Peter van Hees is an Innovation Manager, keynote speaker, and author of AI Agents – They Act, You Orchestrate, a blunt, field-tested perspective on why the future of AI is orchestration.
While much of the AI conversation is stuck debating intelligence, Peter focuses on agency. His work centers on Agentic AI as executional leverage, not philosophical ambition. He challenges the idea that progress depends on smarter models, arguing instead that real impact comes from designing systems that can act, coordinate, and deliver outcomes under real-world constraints.
Peter works inside innovation labs where Agentic concepts are tested, not slides. His focus is early signal detection, rapid experimentation, and proving value fast, often before governance, tooling, or comfort catch up. He thrives where ambiguity is high, stakes are real, and momentum matters.
As a speaker, he is known for high-energy, no-nonsense sessions that confront uncomfortable truths. AI does not fail because it is not intelligent enough. It fails because organizations don’t know how to structure work for agents, how to assign responsibility, or how to let software act without losing control. His keynotes replace hype cycles with mental models, failure modes, and concrete design patterns for Agent-First systems.
AI Agents – They Act, You Orchestrate is his manifesto. The book argues that the next software era will not be user-first or model-first, but agent-first. The competitive edge will belong to those who master orchestration, trust boundaries, and outcome-driven autonomy, long before the tooling feels safe or mature.
Peter has worked across large enterprises such as HP, Microsoft, and KBC, alongside start-ups including IW, Cake and Setle, and institutions like the EU and NATO. Across these environments, he returns to one question: how do you move from powerful models to systems that actually do things?
If you believe AI’s future is less about prompts and more about delegation, coordination, and action, this talk is for you.