strategy
Why most AI plans end up in a drawer
Many organisations have a list of AI ideas and no idea which one to do first. The idea that gets shouted loudest rarely has the best data underneath it. For each opportunity we look at three things: is the data there, can the result be measured, and does it deliver enough to be worth taking into operation later. Whatever does not pass, we drop.
What you get
- 01
Prioritised list of opportunities
For each one: which problem it solves, who will work with it and what it costs to build and to run.
- 02
Data check per opportunity
For every opportunity we look at whether the data is there, where it sits and what its quality is. No data, no project.
- 03
Measurement plan
How we establish per opportunity whether it works: which questions we ask and which answer is good enough.
- 04
Proposal for phase 1
A well-scoped first project with a benchmark underneath, so that at the end you know whether it works.
What we build
View our workFeasibility study
Test technically and on your data whether an AI use case can work, before you build.
Benchmark setup
A yardstick of question-answer pairs so quality is measurable from day one.
Architecture and model choice
A well-founded choice between open and closed source, and between cloud and on-premise.
How we go about it
Our way of working- Step 1
Talk to the people who do the work
We start on the shop floor, not in the boardroom: what takes time now, where does it go wrong and what is being worked around.
- Step 2
Sharpen the problem
What the organisation really needs, not what is being asked for. The question often turns out to differ from the brief.
- Step 3
Data exploration on the most promising ideas
We look at the real data, not the data model on paper. That way you know in advance where it will chafe.
- Step 4
Roadmap with order and cost
What first, what later, what not. With, per step, what it costs and what it has to deliver to continue.
Work in this line
All cases
Making support schemes findable for entrepreneurs.
With Rewire and Dialogic, an AI prototype that improves the findability, transparency and effectiveness of support for SMEs, starting with digitalisation.
A reliable AI assistant for policy officers.
A sovereign AI assistant focused on reliability and transparency, built on soev.ai with Intermax and UbiOps.
Querying construction drawings and IFC files.
Two agents in Google Gemini Enterprise for IFC files and construction drawings, with a citation for every answer.
Frequently asked questions
How long does a strategy engagement take?
That depends on the number of ideas and the state of the data. We agree a fixed lead time in advance and work towards a concrete list, not a report that keeps growing.
Does our data have to be in order already?
No. It is precisely in this phase that we look at how the data really stands. That helps determine which opportunity goes first and which has to wait a little longer.
Can you also build what comes out of it?
Yes, that is usually the next step: phase 1 of a bespoke project, with the benchmark from the measurement plan underneath. You can also carry out the roadmap with another party.
Grab a coffee with Daan. We think along, no strings attached.
We would love to hear what AI challenges you run into within your organisation. Together we figure out what the first step would be.