Make the economic question explicit.
Start with the value pool, the strategic objective, and the consequence of getting the decision wrong.
About · Michael Upchurch
I have spent my career where strategy meets operating reality.
I am an AI strategy and value-creation executive with more than twenty-five years of experience across financial services, enterprise technology, advanced analytics, and AI.
Career throughline
The throughline is older than AI: find the value, build the system, stay for the result.
I have built enterprise ML portfolios, led operating-model change, designed strategies that produced balance sheet growth, owned the P&L of an analytics software company, launched commercial products, developed partner channels, and helped enterprises move technical capability into production and market value.
My work has taken me from Bank of America and Capital One to co-founding Fuzzy Logix and leading financial-services and insurance strategy at Domino Data Lab.
Eddington is where I continue to build, test, and refine the ideas behind that work. It gives me a practical place to stay close to current AI capability while developing operating methods that can be applied inside an enterprise.
How I work
Design the operating system around the work.
Ownership, decision rights, evidence, funding, risk, and production criteria shape whether a technical capability becomes an enterprise capability.
Keep governance close to delivery.
Controls are most useful when they influence design and create evidence as the work happens.
Measure the result and use it.
Production evidence should influence the next investment decision, not simply close the last project.
A personal principle
Every interaction should leave the other person better off than before.
My father taught me that and I have tried to run my career on that principle. It's also useful test for AI: the technology should leave the customer, the team, and the business better off in ways they can experience and measure.