Before you buy more software, find out what's actually broken.
I'm an independent consultant. I map how your work really gets done — the spreadsheets, the handoffs, the workarounds people invented because the official process didn't fit — and then recommend the smallest change that fixes it. Sometimes that involves AI. Often it doesn't.
Book a 30-minute callTwo ways people usually find me
Most engagements start in one of these two places. They need very different things, so it's worth being honest about which one you're in.
Lane one
AI is already in the picture
Either you've deployed something and it isn't behaving, or you're hearing enough about AI that you want to find out whether it's real for you — without betting the budget on it.
- It broke, or it drifted. An automation silently stopped firing. Output quality slid. A rollout stalled because nobody trusted it. I diagnose what went wrong and tell you whether it's worth repairing or retiring.
- You want a trial, not a transformation. One narrow, reversible pilot with a fixed scope, a fixed end date, and an honest read at the end — including "this didn't work."
Lane two
AI isn't the answer — you need someone to look at the actual problem
This is the majority of my work, and usually the more valuable half. The bottleneck is a process, a handoff, or three systems that don't talk to each other. No model fixes that.
- One person is the bottleneck. Everything routes through someone who is manually re-keying data between systems.
- Nobody can see the numbers in time. Budget, inventory or status data exists, but not where the decision gets made.
- The tools you already pay for aren't being used. Often the fix is configuration and process, not procurement.
How I work
Four habits that shape every engagement. They're unglamorous on purpose.
Look before recommending
I sit with the people doing the work and watch the real process, not the documented one. The gap between those two is usually where the problem lives.
Map the current state honestly
Including the workarounds, the shadow spreadsheets, and the steps that exist only because someone left three years ago. You get the real picture, even when it's unflattering.
Validate before you commit
Vendor capability claims get tested against your actual systems before they turn into a purchase order. A surprising number don't survive that.
Start small, scale only if it earns it
Phase one should be low-risk, reversible, and useful on its own. If it doesn't prove out, you stop — having spent a little, not a lot.
When I'll tell you not to hire me. If the honest answer is "fix your intake form" or "you already own a tool that does this," that's what you'll hear — in the first conversation, before there's an invoice attached to it. I'd rather lose a project than sell you a system you don't need. Buying technology you can't absorb is more expensive than doing nothing.
A recent engagement
An independent school came to me about procurement. There was no AI in the solution — the problem was a process that had quietly outgrown the person holding it together.
Not sure which lane you're in?
That's a perfectly normal place to start. A 30-minute call is usually enough to tell.
No sales pressure, no technical jargon, and no obligation to proceed.