I embed inside mid-market companies, identify where AI can solve today's problems and enable tomorrow's growth, and then ship the system that's actually used.
Normalized losses. It wastes money today and limits how fast you can grow tomorrow. It's where I build AI to help.
The AI systems I build do the work itself, instead of telling you to do it.
Project managers were reading multi-volume spec books and assembling submittal packages by hand — 2.5 hours on every project. I built the agent that does it in 22 minutes, validated end-to-end on their own projects.
Estimators retyped every vendor quote line-by-line back into the estimate — 20 to 30 quotes a project, more than an hour each. My system reads any vendor's PDF, maps it to their internal codes, and flags what's missing, mispriced, or misquantified. Faster bids, fewer errors, and more bids out the door.
A partner in the outcome.
We agree on what the system has to deliver before I start, and my fee is tied to it.
No slide decks, no invoice for a workshop, and no consultant pretending they care.
With the people doing the work, until I understand the workflow as well as your employees do. The real problem is never in the RFP.
The solution that will work in your business — not the one that demos well. If AI isn't worth it in a particular area, I'll tell you (and then find where it is).
Custom to your codes, your formats, your data. I write the code. You own the system.
Until your people use it and the results are measurable. Ongoing if you want — so the system keeps working as your business changes.
dkBuilds writes public letters to its board every week. Letter The first letter to the board · August 2026
Read the letters →I believe AI will lead to more fulfilled individuals and uniquely valuable businesses. Before this, I've: