Trust is not added to AI products at the end. It is shaped through thousands of decisions made during design and development. Engineering defines what the system can do, legal determines what it should do, and product teams decide how users experience and control it. Compliance emerges, or fails, in the interaction between these choices.
In the next Trustworthy AI Circle, we will look into what it means to operationalise compliance in practice: how compliance requirements become embedded into engineering, product, and operational processes, and how organisations can move beyond seeing compliance as a purely reactive, bureaucratic exercise.
Join Vibha Mohan (Group Legal Counsel), Baptiste Marien (Director of Product, Pdfools), and Eduard Schäli (VP Engineering, Smallpdf) for a conversation anchored by a live demonstration of AI Redaction, a compliance-first product recently launched by Pdftools. The expert panel will then explore the themes this raises, and in the co-creation session that follows, we will open up the real tensions, trade-offs, and practical approaches for building AI products that are compliant by design, without turning compliance into a purely reactive checkbox-driven exercise.