We should do something with AI — but what?
The normal case and the best starting point: no finished idea required.
Answer · Costs & formats
Often less than you think — and sometimes almost nothing: For many tasks an off-the-shelf AI tool for 0 to 30 euros a month is enough. Above that there are three clearly cut brackets: a one-day workshop at a fixed day rate, ongoing support at a predictable monthly figure, and implementation projects with a fixed price per milestone. “It depends” is not an answer — four cost brackets are.
The four cost brackets
Each bracket solves a different problem. The skill is not to sell the most expensive one — but to spot the right one.
ChatGPT, translators, dictation and meeting-notes tools: for standard tasks without sensitive customer data, this is the right bracket. No project, no consulting — a subscription and one clear rule about what may go in and what may not.
When your team has long been experimenting but rules and real skill are missing: one day on site, on your actual tasks — including the record for the training obligation under Article 4 of the EU AI Act. To the AI Day →
When AI already runs inside your processes and someone has to stay on it permanently: review, guardrails, further development — a fixed monthly figure, cancellable, with accountability for the judgement. To Operator →
When off-the-shelf tools fail against your reality, because your own data, your own systems and interfaces come into play: a project with clear milestones, each at a fixed price. No open-ended invoice, no bottomless pit. To Implementation →
“If a 30-euro off-the-shelf tool is enough, I will tell you so instead of building you something.” That sentence stands exactly like this in my verification record — and it is the reason the initial assessment is free: it is allowed to end with you buying nothing from me.
Do not decide by budget, decide by three questions: how sensitive is the data? How many people work with it? And does something have to be connected to your systems?
The most expensive option, incidentally, is in none of the four brackets: the wrong project. Spending months having something built that a subscription could have done — or deciding nothing at all while the team keeps experimenting without rules.
Not as a tactic. A day rate depends on team size and goal, a monthly figure on the scope of support, a milestone on what is being built. Every format has a fixed, predictable frame — but you get the number in the conversation, calculated for your case, in writing. Not in a footer meant to fit everyone.
“Terms belong in a conversation, not in a footer.”
Frequently asked questions
Often yes — for standard tasks such as copy, translations or meeting notes, where no sensitive customer data ends up in someone else’s systems. If the off-the-shelf tool is enough, that is exactly my recommendation — rather than building you something.
Because serious terms depend on team size, starting point and goal. Every format has a fixed frame — day rate, monthly figure or fixed price per milestone. You get the number in the free first conversation, in writing and calculated for your case.
The wrong project: spending months having something built that an off-the-shelf tool could do — or letting unmanaged use run without rules, which creates liability risks. Both cost more in the end than any of the four cost brackets.
Nothing. The initial assessment takes 30 minutes, costs nothing and ends with three starting points — in writing. Even if we never speak again afterwards, that half hour was well invested.
Free · three starting points in writing · no sales pressure