Artificial intelligence
Where AI makes sense, how to test it with little risk and how to take it to production with your data under control.
What it is
Artificial intelligence can save a great deal of work, or become a cost with no return and a new risk. The difference usually lies in choosing the right problem, data and architecture before the first line of code is written.
We help you decide where it makes sense, test it with little risk and, if it works, take it to production with the data under control and the legal obligations clear.
What we do
- Use cases We identify where AI adds value in your organisation and where it does not, with an estimate of cost, benefit and risk for each case.
- Proof of concept A prototype bounded in time and budget, using your own data and built by your team or a vendor of your choice under our direction, to check that it works before you invest seriously.
- Architecture and model choice Cloud services or models that run on your own infrastructure, depending on how sensitive the data is, the cost and the control you need.
- Data and compliance Which data can be used and how, under the GDPR and the EU AI Act (Regulation (EU) 2024/1689), which sets obligations according to the risk of each use, including telling people when they are talking to an AI.
- AI security Assistants and automations that do not leak information or follow instructions they should not, with the same access control as the rest of your systems.
- Implementation oversight We coordinate development with your team or the vendor you choose, and check that what is delivered does what was agreed.
What you receive
- A prioritised list of use cases, with what each one costs and what you can expect from it.
- If there is a proof of concept, its results and an honest report on them.
- A documented architecture and the obligations that apply to you, in writing.
Frequently asked questions
Do we have to send our data to an outside provider?
Not necessarily. Some models can run on your own infrastructure, so the data never leaves it. If an external service is used, we choose one with proper contractual guarantees and limit what is sent to it.
Does the EU AI Act apply to us?
Almost certainly in some way. Most uses carry limited obligations, such as transparency, but some are prohibited and others are considered high-risk, with strict requirements. The first step is to know which category each use falls into.
How much does a proof of concept cost?
It depends on the case and the data available. After a first conversation we give you a fixed scope, timeline and price.