Service · Training

Teach your team to use AI well - without writing a line of code.

Most people on your team already use ChatGPT, Gemini, Copilot, or whatever showed up in their browser this month. Most of them use it badly - mediocre prompts, leaking sensitive data into public tools, picking the wrong tool for the job, missing the small wins that would shave hours off their week. We come in, sit beside them on the work they actually do, and lift the floor.

Last updated:

The problem

AI is on every desk in your company. Almost nobody is using it well.

There is no shortage of AI tools. There is a serious shortage of people who know how to use them. The gap is not technical - it is everyday craft: how to write a prompt that lands the first time, what context to give the model so the answer is not generic, what to never paste into a public chatbot, which tool fits which task, and how to fold AI into the meeting notes / email drafts / spreadsheet work people are already doing.

The difference between a team that has been told about AI and a team that uses it well is hours a week, per person.

What we cover

Six topics in every engagement.

  • Prompt engineering

    How to ask. What context to give the model up front so it does not have to guess. When to iterate vs start a fresh session. Concrete patterns people can copy from day one.

  • Context and awareness

    Why the same prompt gives different answers in different sessions. How to set up the opening of a conversation so the model performs at its best. When to keep a thread alive and when to throw it away.

  • Sensitive data hygiene

    What you never paste into a public AI tool. What is safe, what is risky, what an EU-hosted model changes. The internal rule of thumb each person can apply in two seconds before they hit send.

  • Right tool for the right job

    Mapped to your team. An accounting team learns Gemini in Word and Excel. A legal team learns long-context models for contract review. A marketing team learns the difference between writing models, image models, and search-grounded models. Real names of real tools, picked for the work your people actually do.

  • Workflow integration

    Small wins that need no automation engineer. The email assistant. The meeting-notes pass. The one-shot document generator. The translation second-opinion. The spreadsheet formula explainer. Each one shaves time off something they already do.

  • On-site walkthrough

    When we run on-site, we sit beside people on the team and look at their actual flow - the inbox, the spreadsheet, the document they are writing right now - and tune the AI step into it together. This is the part that does not transfer from a slide deck.

How this is different

Why most AI training does not stick - and how ours does.

There is a recognisable shape to AI training that the team forgets within a fortnight. Naming it is the first step to avoiding it.

Where most AI training stops
pre-recorded courses · generic templates · slide decks
  • Generic prompts on generic data - never quite fits the work your people do
  • No coverage of sensitive-data rules specific to your industry
  • Tool landscape presented in the abstract; nobody knows which to open tomorrow morning
  • No time at the desk on the actual flows the team runs
  • Knowledge fades within a fortnight without applied use
How we run our workshops
your tools · your work · your sensitive-data rules
  • Prompt patterns built around the tasks your team actually does
  • Sensitive-data rules adapted to your industry and your hosting choices
  • A tool map specific to your business area - accounting, legal, marketing, ops
  • On-site walkthrough on real flows (in 2-day and 5-day formats)
  • A runbook the team owns and adds to after we leave
Three formats

Pick the format that fits the team size.

  • 1-day workshop · €400

    Best fit for teams of 2-5 people. On-site by default; remote on request. Covers the six topics above in workshop form, then we plan 1-2 hours at the desk with each person on real flows. Outcome: the team walks out with everyday craft they did not have that morning.

  • 2-day intensive · €700

    Best fit for teams of 6-10 people. On-site by default; remote on request. Day one is the workshop; day two is one-on-one work - 1-2 hours at the desk per person on the actual files, the actual inbox, the actual reports. Outcome: every person on the team has the craft AND a worked example of what "AI fitted to my job" looks like.

  • 5-day deep-dive · €1,600

    Best fit for larger teams. On-site by default; remote on request. Full week, role-by-role - 1-2 hours at the desk with each person, plus a written runbook of "the AI patterns we use here" that the team owns and adds to. Outcome: AI literacy across the whole team, not just the keen ones.

FAQ

Four common questions.

Can we do this remote?
Yes. The 1-day works fully remote. The 2-day and 5-day are stronger on-site because the second half is sitting beside people on their actual screens - but we run the on-site portion remote when travel is impractical.
What language do you teach in?
English by default; Bulgarian on request. Mixed-language teams handled fluently.
Do we need to prepare anything?
No. We send a one-page pre-call form so we walk in knowing what your team does and which tools are already in use. That is enough.
What if our team uses different tools - paid ChatGPT, free Gemini, an internal AI tool, all of the above?
We work with whatever they actually use. The craft transfers. Prompt patterns, context handling, and sensitive-data rules apply across every tool on the market today.
AI that already runs

Pick a date when the team has 1-5 days clear.

We confirm scope on a 30-minute pre-call and send a written workshop plan before invoicing.

Live · powered by Gemma-4 · running on our hardware in Sofia