Companies keep sending their people to AI courses, and the people keep coming back with prompt tricks, a certificate, and no change in how anything works. Two weeks later the enthusiasm is gone. The problem is not the trainees, it is that the course was about AI in general, and their work in particular never appeared in it.
An AI training workshop built on your real systems
Specialty Tokens runs AI workshops and training as the entry tier of our engagement model, the format we call Training, alongside Tech and Transformation. We are a forward-deployed AI engineering firm in Vienna, and the sessions are taught by the people who build production AI systems for clients: founders Nik Redl (strategy and delivery) and Thomas Schlossmacher (engineering and platform).
The defining choice is what the training runs on. Not slides, not toy datasets, your systems. Before the workshop we connect approved AI tools such as ChatGPT, Claude, Copilot, or Gemini to the tools your team already works in, through the Model Context Protocol (MCP). That means the exercises are your actual work: the report someone assembles every Monday, the inbox that gets triaged by hand, the data that gets copied between the CRM and a spreadsheet. Participants leave with automations that already run, in their own environment.
The program has two tracks, because leaders and operators need different things. The leadership track is about judgment: where AI genuinely carries business weight, what it costs, what to govern, and what to ignore, the questions a board or management team must answer before any budget makes sense. The operator track is hands-on: the people who run the processes learn to delegate the repetitive parts to AI, with the guardrails that make that safe.
We ran this kind of enablement with Speedinvest, the pan-European VC, a team whose job is judging technology, which sets a usefully high bar for the trainers.
A workshop is deliberately the smallest way to work with us, and it is designed to produce a next step rather than a warm feeling. Most sessions end with a shortlist of processes worth automating, ranked by effort and payoff. Some companies take that list and run with it themselves, everything we teach is theirs to keep, with no per-seat fees and no dependency on us. Others hand the top of the list to our AI agent development practice, or take the free AI self-assessment first to see the whole picture in writing.
Either way, the goal is the same: a team that stops treating AI as a course topic and starts treating it as a colleague.
