Instead of starting with workflows and adding intelligence later, foundational models are flipping this approach. They start with a general-purpose AI model—one that can understand and respond in context—and add configurable workflows on top of it. Instead of committing to a traditional SaaS tool, companies might benefit from asking, “What if I started with intelligence (an AI model) and built software around it?” This discussion will see panellists consider the future possibilities of an AI-centric tech stack that might offer more agility and intelligence at its core, rather than layering it on top.