Steve Finikiotis
Three decades building operational systems where being wrong has consequences — now designing how organizations decide around AI.
Steve Finikiotis is focused on a deceptively simple question: how do we make AI actually work for people?
Not for a demo or a benchmark, but for the people who run an operation and the people its decisions affect. In practice, that means AI that works within an organization’s rules, strengthens the judgment of the people who run it, knows when a decision is not its to make, and leaves the people accountable for the outcome with the final say.
The question comes from experience. He has spent three decades designing and scaling operational systems in high-consequence environments — airlines, telecommunications, finance, healthcare, manufacturing, and energy. As managing director of an international carrier and later leading technology development across fragmented, cross-border markets in Africa, he worked in settings where a system that only usually works is a liability, not an inconvenience. His clients have included AT&T, Adobe, IATA, and Federated Department Stores.
He has been building with AI since 2021, well before the current wave. Through South Point he has led system design across science, technology, and complex-policy domains with a consistent approach: define the decision precisely, turn multi-step judgment into durable operations, and keep the system auditable at every layer — evidence, rules, permissions, and human authority intact.
At Arc Capital, a micro-fund focused on early-stage generative AI, he worked at the intersection of technology and operations — assessing portfolio performance while advising teams building systems meant to scale. He has served on multiple boards and mentored founders through Brown University’s entrepreneurial programs.
His conviction runs against the prevailing view: the model is not the system. A larger model does not make AI work for people. The architecture around it does — the information it can see, the logic that governs it, the actions it may take, and the people who remain accountable for the outcome.
His practice brings together a team of decision architects, researchers and data analysts, AI engineers, and domain specialists. They analyze how the operation actually decides and build working systems with the controls and human oversight that real operations require.
His work centers on operations where being wrong is expensive, irreversible, or both. That is where the design of the system matters most, and where getting it right matters most to the people on the other end of the decision.