The Vector Method
Strategy Is a Filter, Not a Wish List
Most AI plans fail before anything's built, because they're built backwards.
"What can AI do?" gets you 34 ideas, a confused team, and six months of pilots that never scale.
"Where do we lose the most money?" gets you the 2–3 that actually matter.
My job isn't to find AI uses. It's to cut hard — so your resources go where they work.
"Here's everything you could do" isn't strategy. "Here's what to do first, and why" is.
What Makes This Work
What Makes the Answer Hold Up
Judgment Over Frameworks
The most useful part of any plan isn't the list of what to build. It's the list of what not to. Chasing everything dilutes everything. I cut hard, so what's left has a real shot.
Kill More Than You Start
The most valuable part of any AI strategy isn't the list of what to build. It's the list of what not to.
Most organisations are already considering too many AI initiatives. Pursuing all of them dilutes focus, drains resources, and produces nothing that works at scale. The Vector Method cuts aggressively — so the initiatives that survive have a real chance.
Assumptions Visible
Every number rests on assumptions, and they're easy to leave unsaid.
If a projection assumes cleaner data than you have, that belongs on the table, before you commit, not after.
You Already Know Something Isn't Right
The AI projects that keep stalling. The options that start to blur together. The debates that go in circles. Let's cut through it.