Service
AI and automation
Advice, strategy and implementation around AI and data. From "we want to do something with AI" to a reasoned plan and something that works.
What it costs
AI work scan
Start here€550
fixed price, half a day
We go through your work and you get it in writing: what is possible, what it costs and what I advise against.
- Half a day together
- Written advice, usable without me too
- "Do not do it" is a valid outcome
- Deducted from follow-up work
Setup and implementation
from €1,500
fixed scope, fixed price
From advice to something that works, including an explanation for the people who will use it.
- Scope and price agreed in advance
- An explanation for the users
- A written record of which data goes where
- No upper limit, larger is fine
Strategy and larger programmes
€750 p/day
Several processes, advice on AI and data strategy, or support during rollout.
- Also when the question is still broad
- First delivery is a plan you can decide on
- Written down in advance what will be delivered
Not sure yet what you want? Then the work scan is the starting point. What you have built afterwards is up to you.
Right now AI is being sold to everyone by people who have never built any. I have. I am finishing my master’s in Artificial Intelligence at Utrecht University, trained my own models for my thesis, and at one of the largest telecom companies in the Netherlands built a tool that uses performance and call data to decide which staff need coaching first.
You do not need to know what you want yet. “We want to do something with AI” is a perfectly good starting point, and it is where most conversations begin. Working out what it should become, or whether it should become anything, is the work, not something you have to have figured out first.
That is exactly why I start with what it will return. A process that is too small, too changeable or too sensitive costs you more than it brings in. You hear that from me before you start anything, along with which process is worth doing.
Where it usually does work
Documents and quotes. Recurring writing where eighty percent is the same every time. This is what I devoted my thesis to: an assistant for tender and quotation documents, including the question of how you measure whether the output is good enough.
Sorting incoming questions. Incoming email and form submissions get sorted and a draft prepared. You read, adjust and send. Nothing ever goes out without a person having seen it.
Searching your own documents. Making manuals, procedures, contracts and old quotes searchable, so your colleagues get answers without having to ask you.
About the work scan
If you already have a process in mind, we take that one. If you do not, then working out which process lends itself best is exactly what the scan delivers. If you then commission follow-up work, I deduct this amount.
The scan stands on its own: you keep a reasoned recommendation you can act on yourself, even if you do not have me do the implementation.
Good to know
Nothing ever goes out without a person having seen it. Customer and personal data only go to services we have agreed on in advance. Before anything gets built, what will be delivered is written down. Arriving with a vague question is fine. That is the normal starting point. What I will not do is start building while nobody yet knows what the result should be: that is how AI projects cost money without producing anything.