How Much Does an AI Consultant Cost?

AI consultants price their work three ways: by the hour, by the project, or on a monthly retainer. Hourly runs from around €75 to €300-plus, projects from a few thousand to tens of thousands, and retainers from roughly €1,000 to €5,000-plus a month. What you should actually care about is not the model, it is whether you end up owning the capability or renting it forever.

Here is how each model works and how to spot padding. Everything is in euro and reflects the Irish and European market.

The three pricing models

Hourly. You pay for time. Simple and flexible, and fine for advice or a quick fix. The risk is that hourly billing rewards slowness, and you rarely know the total upfront. Typical rates run from about €75 for a generalist to €300 or more for a specialist.

Project. You pay a fixed fee for a defined outcome, "build these three agents." The big advantage is you know the cost before you start, and the risk of overrun sits with the consultant, not you. Ranges vary widely with scope, from a few thousand euro for a small build to tens of thousands for something serious.

Retainer. You pay monthly for ongoing work and support. Good for continuous improvement and a business that keeps evolving. The catch is the one below.

The dependency trap

Here is the thing to watch. Some consultants build in a way that keeps you needing them forever: undocumented systems, everything running through their accounts, no handover. The retainer never ends because it cannot. That is not partnership, it is a leash. The better model builds capability into your business so that even if you stay on a retainer, you could leave and still run everything. Ask directly: "If we stopped working together, could I keep running this myself?" The answer tells you everything. If you want to understand the delivery models behind these fees, see DIY vs done-with-you vs done-for-you.

Questions that expose padding

Before you sign, ask these:

• "What exactly will I have at the end, and who owns it?" Vague answers mean vague value.
• "Is this fixed scope or open-ended?" Open-ended is where budgets quietly balloon.
• "Do you run AI agents in your own business?" If they do not eat their own cooking, be cautious.
• "What happens if it takes longer than expected?" On a project, that is their problem; on hourly, it is yours.
• "Can I run this without you afterward?" The dependency test.

Padding hides in fuzzy scope and hourly open-endedness. Clear deliverables and fixed prices flush it out. There is a fuller version of these in how to choose an AI implementation partner.

A tangible example: imagine Liam comparing two quotes

Imagine Liam, weighing two proposals. One is an open-ended retainer at €3,000 a month with a beautiful strategy deck and no clear deliverable or handover. The other is a fixed, done-with-you build from around €1,500 a month that documents everything and hands him ownership within a few months. On paper the first looks like a bigger, more impressive engagement. In practice, it is easy to see how Liam could be paying that €3,000 indefinitely with nothing he owns, while the second leaves him self-sufficient and cheaper. The lesson: judge cost against what you walk away owning, not the size of the invoice.

What good value looks like

The best-value model, in my clearly biased but honestly held view, is one that transfers the capability to you. A with-you build from around €1,500 a month that leaves you owning the system beats a cheaper retainer that keeps you dependent, because you are buying independence, not just labour. Judge cost against what you walk away owning, not just the monthly figure. To sanity-check any fee against the return, run it through the ROI of AI agents, or start on the home page.

What you should actually get for your money

Whatever the pricing model, the thing to judge is the deliverable, not the day rate, so be clear about what lands in your hands at the end. A good engagement leaves you with working agents wired into your real tools, plain-English documentation of how they are set up, the accounts and access in your own name, and enough understanding that you could keep the whole thing running if the consultant vanished tomorrow. A weak engagement leaves you with a glossy strategy deck, a few logins that live in someone else's account, and a dependency you did not ask for. When you compare quotes, ask each consultant to spell out exactly what you will own on the final day, and watch how comfortably they answer. The good ones enjoy the question because ownership is the point of their work; the padders get vague, because vagueness is where the ongoing fees hide. Price matters, but what you own at the end matters more, since that is the difference between buying a capability and renting a habit.

Hourly, project or retainer: which to choose

If you need a quick opinion or a one-off fix, hourly is fine and flexible. If you want a defined outcome with cost certainty and the overrun risk on the consultant rather than you, choose a project fee. If your business keeps evolving and you value ongoing improvement, a retainer makes sense, provided it is built on a system you own and could walk away from. The models are not good or bad in themselves; the only real trap is a retainer that exists because you cannot leave rather than because you choose to stay. Match the model to the job, insist on ownership, and any of the three can be good value.

Red flags that a quote is overpriced

A high number is not automatically a bad one, and a low number is not automatically a bargain, so look past the figure to how it is justified. Be wary of a quote that will not commit to a fixed scope, that cannot tell you exactly what you will own at the end, or that bundles a vague "AI strategy" phase with no working deliverable attached. Be wary too of anyone who cannot show you agents running in their own business, since a partner who does not use what they sell is guessing. On the other side, a clear price attached to a defined, working outcome that you will own outright is usually fair even when it is not the cheapest, because you are buying a capability rather than renting a dependency. Judge the justification, not just the number.