Sean Pan

My career began in hardware — industrial automation, digital signal processing, defense systems, energy infrastructure. Environments where errors don't trigger patch cycles; they trigger shutdowns, safety incidents, and asset loss. That discipline taught me what constraints actually mean: not guidelines or best practices, but structural boundaries that determine whether a system remains correct when it leaves the lab and encounters reality. My BSEE from San José State, and the philosophy of AI course I took at UC Davis, gave me the language for both sides of that problem.

The second act was software. I entered the commercial internet at its inflection point — enterprise integration infrastructure when SOA and web services were being invented, product marketing during the Agile and UML era, and eventually CMO of a CRM, eMarketing, and Business Intelligence SaaS company before the category had mass adoption. These were zero-to-one environments. We were naming things that didn't have names yet. That experience — of building categories from scratch, of going to market before the market exists — is the direct ancestor of the work I do now. In 2001, while at Actional, I published in Enterprise Systems Journal on control broker architecture — real-time enterprise integration at a moment when the commercial internet's plumbing was still being invented.

The third act is where those two worlds converged. I spent over two decades operating inside insurance and financial services distribution at significant scale — managing agencies, teams, P&L, and the acquisition infrastructure that my own research now identifies as structurally vulnerable to AI-mediated discovery. The Great Decoupling isn't a prediction from the outside. It's a field report from inside the system being disrupted.