Article 4 requires documented AI literacy for staff who use AI systems or are affected by them: not certification, not an exam. You come away with a completion record for each person, plus an evidence pack: a training log, a role-to-AI-system mapping worksheet, an AI acceptable-use policy starter, and a one-page memo mapping the course to Article 4.

Our courses are built for baseline literacy: the basics everyone on a team needs. If some of your people are working with more advanced agentic systems, building AI solutions other people rely on, or in roles where a mistake could cause real harm, they need literacy beyond this course. We can help you figure out where that line falls for your team, and get a baseline in place quickly where it’s needed.

We specialize in small and medium organizations. If you’re just starting out, even as a team of one, cohort training needs a few more people to be worth it. The Article 4-focused course may be the better fit on its own. For larger organizations (500+ staff), we won’t turn you away, but we’re upfront that we may not have the team yet to pull off a rollout at that scale as well as we’d want to.

This is the objection we hear most, usually from someone who already knows it’s needed and whose leadership doesn’t yet. A few places this has fit for others: a professional-development or capacity-building line, a security or compliance training budget, or a digital-capacity grant, the kind that funded our work with NPB. If the EU AI Act’s Article 4 applies to you, that’s a compliance argument too, not just a nice-to-have. And the fit call itself is free, so it’s a reasonable first step even before budget is secured.

Some will, but usually alone, more slowly than it looks from the outside, and without the benefit of learning alongside colleagues. People get stuck with no one to ask. And the tools change fast enough that a mental model that was right six months ago can be wrong now. Self-teaching becomes close to a part-time job, and without protected time set aside for it, most people don’t keep it up. If your team has tried this already and it hasn’t taken hold, that’s often a sign the missing piece is shared structure and protected time, not effort.

Common, and usually a sign the training covered AI in general rather than your team’s actual work. Ours starts from one real workflow you choose in the first session, builds the skill on that, and ends each session with a specific, small step to apply that week, so by the end there’s a workflow that’s actually different, not just notes from a class.

That’s the point of starting from your own workflow instead of a generic case study, and of the fit call before anything is booked. We’re not going to know your context better than you do. The course is structured to build the skill using your real work, so the specifics come from your team, not from us.

Normal, and part of what we design around. Cohort programs can run a shared core plus role- or level-based tracks, so a complete beginner and a staff member already building with AI both get something worth their time instead of one generic session aimed at the middle.

Yes. Those are common asks from mission-driven teams and we scope them as applied modules inside a cohort program: how to use AI on a grant narrative without losing your organization’s voice or overstating results, and how to use it for research support without treating output as a finished answer.

Anywhere from a single hour-long session to a multi-week series, depending on what you’re solving for. A one-off session builds awareness and a few concrete skills. A cohort program builds a durable, shared practice. We’ll recommend a length based on your goals, not sell you the biggest package by default.

Either order works. Our training includes AI policy best practices built in, so it doesn’t assume a policy already exists. We support organizations before their policy is in place, during the process, and after it’s in place, including the leaders setting it.

We don’t provide legal advice, a security audit, or a compliance certification. High-stakes AI use in your programs still needs human review regardless of how much training your team has had. See the note below.