Service · Artificial intelligence

AI strategy

Suppose you run a law firm in Nova Gorica with fifteen people. Over coffee you discover that a trainee drafts case law summaries in a free chatbot, the secretary pushes letters for Italian clients through an online translator, and one partner uploaded a draft contract to a tool promising “review in ten seconds”, just out of curiosity. Nobody meant harm, and nobody made a decision. That is exactly the issue: those files carry clients' personal data and material covered by professional secrecy, and nobody can say where it ended up or who can read it now. An AI strategy starts by filling that vacuum. It records what is permitted, what is not, and who gives the green light to a new tool. Next, it captures ideas on one-page cards, scores them on value, practicality and risk, and orders them into a costed plan spanning one to two years. We run it as brief online workshops, so client work does not pause. The outcome is a small bundle of documents that the partners can sign off and the whole team can actually follow.

Tool survey
what is used, fed with which data
Use-case cards
value, practicality, risk
Policy
permitted, prohibited, who approves
Plan
one to two years, costed

Scope of the service

You will not receive a presentation about trends. You will receive documents that leadership signs and each department applies.

Check the scope with an engineer

Which tools, which licences

An anonymous questionnaire plus IT logs reveal everything in play, chat assistants, browser add-ons, meeting transcription bots, and what information goes into each. The point is a truthful snapshot, not blame.

Where the sensitive data sits

Which client files include health or other special-category data, which hold trade secrets, and in which systems they are stored. Only with that map can you say which tools may go near which information.

Use-case cards

Every proposal is captured on a single sheet stating its owner, the success measure, the data it relies on, an approximate cost and the main risks. That lets leadership weigh proposals like for like.

The legal picture

GDPR and ZVOP-2 wherever personal data is involved, guidance from the Information Commissioner on AI as a practical reference, and the AI Act to clarify in what capacity you use each tool. Client confidentiality clauses get reviewed too. The legal opinion stays with your lawyer; we bring the facts together.

Using what you pay for

Plenty of proposals can be met with Microsoft 365 or Google Workspace features already in your subscription. We show the limits of those before any new purchase is considered.

AI policy

Approved tools, categories of information that never enter any tool, how a new tool gets approved, who answers for checking output and whom to ask in case of doubt. No more than two pages, so it gets read.

Phased, costed plan

Phases lasting three to six months, cost estimates, the in-house people each phase needs and decision gates where leadership approves the next move.

How the work runs

Scope is agreed in advance based on how many departments take part. Workshops are short and online, so client matters keep moving.

01

Agreeing the limits

A single meeting with leadership pins down what AI ought to achieve and where it is not welcome.

02

Fact-finding

Questionnaire, interviews with each team, a look at systems and data, and a check of client agreements.

03

Scoring session

Team leads fill in the cards with us, rate them and discard the weak ones.

04

Delivery

Policy, plan, three concrete first actions and, on request, a template reply for clients who ask how you use AI.

A ban without an alternative does not work. Prohibit AI outright and people will carry on using it on their personal phones, invisible to you and outside any contract. A better route is to provide one sanctioned tool under a business agreement with known data terms, together with a brief rule on what must never be typed into it. When staff know what is allowed, shadow use drops off fast.

Your questions answered

At ten or fifteen people it is not a hefty report, just a few pages: the tools you may use, the information that is never entered, and the three trials planned for next year. A few workshops are enough to produce it, and it pays off at once because people stop guessing.

The AI Act is an EU regulation, so it applies directly throughout the Union, Slovenia included. What it asks of you depends on the purpose of the AI system and on whether you build it or merely use it. In the strategy we mark the uses that need a closer look; your lawyer carries out the assessment.

The Information Commissioner in Slovenia has issued guidance and recommendations on personal data in AI tools. In general terms the message is that GDPR continues to apply when a model handles the data, and that the controller must know where it flows. We turn that guidance into everyday rules inside your policy.

The order is the reverse. The strategy reveals which data matters for which proposal, and you tidy it only where it truly counts. Trying to perfect every dataset in advance becomes a never-ending exercise for most small firms.

After an initial conversation we send a quote built on our €50 per hour plus VAT rate. The number of departments and interviews has the biggest influence on the total.

Turn guesswork into decisions

Let us know which AI tools are already around, which ideas keep coming up and whether clients have asked how you use AI. We will propose a sensible scope.

When we work
Weekdays 9:00-17:00 CET, answers the same or next working day
Calls
Video meetings, Teams and Meet both welcome

We set essential cookies only: the ones that keep the site running and recall which town you chose. Nothing for ads and nothing that follows you around the web. Details are in our privacy and data protection notice.