Do You Need an AI Strategy, or Just an Automation?

Almost every first call opens with some version of the same question. Sometimes it’s “we need an AI strategy.” Sometimes it’s “can you automate our invoicing.” About half the time, the person asking has named the wrong one.

They’re not similar purchases. They answer different questions, for different buyers, at different prices, and the failure mode is quiet, because you get a perfectly good deliverable that solves a problem you didn’t have.

The one-sentence test

Can you already name the process that’s costing you?

If you can, and the answer is “our team spends two days a month rebuilding the same report,” you don’t need a strategy. You need someone to price the fix and build it. Strategy work would spend five weeks arriving at a conclusion you could have written on day one.

If you can’t, and the honest answer is “leadership is asking what our AI plan is and we don’t have one,” then automating something is premature. You’d be optimizing a process that a strategy might have told you to eliminate.

What automation work actually answers

Bottom-up, at the process level: what should we automate, in what order, and what is each one worth?

It’s specific and it’s fast. Volume counts, minutes per instance, loaded hourly rate, then a ranked list where every item carries a price. Real candidates are almost always the same three shapes: document handling, lead routing, and reporting nobody trusts. Weeks, not quarters. The buyer is usually the owner or the operator, because they’re the one watching the hours disappear.

What strategy work actually answers

Top-down, at the business level: where does AI change how we operate?

That means capability, org design, who owns this, what gets funded, and in what order. The output is a roadmap and an investment case with confidence ranges rather than a single ROI number nobody can defend. The buyer here is the executive team or the board, because the decisions are theirs to make.

Both mistakes are real

Buying strategy when you needed automation gets you a thoughtful document and a business that runs exactly as it did before. A strategy nobody funds is a document nobody reads.

Buying automation when you needed strategy is worse, because it feels like progress. You ship three automations, they work, and eighteen months later you’ve made a fundamentally wrong operating model slightly more efficient.

The question underneath both

Before either one, there’s a harder test: can you name the decision this work informs, and the date it has to be made by?

If you can’t, buy neither yet. Work that isn’t attached to a decision becomes shelf decoration no matter how good it is, and that’s true whether it’s $2K or $200K.

If you can name it, the answer to strategy-or-automation usually names itself. Then it’s just a question of scope, and what each one costs is published.