What an AI Consulting Engagement Actually Costs (and What You Get)

Why this comes before the build

Two things make AI projects expensive, and neither is the modelling. The first is starting with a technology and hunting for a use for it. The second is discovering months in that the data needed to train on was never captured consistently. A consulting engagement exists to surface both before a build budget is committed.

What it actually covers

• A process review — which decisions in your operation are repetitive, high-volume, and rule-poor enough for a model to help with
• A data readiness check — what you hold, where, and what shape it is in
• A build-versus-buy comparison, including whether an off-the-shelf tool already does this cheaper
• Running-cost modelling at your actual volumes, not a demo's
• A staged roadmap, sequenced so the first stage pays off on its own

How it is priced

Fixed price against a written scope, not a day rate that runs. What moves the number is how many processes are in scope and how many source systems hold the data — a single-process review is at the small end, an operation-wide assessment across an ERP, a POS and a stack of spreadsheets is at the larger end. You get the number before the work starts.

What you get at the end

A written assessment, not a slide deck — meant to be read by whoever holds the budget and by whoever would build the thing. It includes a prioritised list of opportunities, an honest data-gap list, a recommended approach per opportunity, and a phase plan sized so the first deliverable is small enough to prove or disprove the case on its own.

When the answer is not to build anything

Sometimes the finding is that the data does not support the idea, or the volume is too low for the running cost to make sense, or a simpler rules engine solves it for a fraction of the price. That is not a sales objection to be handled around — it is the finding that most often turns out to be the one worth paying for.