AI workshops for teams
Berlin and remote
Most teams use a sliver of the AI tools they already pay for, slowly, and they don't trust the output. I run a hands-on workshop built on your team's real work, so they leave using AI with confidence and good judgment the same week. AI that speeds the work up without quietly wrecking the quality. Optional internal AI guideline as a second step.
A 30-minute call. I learn how your team really works. No pitch, no obligation.
They paste in a vague prompt, get generic output back, and either ship it or spend twenty minutes fixing it. Harvard researchers named this workslop: low-value AI content that looks like progress and creates none. In one 2025 study, people estimated around 15% of the work they review is exactly this. Speed went up, quality went down. The fix is context and judgment, and that's a skill you can teach in a day.
Real skills on real tasks: prompting, editing with AI instead of letting it write for you, and a fact-checking layer so nobody publishes something a tool made up.
The reason teams stall on AI is rarely the tools, it's permission. I name the fears in the room and give people a safe hour to try AI on something that matters.
Accuracy, data protection, and staying the decision-maker, so AI speeds the work up without degrading it.
Time back, every week.
Built on your team, not a template.
Work teams actually find useful.
Bring your numbers to the call and I'll size the return to your team. The page earns the call, the call sizes the value.
Role, task, audience, context, format.
Live on your own work.
A scoring pass, then a fix pass, on work a human drafted.
A second tool for real-time verification.
Set it once in a project, stop re-explaining yourself.
Get the model to interview you first.
The fears, the data limits, and how adoption sticks.
You are the decision-maker. Never outsource your reasoning.
I learn how your team uses AI now and where they're stuck.
Eight to ten questions that map real tasks, tools, and fears.
I build the workshop around your real use cases. No generic deck.

A hands-on session with a thread on judgment and responsible use running through it.
An internal AI guideline or a wider rollout, as a separate step.
This survey-first method is the same one a national newsroom and the AI working group at GIZ Türkiye asked for.
I'm Magda Sokolovic. I build websites and AI automations for a living, so practical AI is something I do daily, not a slide I borrowed. I'm also a yoga and meditation teacher, which is where the attention and judgment angle comes from. In 2026 I gave a talk at re:publica, Europe's largest conference on digital society, on training your attention in the AI internet. Since then I've trained a full newsroom team at Prabhat Khabar in India and I'm in discovery with GIZ Türkiye.

.jpg)

Designed around your team's real tasks from a pre-survey
Teaches judgment and responsible use, not just prompts
Built by someone who delivers AI work daily and trains attention
You keep your voice, your accuracy, and your decision-making
One-size deck, same for every audience
Prompt tips you can find free on YouTube
Theory from someone who doesn't do the work
Faster output, quietly lower quality
You lead a team that knows it should use AI better and doesn't know where to start. Whether you're a newsroom, an NGO or development agency, an L&D or People team, or a founder-led company, the workshop is built around your real work.
You want a one-hour motivational keynote, someone to promise AI replaces half your staff, or generic prompt tips.
Start with a 30-minute discovery call. I'll learn how your team really works, and if there's a fit, I'll send a short, scoped proposal with a clear price. No pitch on the call, and no obligation after it.
Practical. Responsible. Built on your real work.




