AI Research & Execution
Four ventures, one job: find the work that shouldn't be manual anymore, then build the thing that does it. Voice agents, clinical SEO, and a multi-agent suite headed for white-label.
Most AI projects fail because someone automated the wrong thing. I start at the other end.
Not the flashiest task — the one you're doing for the third time this week and resent. That's where automation pays for itself in a month.
One working system beats a roadmap. I ship something you can use, watch it run against real work, then widen it.
Monitoring, fallbacks, and a loud failure when something breaks. A system nobody trusts is a system nobody uses.
Four ventures at different stages — one platform in beta, one agency profitable, two growing.
A multi-agent suite for audiology practices — company knowledge base, business intelligence, and agents that handle the work a front office repeats every day. In beta with real clinics, headed for white-label.
Clinical SEO, AEO, and website rebuilds for audiology and hearing practices. Per-client AI agents run the research, citation tracking, and reporting that agencies usually bill hours for.
apexailabs.ai →Apex's sister brand, pointed at everyone else — the same SEO and answer-engine machinery, tuned for general business rather than clinics.
AI voice agents that make and take real phone calls, plus Audibly Academy — where people learn to build and run agents of their own instead of buying one they can't change.
audiblyai.com →I'm a business management student at BYU's Marriott School, and for the last couple of years I've been doing the other half of my education in production — researching how to build AI systems, then actually building and running them for paying clients.
Most of what I know came from shipping something, watching it break, and fixing it. I taught myself AWS because a platform needed to live somewhere. I learned agent architecture because a client needed work done overnight. That's still how I prefer to learn.
The thread through all of it: I'd rather build the system once than do the task a hundred times. If you're sitting on work that feels like it should already be automated, that's the conversation I want to have.
Thirty minutes, no pitch. Bring the problem — we'll figure out together whether it's worth automating.