Real Problems, Real Solutions
I've spent 15+ years solving messy problems in financial services—from mortgage compliance that actually works to AI systems that don't just alert, they prevent. Here's how I think through the hard stuff.
These aren't theoretical case studies. They're based on real problems I've solved at companies like USAA, LoanCare, and ServiceMac—where I led teams building AI-driven compliance platforms, predictive models, and automation systems.
Each case study shows my approach to product thinking: finding the real problem, designing for adaptability, and delivering measurable business impact. They're deliberately concise—I'm optimizing for showing how I think, not proving I can write a novel.
Businesses spend hours tweaking prompts and managing multiple AI tools, ending up with generic content that needs heavy editing.
Users don't want 'AI'—they want magic that works. The difference between good and great AI products isn't the model—it's the product thinking around it.
When you move fast, you lose track of what you've built. CollegeCompass needed clarity on 32 features to plan their next phase.
Missing features aren't always critical. Prioritization matters more than completion. Sometimes you build better than you planned.
$4.2B in penalties in 2023—not from bad intent, but from systems that couldn't keep up.
Most compliance platforms try to catch everything. Wrong game. Better to make violations structurally impossible.
Customer acquisition costs 5-25x more than retention, yet most companies focus on the wrong metrics.
Churn prediction models are backwards. Don't predict who will leave—predict who's quietly unhappy but hasn't decided yet.
Servicers spend millions handling requests with 15-20% conversion rates. The rest is wasted effort.
Call centers optimize for speed. They should optimize for intent. One high-intent call is worth ten tire-kickers.
Generic experiences hide massive costs in low conversion rates—typically 1-3% when personalized approaches hit 8-12%.
Most companies fear hyper-personalization because it seems complex. But generic experiences are actually more expensive.
Most "problems" are symptoms. I dig deeper to find what's actually broken—often it's not what everyone thinks it is.
In fintech, regulations change, customer behavior shifts, and markets evolve. Build systems that learn and adapt, not rigid solutions.
Financial services requires caution, but too much caution kills innovation. I find the sweet spot where you can move fast without breaking things.
Every product decision should tie to business metrics. I focus on ROI, not just features—because pretty solutions that don't work aren't solutions.
Whether you're building compliance platforms, retention systems, or operational tools—let's talk about how to approach the problem differently.