Case study zero
The system we built for ourselves
The system we'd build for you, we built for ourselves first: what it does, why it's built this way, and what we learned building it. This is the system running, described from the inside.
It is why we know what this work takes.
Who runs it
I'm Luke Morton.
I grew up working on my dad's farm in Ohio. Through high school I shot photo and video for farms and wildlife ranches, and after it I built websites for roofing companies. Small businesses have been the whole of my working life since. Business Administration degree, Digital Marketing and Advertising, Liberty University, 2019.
I ran the marketing for a basement repair and refurbishment company: site, ads, content, video, and the CRM. That is where I first watched the thing this practice exists to fix. Good leads died because filtering them was tedious and nobody got to it. Invoices went unpaid for weeks because chasing them was somebody's least favorite job. Both were costing real money every month, and neither was anybody's actual job.
Alongside it, brand and packaging for a portable air conditioning line, and freelance web and graphic design. Then four years running my own Amazon store: inventory, supplier documents, spreadsheets, product research. That operation was mine, so the admin was mine too, which is a different kind of education.
Testing AI against that store is how this started, a few years back now. I wasn't studying it. I had data to organize and products to qualify, and doing it by hand was eating my week.
The pattern held in every one of them: different industries, same waste. What I took from that is narrower and more useful than "AI helps." It is knowing which work should be removed, which should be simplified, and which is worth automating, without the quality dropping. Those are three different answers, and most of what gets sold as automation gives the same one to all three.
You would be working with me directly. The person on the fifteen-minute call is the person who runs the assessment, writes the document, and builds whatever comes out of it. I work from Columbus, Ohio, and the work itself is remote.
What we run on ourselves
Everything Maseron knows lives in one structured system rather than scattered across drives, inboxes, and someone's memory.
- Capture is separate from filing. Anything worth keeping gets dropped in raw (a note, a transcript, a clipped article) and sorted afterward. Deciding what a thing is before you can save it is exactly the friction that kills these systems, so we removed it.
- Every page carries its provenance. Each entry records which raw material it came from. A claim can be traced back to what was actually said, by whom, on what date. Pages making factual claims with no source attached are treated as a defect, because they usually mean a merge went wrong or something got invented.
- Uncertainty gets written down. Where a source is unclear, contradicts another source, or was probably mis-transcribed, the entry says so. An entry with a flagged gap is more useful than a confident invention, and the flag is what stops the invention from hardening into fact three months later.
- It gets checked on a schedule. A standing health check looks for broken links between pages, orphaned entries, stale material, and the important one, contradictions between pages that both claim to be current. Contradictions get reported rather than auto-resolved.
The positioning on this website came out of that system. So did the pricing structure, the method page, and the decision to say no to three of the four things Maseron could have sold you.
Why it's built this way
Most company knowledge bases fail the same way, and it isn't a technology problem.
- They live outside the work. A wiki you update after doing the job is a chore that gets compressed away the first busy week, and once it’s stale nobody trusts it, and once nobody trusts it nobody updates it. The system has to sit inside the loop where work actually happens.
- They’re organized for people, not for retrieval. The tidy folder hierarchy that feels intuitive in a meeting frequently does nothing for whether an answer can be found. Better to put material in, watch what can’t be located, and reorganize from that.
- They save the output and throw away the reasoning. Keeping the final document but not the constraints, the rejected options, and the corrections that produced it discards the part that teaches the next person. The method is what compounds.
- They get built as a destination. A portal everyone is supposed to visit is one more thing to open. People should keep working where they already work; the knowledge sits underneath.
Getting it used
A system nobody uses is worse than no system, because you paid for it. So this is designed before the build starts, not addressed afterward with training.
- We remove tools rather than add one. If the new system becomes the eighth thing in the stack, it’s asking your team to do more on day one. It should replace things.
- Familiar views, different engine. A board stays a board. A pipeline stays a pipeline. Nobody should have to learn a new mental model to do the job they already do.
- Nobody has to be good at AI. The work runs in the background and people approve outcomes. There’s no prompt box for anyone to quietly avoid.
- We watch for the shadow spreadsheet. When someone keeps their old tracker running as insurance, that’s information: either they weren’t properly trained, or the system genuinely doesn’t handle their case. Both are cheap to fix in the first month and expensive after three.
- We measure the work, not the logins. Seat counts and usage stats say nothing about your operation. What matters is manual hours remaining against the original map. If a workflow is still being done by hand, the build isn’t finished.
Whose it is
It lives in your accounts.
Your infrastructure, your storage, your credentials. We take scoped access for the engagement and give it up at the end. We don't host your knowledge base and we don't hold your documents.
Nothing is shared between clients. Ever, in any direction.
You can revoke our access at any time, without asking and without explanation.
If we manage it for you afterward, that's a service you're choosing. It isn't a dependency we've built in.
What arrives every week
If we're running something for you, a standing report lands on a fixed day. Not on request, not when there's news:
- what got done
- what that work used to take
- what it took this time
- anything that failed or needed a human
- what we’re looking at next
Two numbers get agreed before we start: how long each task currently takes you, and what an hour of that time is worth. That's what makes the report a dollar figure rather than a generic estimate — it's your arithmetic, not ours.
The point is that you shouldn't have to remember why you're paying us. It should be on the page.