AI is a force multiplier. What it multiplies is the quality of thinking behind it.
I sit at the intersection of product design, systems building, and AI strategy. Over the past few years I've moved from designing complex products and internal tools into defining the operating models, AI-integration strategies, and executive-facing systems behind them — embedded today inside Assurant's executive strategy function. Most days that means turning an ambiguous leadership need into something real: a working tool, a strategic framework, a patent-worthy program, an executive presentation.
Before this, four years designing in financial services. Before that, eight years as a US Army Cavalry Scout and PSYOP — where I learned to make decisions before all the information arrives. That thread runs through everything I build.
Most organizational knowledge lives in the wrong place, belongs to the wrong people, and degrades the moment it stops being actively maintained. This system maintains itself — structured around how people actually search.
Access requiredWhen a transaction needs a licensed specialist, the usual answer is scheduling: customers wait, specialists sit idle, or both. The lift was a coordination model that routes the right specialist to the right transaction in real time without breaking the compliance trail. Patent pending on the underlying system.
Access requiredLeadership reports are a synthesis problem dressed up as a writing problem. The lift was the schema — a structured representation the AI could plan against, not paragraphs it had to author from scratch.
Access requiredAs product designers, we've always relied on familiar patterns to create clear, intuitive experiences.
We are experiencing a significant shift in product design due to AI advancements.
When Karpathy introduced the idea of "vibe coding," it clicked with something many of us in tech had started noticing.
Tactical Decision Aid (TDA) — for when the room still wants an answer.
TDA · 1965 Army doctrine. The decisions still aren't clear.
The most common AI mistakes I see aren't about the wrong model or the wrong tool. They're about bringing AI into a problem before the problem is understood — and treating speed as the goal when speed in the wrong direction just gets you somewhere wrong faster.
The hardest part is rarely the model. In regulated industries especially, the gap between a working prototype and something that survives real conditions — compliance, audit trails, edge cases that never showed up in the demo — is enormous. Closing that gap takes someone who reads both the technology and the terrain.
If you're working on something that doesn't have an answer yet, I'd like to hear about it. abraytondesign@gmail.com.
— AB · HOU · 2026