Skip to content

STRATEGICRepublic of San Marino

AboutContact
ItalianoEnglish
Client area
Contact

AI training

AI training by role and function: decision-makers, developers and users learn different things

In an organisation, the person preparing a bid uses AI, IT connects it, the legal team assesses it and the board is accountable for it. Each of them has different questions. That is why we build one track per function, each with its own practical exercise, and we keep a record of every activity.

Let’s talk about your programmeThe tracks
Example of an AI training programme. Five functions, from the board to the internal champions, each have common foundations, a function module, a practical exercise and a refresher. Beside it, the training record and the AI systems the programme relates to.
Example · the functions, activities and systems shown are not real
Tracks by function
board, IT, legal and HR, everyday users, internal champions
Practical exercise
every track ends with real work from that function
Training record
who completed what, on which system

01 · copy the link to this sectionThe situation

The board, IT and procurement don’t have the same questions

The board doesn’t need to learn how to phrase requests to an assistant. The IT team doesn’t need the introductory lesson.

The board wants to know which systems are in use and who is accountable for them. IT has to connect an assistant to the ERP without opening security holes. The lawyer has to read a contract with clauses they have never seen before. HR receives an offer for a tool that screens CVs.

Without answers, the price is paid later. A project gets approved without the right questions being asked. An agent is given one permission too many. A contract is signed without anyone reading the data clauses. On top of this, the AI Act requires organisations to prepare the people who use AI (the law calls this AI literacy). Since 3 August 2026, national authorities can check compliance.

Useful training starts here. What does each function decide, build or use, and what does it need to know to do that well?

Typical scenario

The organisation activates licences for an AI assistant and runs an introductory course. It is the same course for the chief executive, the systems administrator and HR. Everyone leaves with a general idea. Nobody leaves with the answers they need.

What the law requires

The same AI assistant in the centre and six functions with six different questions. The board asks who is accountable, IT which permissions to grant, legal whether the vendor uses the data, HR whether the screening tool is high-risk, users whether they can trust the answer, and the internal champions which task to start with.

02 · copy the link to this sectionThe tracks

Five tracks, one for each function

They all start from the same foundations: what AI can do, where it goes wrong and which systems the organisation uses. Then each function goes its own way.

AI

Leadership and the board

Which AI systems the organisation uses and who is accountable for them. How to read a risk without becoming a technician. Which questions to ask whoever proposes a project, before approving it. Practical exercise: a real proposal, examined as if it were on the agenda. For the board
AI IT and development How assistants that answer by citing documents work (the technique is called RAG), and how agents carry out actions in your systems. Permissions, logs, defences against instructions hidden in content, and how to measure quality and cost. Practical exercise: a small assistant built and tested in a separate environment, with test data.
AI Legal, HR and procurement What the AI Act and the GDPR require of organisations that use these systems. Which HR tools count as high-risk. Which clauses to look for in a vendor contract. Practical exercise: an AI service contract read as a group, with the changes to request.
AI Everyday users The usage rules for their own function, written down. How to phrase a clear request, how to check an answer before relying on it, and when to pass an answer to an expert colleague. Practical exercise: a few typical tasks from the department, done with the approved tool and then reviewed together.
AI Internal champions The people who will stay on as the point of reference for their department. They work with us on a real case: choosing the task, setting the criteria for whether it works and checking the results. Practical exercise: a case from the department, taken through to a trial with written criteria.

Usage rules by function and the systems registerBefore signing with a vendor

03 · copy the link to this sectionLeadership

For the board: the questions to ask before approving

Board members don’t need to know how to configure an assistant. They need to ask the right questions and recognise a vague answer.

The leadership track works on a real proposal from the organisation, anonymised if necessary, and examines it as if it were on the agenda. At the end, the board has a list of questions to put to every new project, and knows which function to put them to.

The questions we work on:

  • What will we use it for, and with what data?
  • Who is accountable for it, and who oversees it day to day?
  • What happens if it gets something wrong, and how will we notice?
  • What will it cost over the years, not just in the first?
  • How do we leave the vendor, if we need to?

When an answer is not convincing, the tools are there: a second opinion on the proposal before signing, and the rules and systems register before go-live.

Second opinion on AI projects and vendorsAI security and governance

Example: a board meeting agenda with a proposal for an AI assistant. It shows the questions the board asks before voting with the status of each answer, the decision postponed pending clarification, and the leadership training module completed.

04 · copy the link to this sectionHow we work

How a programme is built: from the map of roles to the record

The programme grows out of the map of systems and roles, not out of a course catalogue.

Map of roles and systems

Which AI systems are in use or on the way, who uses them, who administers them and who is accountable. If there is a register of AI systems, we start from there. If not, we sketch one out together.

Programme by function

For each function we set the objectives, content, practical exercise, schedule and format. The programme is approved before we start, with the people who handle the topic for HR and for IT.

Sessions

The sessions use the organisation’s real systems and documents, on a schedule agreed together, on site or remotely.

Practical exercise and record

Each track ends with an exercise based on the function’s own work. The record notes who took part, in which track, with which materials and with what outcome.

Updates

Tools change, new people join and roles shift. The programme is reviewed at an agreed interval, and the record carries on.

Timing
The schedule depends on how many functions and people are involved. It is set in the programme, before we start.
Who does the work
The proposal names who runs the sessions, function by function. On the organisation’s side, we need one contact person to follow the programme.
Costs
There is no price list. Costs depend on the functions, people and sessions involved, and on how far the materials need to be built around your systems. The first conversation carries no commitment. If there is work to be done, you receive a written proposal for the initial assessment, usually one or two days of work with the people involved. After the assessment comes the proposal for the project, with phases, timing and costs.
If something isn’t working
The practical exercises show it. We adjust the track for that function before moving on to the next ones.

Engagement formats

05 · copy the link to this sectionThe record

The training record: who was trained, and on which system

Sooner or later someone will ask what has been done. The record lets you answer with a document, not from memory.

For each person or group, the record notes:

  • function and role
  • the relevant AI system, taken from the systems register
  • the track followed and the materials used, with their date
  • when and how, on site or remotely
  • the outcome of the practical exercise
  • the next planned update

The record stays with the organisation, together with the approved programme, the materials and the completed exercises. It comes in a format that HR can keep with the rest of its training documentation.

It is not a certificate and it confers no titles. It documents what was done, for whom and on which system.

Example of a training record. For each function it shows the relevant system, the track, the format, the practical exercise and the status, with high-risk systems highlighted. Below, what is recorded and what is kept on file.

06 · copy the link to this sectionReference

What the law requires

Updated 16 September 2026

The duty to prepare people does not only apply to those who develop AI systems. It also applies to organisations that use them in their work. In brief:

LawWhat it requiresFrom when
AI Act, Article 4 (wording in force since 27 July 2026)Organisations that provide or use AI systems help their staff, and others who work on their behalf, to develop the knowledge they need. There is no fixed level to reachChecks from 3 August 2026
European Commission, questions and answers on AI literacy (updated 27 July 2026)No certificate is required. The organisation can keep an internal record of training and other initiatives. ‘Other persons’ include, for example, external contractors, service providers and clients. Supervision is carried out by national authoritiesSupervision and penalty rules apply from 3 August 2026
AI Act, Article 26(2)Organisations that use a high-risk system assign its oversight to people with the necessary competence, training and authority. According to the Commission, this obligation remains after the 2026 amendmentsAnnex III systems, including recruitment and staff management, from 2 December 2027. Annex I systems from 2 August 2028

This box is for information only and is not legal advice. How it applies to a specific case is assessed with your lawyers and your data protection officer. For businesses in San Marino the framework is different and must be assessed case by case.

Sources (consulted on 16 September 2026)

07 · copy the link to this sectionPrinciples

What we don’t do

  • We don’t sell pre-recorded video courses that are the same for everyone.
  • We don’t issue ‘AI expert’ certificates or similar titles. The record documents what was done. It does not certify anyone.
  • We don’t teach people to use tools the organisation has not approved.
  • We don’t replace your lawyer or your data protection officer. In the legal and HR tracks we explain the rules. The assessments remain theirs.
  • We don’t promise productivity gains. Through the practical exercises, we check what people can do by the end of the track.
  • We don’t use artificial intelligence to assess the people taking part.
A well-prepared board doesn’t know how to configure an assistant. It knows which question to ask before approving one.

08 · copy the link to this sectionFrequently asked questions

Frequently asked questions about AI training for organisations

Who should be trained first in an organisation?
It usually makes sense to start with the people who decide and the people who administer the systems. The board sets priorities and responsibilities, and IT makes them workable. Next come the functions that handle personal data or confidential documents, then the other departments that use the tools. The final order comes from the map of roles and systems in use, drawn up at the start of the work.
What changed for AI literacy with the 2026 amendments to the AI Act?
Since 27 July 2026, Regulation (EU) 2026/1744 has replaced the wording of Article 4 of the AI Act. Organisations that provide or deploy AI systems must take measures that help their own people, and those who work with the systems on their behalf, to develop AI skills. The text sets no minimum level. Preparing the people who oversee high-risk systems remains mandatory. Each specific case is assessed with the organisation’s lawyers.
Does every trained person need a certificate?
No. In its answers on Article 4, the European Commission makes clear that no certificate is required. The organisation can document what it has done with its own record of courses and guidance initiatives. That is why our programmes produce a record showing the function, the relevant system, the track, the materials, the dates and the outcome of the practical exercise.
Do you also train the IT team on the technical side?
Yes, and it is one of the main differences from a general course. The IT track explains how assistants that answer by citing documents work and how agents act on systems. It covers how permissions are assigned, how logs are read, how to defend against instructions hidden in content, and how quality and costs are measured. The work takes place in a separate environment, with test data.
How do you prepare the people who must oversee a high-risk system?
The AI Act requires organisations that use a high-risk system to assign its oversight to people with the necessary competence, training and authority. The track starts from the system’s instructions for use and its purpose. It covers what to check, when to step in, how to stop the system and how to record decisions. For Annex III systems, such as those used in recruitment, the obligations apply from 2 December 2027.
How much does a training programme cost, and how long does it take?
There is no price list. It depends on how many functions and people are involved, the number of sessions and how far the materials need to be built around the organisation’s systems. The schedule is set in the programme. The first conversation carries no commitment. If there is work to be done, you receive a written proposal for the initial assessment, usually one or two days of work with the people involved. After the assessment comes the proposal for the project, with phases, timing and costs.

09 · copy the link to this sectionObservatory

From the Observatory: artificial intelligence

Analysis and guides from the Strategic editorial team.

All articles on artificial intelligence

The essential parts of this page print as a short summary of a few pages, without images or invitations to get in touch.

10 · copy the link to this sectionContact

A programme for your organisation starts with a map of roles

A few pointers are enough. Tell us which AI tools are in use or on the way, which functions to involve and whether there is an internal deadline. We reply by email, usually within two working days.

Please do not put employees’ personal data or confidential documents in the form.

Or write to us:WhatsAppstaff@strategic.sm

On WhatsApp, messages are answered by Strategic’s AI assistant.

Your requestAIAI training

Do any of these questions sound familiar? tick the ones that apply to you

The form opens with the topic and the questions already filled in.