Most companies are adding AI to the product they already built. The more interesting work is imagining the product that assumes it — and then actually building the thing.
Built for businesses that don’t call themselves tech companies — where the gains are largest and the least glamorous.
How we work together
Different shapes for different problems. The common thread is that each one ends in something running, not something presented.
Imagine the product that assumes AI, then build it.
Not a feature bolted onto an existing roadmap. We start from what your business actually does, ask what becomes possible when intelligence is assumed rather than added, and ship a working version you can put in front of real people.
Ongoing technical leadership for teams without a CTO.
Architecture, hiring, vendor decisions, and the judgment calls that are expensive to get wrong. Enough presence to be genuinely accountable, structured so you are not carrying a full-time executive salary to get it.
Fixed scope. Find where modern tech removes real cost. Ship it. Prove it.
A defined engagement that ends in something running, with a number attached. Built for operations that still move on spreadsheets, email threads, and institutional memory — where the gains are largest and least glamorous.
Approach
The question is never where we can put AI. It is what the business actually does all day, and which parts of that should no longer require a person.
Anyone can add a panel to an existing screen. The harder move is redrawing the process itself, so the intelligence is load-bearing rather than decorative.
Working software beats a deck. The fastest way to find out whether an idea holds is to put a real version of it in someone's hands and watch what they do.
Novel at the surface, conventional underneath. The interesting risk belongs in the product. It never belongs in the infrastructure.
Currently
No case-study theater. This is what is actually being built, poked at, and read about right now — which says more about how the work goes than a finished logo wall would.
A short, plain description of what this is and who it is for. Two or three sentences is the right length — enough to make the idea concrete, short enough that someone scanning the page still reads it. Say what it does before you say why it matters.
Something you are actively poking at but have not committed to. The Exploring tag does real work here: it lets you show early thinking without implying a launch date, which is exactly the honest position for work at this stage.
An open question you are spending time on. Research entries signal the direction of your attention, which is often more persuasive to a prospective client than a finished case study — it shows how you think, not just what you shipped.
Contact
Two doors, depending on how formed the idea is. Neither one obligates you to anything.
