About Ray
I build the tools I wish I'd had.
Garrett Business Systems is named after me on purpose — every system I sell, I've used myself first, on my own work and my own business.
About Ray
Garrett Business Systems is named after me on purpose — every system I sell, I've used myself first, on my own work and my own business.
I climbed the ladder inside a Fortune 500 company — individual contributor, middle manager, eventually Senior Director of a division. At every level up, I expected the workload to get easier to delegate. Instead, the repetitive, low-return busywork just got bigger and higher-stakes. The difference was that at each level, I could afford slightly better tools — and that gap is exactly what this business exists to close for people who aren't there yet.
Repetitive reporting and manual busywork ate the parts of the day I actually wanted for real work — and there was no budget or standing to change how things were done.
More responsibility came with more cross-team coordination — most of it manual. I started building small automations on the side, eventually designing training and enablement systems used by thousands of people, just to keep my own head above water.
Directing a business unit meant real scale — teams of dozens, multi-million dollar budgets, systems processing hundreds of thousands of transactions a year — and finally the standing to build real infrastructure instead of workarounds. Along the way I helped take forecasting from "wildly unreliable" to routinely accurate, using the same AI-enabled, systems-first thinking I now build for clients.
I also own and run a small business on the side, where I use the same category of AI automations and agents I build for clients — inbox triage, reporting, and more — in my own day-to-day operations, not just as demos.
Coworkers and colleagues started asking me to build what I'd built for myself. That's this business.
Two from my own business, one from a personal project — all still running today.
I built an AI agent to read my own business's inbox — email, chat, and a support tool — and turn anything that needed action into a task-board card, sorted by Getting Things Done principles. The first version worked great, until one recurring email topic quietly generated more than 25 duplicate cards, one for every reply in the thread. The automation was doing exactly what it was told; it just wasn't told to check the board before creating something new. Adding that one dedup check — search first, only create if the topic isn't already open — fixed it permanently. That fix is baked into the free AIVA toolkit below.
Building an automated insights report for my business's task pipeline, I found that one metric — average cycle time — was reading 49 days when the real number was closer to 4. The cause was a wrong "created" field being used as the reference date. The same build separately caught a stale stage field that was quietly understating a pipeline count by 13 tasks. Neither error was dramatic on its own — but both would have driven real decisions off bad numbers if the system hadn't been built to check its own assumptions.
For my own investing, I built a tool that pulls 10-Q and 10-K filings straight from SEC EDGAR and turns them into structured financial summaries — flagging the numbers and trends worth a closer look instead of requiring a full manual read-through of every filing. Same instinct behind every build here: let the AI handle the repetitive first pass, and save your judgment for the parts that actually need it.
Currently preparing for Anthropic's Claude Certified Architect certification through the Partner Academy.
Systems built to run at real scale — hundreds of thousands of transactions a year, for large organizations.
Recognized as a top 10% user on an early enterprise AI platform, back when most people were still asking whether AI was a fad.
An MBA with a corporate finance specialization and a minor in Information Systems means I can build tools for real financial analysis, not just inbox triage.