Article
The short answer
It depends on what you build, and anyone who quotes a price before understanding your operation is guessing. A well-run engagement starts with a small, fixed-scope discovery that produces a costed roadmap, so every step after it is a decision you make with the numbers in front of you.
Why nobody publishes honest numbers
AI consultancy pricing is rarely transparent, and there is a reason: the same question costs wildly different amounts depending on what sits behind it. An AI agent that answers customer emails from your live stock and pricing data is a different project to a document pipeline that reads supplier paperwork, even though both are one sentence long when a director describes them. Any firm that quotes from the sentence alone is either padding the price to cover the unknowns or planning to discover the real scope on your invoice.
What actually drives the cost
Four things move the number more than anything else:
- Where your data lives. Clean data in one system costs less to work with than the same data spread across spreadsheets, inboxes and a legacy ERP.
- What the AI is allowed to do. A system that drafts for a person to approve is simpler than one that acts on its own, because acting safely means validation, logging and controls.
- What it must connect to. Every system the AI reads from or writes to, from Business Central to a courier portal, adds integration work.
- Who runs it afterwards. Software that is built and handed over ages quickly. Software that is run, monitored and improved keeps paying back, and that ongoing attention is part of the true cost.
How we keep the decision in your hands
We structure engagements so that you never have to take a large number on faith. The first step is a discovery exercise with a fixed scope: we sit inside your operation, find where the time actually goes, and produce a costed roadmap of the systems worth building, each with the saving it targets. The roadmap is yours whether or not you build with us.
First projects are deliberately small. One process, delivered in weeks, with a measurable result: orders processed in seconds instead of minutes, replies drafted before a person opens the inbox. If the first system does not pay for itself, you have lost weeks rather than a transformation budget, and you have learned something true about where AI pays in your business.
Questions to ask any consultancy
- What will the first project measurably change, and how will we both know?
- What happens after delivery: who watches it, fixes it, improves it?
- Do we own the system and the integration layer, or are we renting it?
- What did this cost your last client of our size, and what did they get for it?
A serious firm answers all four without flinching. If you would like the same honesty about your own operation, start the conversation and tell us where your team loses the most time.