From “black box” to co-pilot: Bricsys opts for transparent AI in design software

Transparent AI that really helps engineers

AI and automation promise efficiency gains, but can also create extra work when users do not trust the results. Bricsys, a developer of design software for engineers, noticed this clearly in practice. Although AI tools could automate designs in a single click, their use remained limited: users still had to manually check and adjust the output.

From automation to support

Bricsys saw that designers and engineers spent a disproportionate amount of time on repetitive tasks (such as adding parameters, making minor adjustments and routine work) that contributed little to the core value of the design. Automation seemed to be the solution, but the “black box” approach was counterproductive.

In collaboration with FlandersMake@UHasselt, Bricsys is therefore switching to a human-centred approach. Professor Raf Ramakers, a specialist in human-machine interaction, is working with the company to develop explainable AI: a system that does not make automatic decisions, but presents choices, explains implications and lets the user determine the direction of the design. This keeps the engineer in control, while the AI acts as a smart co-pilot that provides support with relevant information at the right moment.

From research to application – and beyond

The collaboration demonstrates how research can migrate to practice and how companies play an essential role in realistic test environments. Moreover, this is not a one-off project: Bricsys and Flanders Make have previously collaborated on tools that support designers in assembly planning, including practical feasibility checks during assembly.

The principles behind this transparent AI are widely applicable and form a springboard to other sectors. Explainable AI not only makes digital innovation faster, but also more reliable and usable – exactly what is needed for industrial applications to succeed at scale.