AI Foundation
A basis your management can actually decide on.
Before the first application gets built: a prioritised list of use cases, a vendor-independent technology recommendation, a data basis that holds, and a written AI policy.
About this service
Most companies have already tried AI — and most of those attempts stayed pilots. Not because the technology was missing, but because the foundation was: unclear ownership, unstructured data, no answer to the data protection question, and a model chosen on instinct. The AI foundation starts before all of that. What you have at the end is a document a budget decision can be made against — and one that works with any other provider too.
What's included
Analysis and selection
- Taking stock of processes, systems and data
- Prioritising use cases by effort and effect
- Evaluating suitable language models
- Cost-benefit analysis of the infrastructure options
- A vendor-independent recommendation, reasoned in writing
Foundations and rules
- Structuring unstructured company data
- Checking GDPR compliance per use case
- Developing an AI policy for internal use
- A role and ownership model
- A roadmap with sequence, effort and dependencies
How the project runs
Four phases to a basis that can carry a decision. What you have at the end is a document, not an impression.
Taking stock
We look at which processes, systems and data actually exist — and where ownership currently ends. Without this step every roadmap is a guess.
Prioritising use cases
Together we assess candidate use cases by effort and effect. The result is a short list ordered by sequence, not by enthusiasm.
Technology and model selection
Evaluating the language models, a cost-benefit analysis of the infrastructure, a recommendation reasoned in writing. We do not sell licences, which is what lets us be straight here.
Roadmap and AI policy
An implementation plan with sequence, ownership and effort — plus an AI policy setting out which tools are permitted for which content.
Where it is used
The foundation is worth it wherever more than one department is pursuing its own AI ideas and nobody is setting the order.
- A first AI initiative with no prior experience in house
- After a pilot that never reached production
- Ahead of a larger budget decision
- With several competing ideas from different departments
- Before choosing a provider or a platform
- When data protection is the open question
What to watch for
Three points where work like this typically gets difficult — and how we handle them.
An analysis with no delivery stays paper
Which is why the roadmap ends with concrete first steps and named owners. Whether we deliver them is your call afterwards — the document works with any other provider too.
Data quality is routinely underestimated
We check it in phase 01 rather than when an application fails on it. Where substantial work is needed it appears in the roadmap with an estimate.
Without the departments there is no basis
Prioritisation happens with the people who run the processes daily. A list produced with IT alone prioritises the wrong things.
Why BLAID for this
Four reasons this differs from a vendor presentation.
No licences in our portfolio
We do not sell software licences and take no commissions. That is what makes a vendor-independent recommendation possible at all.
We also deliver
The recommendations come from people who build and run these systems themselves, not from consulting with no delivery experience.
A result in plain language
The document is written for management, not for a technical audience. Whoever reads it can decide afterwards.
Works without in-house IT
We regularly work with companies that have no IT department, coordinating directly with management and the relevant departments.
Where should your first AI project start?
Let us look at your starting point and design an order that holds.
Advice from specialists
A conversation with people who build these systems themselves.
Predictable delivery
4–8 weeks to a finished roadmap, with results along the way.
Secure and accountable
Data protection assessed per use case, an AI policy as a written deliverable.
What you get
- A prioritised list of real use cases
- A reasoned technology recommendation
- A roadmap with effort and sequence
- An AI policy for internal use
- Clarity on the data protection question
- A basis that works without us too
Free first call · no obligation · an individual solution