The problem

A public university holds $90M+ in equipment across 127 departments, and the system of record was static. Finding an asset, auditing a department, or clearing surplus meant manual lookups, email chains, and multi-tab spreadsheet work. Audits crawled. Nobody could answer "is this equipment actually being used?"

The system

Strata is built on Smartsheet in four layers, each answering a different question. Compliance: is every asset accounted for and audit-ready? Surplus workflow: what should leave, and how does it exit cleanly? Utilization scoring: is this equipment earning its footprint? Commercial intelligence: what is it worth if it moves?

Data flows in from Oracle ERP and Banner Finance without manual handoffs. AI tooling executes next steps in receiving, compliance follow-up, and surplus disposition rather than just summarizing status.

The result

Audit workflow cycle time dropped 90%. A live dashboard tracks open items, discrepancies, and hours saved, so the system proves its own value quarter over quarter. It's in daily use across university operations.

The problem

Institutional policy drifts. Documents reference other documents that changed, deadlines pass silently, and a full review takes a committee most of a semester. Nobody reads the whole policy library. So nobody sees the whole picture.

The system

Policy Studio is regulatory cartography for higher education: an AI system with five commands that read, cross-reference, and audit the policy library at machine speed. It maps how policies connect, flags contradictions and stale references, and turns a shelf of PDFs into something you can actually query.

The result

A single sweep surfaced 26 findings across 16 policies. The kind of review that used to be a committee's semester is now an afternoon, with the human judgment saved for the findings instead of the reading.

The problem

Context is the tax on everything. Projects, data, and decisions live in a dozen tools, and every task starts with re-explaining the situation. AI is only as useful as the context it can see.

The system

AOS is an AI operating system built in Claude Code with three layers. Context: who I am, what I'm working on, how I decide. Data: the live information those projects run on, from operations to health tracking. Skills: repeatable procedures the system can execute, not just describe.

One environment where everything compounds instead of resetting.

The result

In active development and moving toward daily-driver status. It's also the proving ground: every pattern that works here becomes something I can build for someone else.

The problem

Local news is disappearing, and what's left is scattered across Facebook posts, event calendars, and press releases nobody reads. The Upper Cumberland needed one place that pulls the week together. But a one-person newsroom only works if production is a system, not a grind.

The system

The Cardinal is a weekly newsletter on beehiiv with an AI-assisted production pipeline behind it. Zapier automations connect sourcing, content tools, email, and publishing. AI handles research, drafting, and assembly; I handle judgment, voice, and the final read. Every issue follows the same section architecture, so the system knows exactly what to build.

The result

A reliable Monday-morning publish, every week, from a team of one. What used to be a multi-hour manual process keeps compressing toward a sub-2-hour production target. Media, treated as an operations problem, behaves like one.

The problem

My grandmother, Miriam McClung, is a Birmingham painter with a 75-year career and almost no digital footprint. The work was extraordinary. The audience was whoever had walked past it.

The system

I built her a content operation: a repeatable pipeline for turning paintings and the stories behind them into short-form video and carousels for Instagram and TikTok. Same discipline as everything else I run. Consistent formats, systematized production, and a clear voice, so a one-person team can publish like a studio.

The result

The account went from nothing to 40,000+ followers, with content driving millions of views. A 75-year body of work now has an audience measured in the millions, and the growth engine runs on a system, not luck.

The problem

Could I take a business from zero to shipped at volume, solo? Wisecrackle was the test: a print-on-demand brand where every function most companies staff, I had to systematize instead.

The system

The full supply chain runs lean: vendor coordination, drop-ship fulfillment, shipping, cost control, and customer resolution, all handled as designed processes rather than daily firefights. Product design and listings run on repeatable templates so new products ship fast.

The result

4,000+ orders since founding in January 2024, run by one person. It's the smallest asset on this registry and the clearest proof of the thesis: systems beat headcount.