Intentional Architecture in the Age of AI: Defining and Automating the Guardrails

As AI-assisted development becomes part of everyday engineering work, the speed and scale of change are beyond what many teams are prepared to manage. Generative systems can produce code, tests, designs, and entire subsystems in seconds, but without intentional guardrails they can also automate ambiguity, amplify architectural drift, and create more complexity than they remove.

In this keynote, we explore why the future of software engineering depends not on generating more artifacts, but on defining the boundaries that make those artifacts trustworthy. We will briefly revisit a few moments in scientific and engineering history where breakthroughs emerged only when bold ideas were paired with disciplined constraints. These stories frame the central challenge of the AI era: AI can generate output at unprecedented speed, but only engineers can supply the intent, structure, and feedback loops that keep that output aligned with reality.

We will introduce a path forward: how input, process, output, and feedback guardrails, combined with a clear and shared understanding of architectural goals, can help teams navigate AI-accelerated development without losing clarity, ownership, or control. You will leave with a direction to explore and a way to think about AI-assisted engineering that keeps human intent at the center while enabling teams to move quickly with confidence.