People and skills
Programs per role, from basic understanding for everyone to depth for the people who assess and design AI. Documented and maintained, in line with Article 4 of the AI Act.
Since 2 February 2025, Article 4 of the AI Act requires organizations to ensure a sufficient level of AI literacy among staff who work with AI. Apart from that obligation, it is the precondition for everything else you want to do with AI: people who understand what the technology can do use it better and ask sharper questions about it.
One training for everyone rarely covers the obligation. An executive who decides on AI investments needs different knowledge than an employee who works with Copilot every day, who in turn needs different knowledge than the buyer who assesses a vendor. That is why we organize AI literacy per role, and record who holds which level.
The law describes AI literacy as the skills, knowledge and understanding needed to deploy AI systems in an informed way and to recognize their opportunities and risks. The required level depends on context: the employee's role, the systems they work with and the people affected by the outputs. To us it means that an employee can do four things, in increasing depth.
Understand. Know how a language model or another AI system arrives at its answers, why it can sound convincing and still be wrong, and which rules the organization applies. This level is for everyone.
Apply. Use the tools the organization has well: formulate a task, give context, know which data may go in and check the output before it goes out the door. For everyone who works with AI.
Assess. Be able to vet applications and vendors on risk, quality and compliance, organize human oversight and decide on deployment. For management, HR, procurement, compliance and information security.
Design. Build applications, assess data and models and draw up the technical documentation the law requires. For IT, data and innovation.
| Role | Understand | Apply | Assess | Design |
|---|---|---|---|---|
| Board and executives | Required | Basic | Required: Decide on deployment, risk and investment. | Not needed |
| Managers and team leads | Required | Advanced | Required: Assess applications in their own process and organize oversight. | Not needed |
| HR, compliance and procurement | Required | Advanced | Required: Vet vendors and high-risk applications. | Basic |
| Knowledge workers | Required | Required: Daily use: formulate tasks, check outputs. | Basic | Not needed |
| Operational staff | Required | Advanced | Not needed | Not needed |
| IT, data and innovation | Required | Required | Required | Required: Develop applications, assess data and models. |
We start with an inventory: which roles work with which AI systems, formally and informally, and which level of knowledge does that require? Often an organization turns out to use more AI than the board knows, because vendors have switched on AI features in existing software. The inventory ties in with the AI register from our governance service. Those who already have one save time here.
Program per role. Based on the inventory we compose a program: a short basic module for the whole organization, practical workshops per team with their own cases and work material, and in-depth training for the roles that assess or design AI.
Delivery with School of Data Science. The training is given by School of Data Science instructors who work with the technology every day. The training AI literacy and the training AI for management teams and boards form the basis. Where needed we add the Copilot series or tailored sessions for a specific role.
Working agreements. Every practical workshop closes with agreements the team makes itself: which tools for which work, which data may and may not go in and how outputs are checked. Those agreements are short, concrete and the team's own. That is why they are kept.
Recording. We record the required level per role, the training each employee has completed and how the organization keeps this up to date. That way you can demonstrate that Article 4 is met, without a certificate the law does not ask for.
Who it is for. Boards and management teams that want to know where to begin, organizations that get stuck after a number of pilots, executives preparing for the obligations of the AI Act and organizations that want to build an AI roadmap on an honest foundation.
Besides the training itself you get an overview the organization can maintain on its own: which role needs which level, where everyone stands and what has to happen each year to keep it that way.
Deliverable
The AI literacy program
One overview with which the organization organizes, demonstrates and maintains AI literacy.
A 10- to 15-page program document, a register per role and employee, and an annual review.
Roles, AI systems and the required level per role.
Modules per level, formats and planning per team.
Per team: which tools, which data and how outputs are checked.
Who has completed which training and when.
Onboarding, new tools and the annual review.
How the whole meets the obligation, ready for a regulator.
The overview is both the evidence for a supervisor and the starting point for the rest of your AI roadmap. Staff who understand what AI can do come up with better ideas and recognize risks earlier. How to meet the obligation per role, and which mistakes organizations make along the way, is explained in the article AI literacy has been an obligation since 2025. Here is how to approach it.
AI literacy is not a one-off. Tools change, staff come and go and the law is being worked out in guidelines. That is why we agree at the start how the organization keeps the level up.
New staff. The basic module becomes part of onboarding, so that nobody works with AI without knowing the rules.
New tools. For every new application or newly enabled AI feature, we assess which roles need extra knowledge, and that knowledge is arranged before the tool goes into use.
Annual review. Once a year the overview per role is updated, the working agreements per team are revisited and the program is adjusted to what has changed.
Staff who use AI purposefully and safely, managers who can assess applications and an organization that demonstrably meets Article 4 of the AI Act. And a foundation for every next step of your AI roadmap: people who know what they are doing.
Duration. The inventory and program design usually take four to six weeks. Delivery of the training follows in the months after that, at a pace that suits the organization.
Who it is for. Organizations that want to meet the AI Act obligation properly, management teams that want to make better decisions about AI themselves and teams that work with AI tools and want to get more out of them.
Current
Microsoft 365 Copilot implementation
Implementing Microsoft 365 Copilot in five phases: Readiness, Strategy, Pilot, Rollout and Scale, with security, governance, adoption and measurement as continuous layers.
A free one-hour conversation
A free one-hour conversation with a Synaura consultant. You leave with an honest picture of where your organization stands and what a sensible first step looks like.
What you get
Other services
A free one-hour conversation about how ready your organization is for AI, with a report within a week: opportunities, risks and a concrete starting point.
An AI strategy with real choices, an opportunity assessment of your operations and a roadmap that puts opportunities, risks, people, processes and technology in the right order.
Redesign of work processes with the people who do the work, vendor-independent set-up of tooling and guidance of the rollout until the new process is in place and the results are measurable.
A light and workable governance framework for AI: an AI register, a policy people actually read, clear roles, AI Act compliance per system and a vendor check, fitted to how your organization decides.
