Who fits which buyer — vendor comparisons, not rankings

How these lists are put together

Every vendor list on the internet has an agenda, and the honest thing is to state ours: we are an AI engineering practice, we sell work in every category below. So we are not in any of the comparisons: each page discloses where we fit and where we do not in a separate, labelled box, with only evidence you can check yourself. Where another firm is a better fit for a particular shape of work — a large regulated programme, a specific vertical, a geography we do not cover — that is said plainly, because a list that never recommends anyone else is an advertisement wearing a list's clothes.

The entries describe what each option is actually good at and who it suits, rather than scoring everyone against a single scale. "Best" is not a property a firm has; it is a relationship between a buyer's constraints and a supplier's shape. A boutique that does one thing exceptionally is the right answer for a team that needs that thing and the wrong answer for a team that needs breadth. The comparisons are grouped by buyer type to help you recognise which situation you are in, which is more useful than a rank order.

We do not publish other companies' prices or license terms unless we can point at where they say so themselves. Prices go stale within months, a rate quoted second-hand is unverifiable by the reader, and a claim about someone else's licensing is a factual assertion made on their behalf. Where the shape of an offering matters to the decision we describe the shape — fixed scope versus retainer, audit-first versus build-first — without attaching a figure we cannot source.

Lists are revised when the market moves rather than on a schedule, and each carries the date it was last reviewed. A category where nothing has changed does not get a cosmetic refresh to look current. If you think an entry is wrong or an obvious option is missing, we would rather hear it than not — the correction costs us nothing and the omission costs a reader something.

What a shortlist cannot tell you

The limit of any list like this is that fit is discovered in conversation, not in a page. Two firms with near-identical descriptions can feel completely different once you are on a call with the people who would actually do the work, and that difference — whether they ask good questions about your constraints, whether they push back on anything, whether the person selling is the person delivering — predicts the outcome better than anything we could write here. Use the lists to build a shortlist of three, then spend an hour with each.

The question worth asking every one of them is what they would not take on. A supplier who claims to be a fit for everything has either not thought about it or is not telling you, and an honest answer about where they are weak is the strongest available signal that the rest of what they say is reliable.