AI Readiness Assessment
Find out where AI pays off in your company, before you spend on it.
A fixed-scope review of your processes, data, systems and team. You get 3–5 use cases ranked by value and feasibility, the risks, the architecture options and a written roadmap with a business case.
AI READINESS ASSESSMENT
Start with the right problem, not with a tool.
A short, fixed-scope engagement that tells you what to build first, what to skip and what it takes.
Most companies we talk to know AI could help them, but not exactly where. The typical first attempt is a chatbot pilot nobody uses, a licence for a tool that does not fit the data, or a project that stalls the moment someone asks where customer data is processed. The cause is almost always the same: the work started from the technology instead of from a process with a real, measurable cost.
The AI Readiness Assessment puts things in the right order. Over an agreed, limited period we look at how work actually gets done, what data exists and in what shape, which systems need to talk to each other, and which rules apply. It is built for SMEs and mid-market companies that want to start with AI without spending budget on experiments that lead nowhere. Our team has been building software for companies since 2013 — integrations, e-commerce, mobile apps and custom systems — so we treat AI as one more part of your systems, not as a separate experiment.
Use cases, ranked
We collect candidate use cases from your processes and score each one on business value, feasibility, data availability and risk. You get the 3–5 worth doing, in order, plus the ideas worth parking.
Data and systems readiness
We check where the relevant data lives — ERP, CRM, documents, email, databases — how clean and complete it is, who can access it, and what integration work each use case needs before any model is involved.
Risks and architecture
GDPR and EU AI Act classification for each use case, and a clear recommendation: cloud API models, a private model on dedicated infrastructure, or a mix — with the trade-offs in cost, control and effort.- Interviews with process owners and key users
- Map of the processes in scope and where time is lost
- Long list of AI opportunities, scored on value and feasibility
- 3–5 prioritised use cases with success metrics
- Data inventory and data-quality findings
- Integration map: ERP, CRM, documents and internal tools
- GDPR and EU AI Act risk classification per use case
- Architecture options: cloud API, private AI or hybrid
- Business case with cost drivers and expected benefits
- Phased roadmap: pilot, production, scale-up
HOW IT WORKS
Six steps, a fixed scope, one written result
Short enough to keep momentum, thorough enough to base a decision on.
Kick-off and scope
We agree the departments, processes and questions in scope, who we talk to and the delivery date. Scope and timeline are fixed before we start.
Interviews and process review
Working sessions with the people who do the work: where time goes, where errors happen, which decisions wait for information.
Data and systems review
With read access or guided sessions we look at the systems and data behind each candidate process: availability, quality, structure and access rights.
Scoring and prioritisation
Each candidate is scored on value, feasibility, data readiness and risk. We review the ranking with you and challenge it together.
Risk and architecture
For the top use cases we classify GDPR and EU AI Act obligations and outline the architecture: which models, where they run and how they connect.
Roadmap and business case
A written report and a presentation for management: what to do first, what it needs, what it should return and how to measure it.WHAT WE LOOK AT
Cloud API or private AI? An agent or a simple automation?
The right answer depends on the data, the process and the rules — not on the latest model.
Not every use case needs the same setup. Sending a public product description to a cloud model is low risk; sending contracts, health records or production data is a different matter. For each use case we recommend where the model should run: a cloud API (OpenAI, Anthropic Claude and similar) when quality and speed of delivery matter most and the data allows it; open-weight models such as Llama, Mistral or Qwen, served with tools like vLLM on dedicated infrastructure, when data has to stay under your control — including on our own GPU cloud in Romania, on our own network (AS56951). See private AI for how that works.
We are also direct about the type of solution. Some problems are best solved with a well-designed integration or a rule-based workflow, with no language model at all. Others need search and answers over your own documents, and some justify AI agents that take actions across several systems. The assessment says which is which, so you do not pay for more complexity than the problem needs.
On the regulatory side we flag where a use case could fall into a high-risk category of the EU AI Act, which transparency obligations apply, and what GDPR requires for the data involved: legal basis, processing agreements, retention and where the data is processed. If deeper work is needed, it continues in EU AI Act compliance. We are engineers, not lawyers — we work alongside your legal adviser, not instead of them.
FAQ
Questions we usually get
How long does the assessment take?
It depends on how many departments and processes are in scope, which we agree at kick-off. For one or two business areas it is a matter of weeks, not months, and the delivery date is fixed before we start.
What do you need from our side?
A sponsor on the management side, a few hours from the people who run the processes in scope, and read access or guided walkthroughs of the relevant systems. We do not change anything in your production systems during the assessment.
Do we have to continue with you afterwards?
No. The report is yours and is written so that any competent team, internal or external, can act on it. If you do continue with us, it becomes the starting point of the delivery phase — see our AI engagement models.
How is the assessment priced?
As a fixed price for a fixed scope, agreed in writing before we start, so there are no open-ended hours. For the implementation phases that follow we also offer value-based models.
How do you handle our data during the assessment?
We sign an NDA and, where personal data is involved, a data processing agreement. We work with read-only access or screen-shared sessions, we do not need copies of production data, and nothing is sent to external AI services without your explicit agreement.
Not sure where AI fits? Start with an assessment.
Tell us about your company and the processes you have in mind, and we will propose a scope and a delivery date for the assessment.