who this matters most to
A Structural Engineer or MEP Engineer uses generative design tools to explore design options, but still applies professional judgment to select and refine the final design from what the software generates.
Where this goes wrong
A design team adopts a generative design tool's top-ranked structural layout without reviewing it for real-world constructability, trusting the software's optimization score as the deciding factor. A contractor's early cost estimate reveals the selected layout, while structurally efficient on paper, requires unusual formwork that makes it significantly more expensive to actually build than a marginally less optimal option would have been.