Atlanta Metro & Southeast
I run ethnographic discovery with small and mid-sized businesses across Atlanta Metro and the Southeast, then ship the production AI workflows that come out of it. On-site in Cumming, Alpharetta, Atlanta, and anywhere within a couple hours of Forsyth County.
A practice built around watching how the work actually happens, then building tools that fit the way your team already moves. The first day is a half-day on your floor with the people doing the work. The first week is a working prototype on your real production data.
The clients I do my best work with run between five and a hundred people. They have a workflow that drives them crazy. They want something running this month.
Step 1
I sit with the people doing the work and watch where the day actually breaks. The interesting friction is rarely where leadership thinks it is.
Step 2
We name the friction. Whichever workflow breaks weekly and costs your team the most hours is usually the right place to start.
Step 3
Working tool inside the first week, against your real production files. You see it run on the inputs your team is already wrestling with.
Step 4
Once the prototype proves value, we polish and integrate it. Scope grows from what worked, and you decide where it goes next.
A short read of the engagements that shaped the practice. Some have full case studies. Some are still in flight or under NDA.
Industrial valve distributor. Built a contact and territory intelligence engine: 600+ contacts and 500+ facilities, scored and segmented. Sales reps walk into meetings already knowing the account and the right angle. The same field team now covers more ground per week.
Sole publishing operations for a profitable indie game. Built creator-discovery AI, automated pricing analysis, and live campaign reporting. 180,000 units sold in 2025 against a 200,000 target.
Read the case study →On-site discovery, then a SolidWorks pipeline that produces dimensioned drawing packs from native assemblies in under a minute, validated to 0.001 inches against hand-engineered references. Engineers got their afternoons back. Drawing packs that used to consume an entire workday now ship in seconds.
Read the case study →Built a Coming Soon page for a specialty pipe distributor. From first conversation to deployed for feedback in 24 hours.
See the page →Based in Cumming, GA. Most engagements include on-site time at your facility. Atlanta Metro is a same-day drive and the Southeast is a short flight. Remote-only works for follow-on phases once the relationship is real.
Most engagements start with a half-day on site. I sit with the people doing the work and watch where their day actually breaks down. We pick one workflow that hurts. I prototype something that does it for them, and you see it run on real data within the first week. Scope grows from there based on what worked.
Day one or day two for a first running prototype. Most engagements have something meaningful on real production data inside the first week. Polish and integration come after the value is proven.
Code. I ship the working tool. Slides are for the part where I explain to leadership what we built and what it does. Most weeks the ratio is 90% building, 10% communicating up.
Small and mid-sized businesses. Owner-operator shops up to teams of about 100. Below five people there usually isn't enough repeatable process to automate. Past 100 you're in enterprise-architecture territory with different timelines.
$100/hr is the standard rate for hourly work. First engagements usually start with a $500 trial run: 10 hours of whatever your business needs most. After the trial, fixed-price options tie to specific outcomes. A workflow built and shipped, a discovery phase with a working prototype, or a full automation rollout. Cost lands once we know what you're actually trying to fix.
Yes. Most engagements include on-site time at your facility, especially during discovery. I'm based in Cumming and travel anywhere in Atlanta Metro, North Georgia, and the Southeast. Remote-only works for follow-on phases. Discovery has to happen in the room.
Claude (Anthropic) is the primary model. Tooling depends on the workflow. Python, TypeScript, web scraping, API integrations, Excel and CAD where the work lives. I work with whatever stack you already run.
That's the normal starting point. Most of my clients had never used AI in production before our first engagement. The discovery phase is built around watching how your team works today. Whatever I build fits how the team already works.
Tell me what's eating up your time. I read every message and write back the same week.