From curiosity to capability: How Henning Larsen turns AI exploration into practice ?

AI adoption in architecture is no longer a question of access, but of approach.

While many practices experiment with tools, few have established a clear model for embedding AI into everyday work.

This session presents Henning Larsen's internal model for AI adoption, developed and tested across multiple international offices. It is structured around four core principles: literacy, transparency, agency, and accountability, forming a practical framework for moving from isolated experimentation to meaningful integration.

A key aspect of this model is the balance between bottom-up exploration and organisational structure. Employees are encouraged to test, adapt, and challenge AI tools within their own workflows, while shared principles and training ensure alignment across teams.

The session also introduces the distinction between horizontal and vertical adoption. Horizontal adoption focuses on broad accessibility and capability building across the organisation, while vertical adoption targets specific workflows where AI can meaningfully enhance design, analysis, and communication.Drawing on real use cases from practice, the talk demonstrates where AI adds value, how teams are trained to engage with it critically, and how ethical considerations are translated into practical decisions.

The takeaway is a replicable model, not a set of tools, but a way of structuring adoption, for AEC professionals ready to move beyond ad hoc experimentation towards something more intentional and lasting.