Insights · April 2026 · 4 minute read

Discovery or deployment? Don’t let anyone sell you the wrong one.

The most expensive mistake in AI adoption isn’t buying the wrong tool. It’s buying a deployment when you needed a discovery, or buying a discovery when you needed a deployment. They are different products with different deliverables, different prices and different people doing them — and they get sold to the same client by the same vendor inside the same proposal.

The two questions

A discovery answers two questions and only those two: should we? and where? It is a paid review, written down, that ends with a recommendation. The recommendation can be “not yet”, and a good discovery says so when that’s the right answer.

A deployment answers three questions: how, who, by when. It is a build — a website, an AI helper, a workflow. It produces a thing your clients see, your team uses, or your accountant can put on the balance sheet.

Confusing them is how a firm ends up with a six-month build that nobody asked for, or a beautiful slide deck and no working software.

Three signs you are being sold the wrong one

What we do

We sell them as separate products on purpose. A discovery is £495, five working days, a written brief, and a recommendation that can be “don’t do this”. A deployment is fixed-price per scope, built to a written specification, and handed over in your repository on day one.

Most clients who work with us start with a discovery. About one in five comes back with a deployment. The rest take the brief and act on it themselves, or decide the timing isn’t right yet. All three are honest outcomes.

The trick isn’t to be the firm that sells you both. It’s to be the firm that tells you which one you actually need, in writing, before you’ve signed anything.

← Back to insights