Every ML system should start on one page.
That page is the Machine Learning Canvas: what to predict, what decision the prediction feeds, and what a mistake costs, laid out before anyone builds anything. Teams use it to choose which ML projects to fund, and to agree on what each one should do. 15,000+ people have downloaded it since 2015.
Where the canvas has been used:
Free, 30 minutes, on a project you're scoping. Planning ML across several teams? Let's talk about that instead.
The project is yours to land. The definition isn't yours alone to fix.
3 weeks in, the team still doesn't agree what counts as churn, who acts on the score, or what a false alarm costs… You can't ship past a disagreement like that. And you can't settle it on your own: the people who own those answers have to be in the room.
You inherited the scope, and you're the one accountable for it landing.
"AI-driven churn prediction" meant 5 different things to the 5 people who approved it.
It shipped. Nobody agrees whether it worked.
20 AI ideas on the list, enough people to build 5, and no common way to compare them.
1 page. 10 boxes. Everything that has to be agreed before anyone builds.
From what's being predicted through to what happens after the system ships.
Filled in together, in one place, before the first line of code.
What running the canvas gets you
-
What a false alarm costs, what a miss costs, and the accuracy a model would have to reach before it's worth building at all. You can decide that before anyone writes code, and compare projects on the same terms.
-
Precise enough to work from directly, instead of reconstructing it later from a deck and a chat thread. It also answers what an AI coding assistant would otherwise have to guess: what's knowable at decision time, and what counts as success.
-
The canvas holds the intent, data, decision and monitoring documentation that regulatory model reviews ask for. Written during scoping, instead of reconstructed months later.
-
A consultancy building anomaly detection for a US telecom manufacturer stopped the modeling after several rounds: the problem wasn't defined well enough to build against. Fixing the definition first is what got the system live.
-
What counts as a hit, who acts on it, what threshold to use. Those normally surface halfway through a build. Settle them first and NYK Group's R&D lead describes what happens: the whole process from use case idea to deployment gets expedited.
"MLC played an essential role in a positive project outcome. It is now being used throughout the organization."
— Daitan, now part of Coforge (global IT services firm)
Planning ML across several teams?
Let's talk about that instead.
Bring the project you're scoping
In 30 minutes, you'll know where it's likely to break, and what to check before anyone builds it.
Before you book
What do I actually get?A free 30-minute project review. I draft your canvas from a few lines, or mark up yours, before the call. You leave knowing where the project is likely to break, and what to check before anyone builds it. If you want to go further, we'll discuss working together. See how a review runs
What do I keep?The filled canvas is yours, and so are your numbers, or your client's. We sign your NDA. If you're a consultant, client-facing work can be co-branded. Your team can learn the method too.
Why not just fill it in myself?You can, and the canvas is free. The hard part is spotting a gap in a canvas that looks finished. A box that says the goal is to reduce churn looks right to anyone who reads it. But if none says how flagged customers are contacted, the predictions have no impact. That's what the review looks for.
"Louis offers a unique systems perspective, so that you can see how all the pieces of the puzzle fit together. This is ideal for consultants, or for people who have to plan or manage a machine learning project in their organization."
— Dr Zygmunt Szpak, co-founder, Insight Via Artificial Intelligence (2020)