Skip to content
SSyntalith
PLEN
Book a scan

Your team already uses AI, just without boundaries or a trace.

Client data goes into random tools, no one can vouch for the output, and a decision leaves no trace.

Training is led by engineers with AI systems in production, not career trainers. Instead of trend slides we work on your real process: context, tool choice, boundaries, verification, and trace.

  • You talk to an engineer, not a salesperson.
  • 30 minutes, no sales presentation.
  • No prep or homework on your side.

In short

Price
from 4 900 PLN / day neta band of 4 000–6 000; quoted per scope after one call
Format
1 or 2 days, online or on-sitea closed team session up to ~15 people, multiple groups in a larger company
Who teaches
Engineers with systems in productionthe program comes from deployments, not slides; work on your process
For whom
Boards and teams that already use AInot for a trends lecture without work on your own tasks
Let's talk about training for your team

Engineers who deploy, not trainers.

The AI course someone on the team has already sat through usually ends in slides and prompt tricks. Here you are taught by the engineers who build and run those same systems in production for Polish companies.

Who teachesA typical AI courseA training facilitatorSyntalith trainingAn engineer who ships AI systems in production
On whatA typical AI courseGeneric examplesSyntalith trainingYour real process and data
What it teachesA typical AI coursePrompt tricksSyntalith trainingContext, boundaries, escalation, trace
ComplianceA typical AI courseSkippedSyntalith trainingOperator duties under GDPR and the AI Act
After the sessionA typical AI courseSlidesSyntalith trainingMaterials and a way of working that stay with you

See the deployments →

What your team learns.

Each module pairs hands-on work with Claude with the discipline that makes it trustworthy in a real process. We fit the scope to the team and the goal.

  1. Theory · the foundation

    How the models actually work and where they fail, on neutral examples before we work on your tasks.

    01Foundations · Model

    AI foundations: how models work and where they fail

    How AI models actually work and where they fail, before we touch your data: where errors and the only-plausible answer come from.

  2. 02Claude · Context

    Claude cowork: context, task and working together

    How to give the model the right context and task, how to run longer work, and how to recognize an answer that is only plausible.

  3. Practice · the drills

    Concrete drills on your process: tool, boundaries, escalation, trace.

    03Documents · Analysis

    Claude in the team's daily work

    Documents, offers, analyses, correspondence, your team's real tasks moved onto the model, with verification of the result.

  4. 04Tools · Boundaries

    Agentic workflow: prompt, automation, or agent

    When a conversation with the model is enough, when a repeatable workflow is better, and when an agent should use tools, and where a human must approve the result.

  5. 05GDPR · Escalation

    Data safety and oversight

    What can go into AI, what cannot, where the model must stop, and how to set escalation before something goes wrong.

  6. 06Trace · AI Act

    Trace, quality, and AI Act duties

    How to leave a decision trace, measure the quality of work with AI, and how training supports the literacy duty in Article 4.

We run most trainings on Claude, because that is what we build with in production, but we also work in OpenAI's tools (ChatGPT) and others, and if your team prefers a particular tool, we teach on it.

Three groups, three scopes.

Who will be in the room: the board deciding where AI makes sense, or the team that has to use it from tomorrow? Each group needs a different conversation, so we fit the scope to who is in front of us.

Boards and decision-makers

You decide where AI enters the company and at what cost, and the vendor offers all sound the same.

  • Where Claude and agents are an advantage, and where they are a risk
  • How to read an AI vendor's offer and what to demand in the contract
  • What the AI Act means for a company deploying AI
  • A short session and operational decisions, not a trend lecture

Operations teams

Your people already reach for AI on the job, each in their own way and with no rules.

  • Hands-on work with Claude on your real tasks and data
  • Context, boundaries, and catching the plausible-but-wrong answer
  • What can go into the model and what can't, data-handling rules
  • Materials and a working discipline that stay with the team

Team leads

You have to roll AI out in the team so it sticks and doesn't fracture into 20 styles.

  • How to spread AI work across the whole team
  • Shared rules and a working discipline across the whole team
  • What stays in the team after the training, and how to keep it
  • Keeping quality and a trace in daily work

Formats and prices

from 4 900 PLN

per day · net

One day or two. A per-day price, stated up front.

The same formats, online or on site: for one team (up to ~15 people) or several groups across the company, on one consistent way of working, not 20 styles. The final quote depends on scope, headcount, and the number of groups.

1 day · introduction

Boards, decision-makers, teams starting with AI

One day on your own examples: what Claude and agents do well, where the boundaries are, and which data-handling rules to adopt from tomorrow.

from 4 900 PLN

per day · net

2 days · workshop on your process

Teams that will use AI daily

Day one: hands-on work with Claude on your tasks. Day two: boundaries, escalation, trace, and a working discipline that stays with the team.

from 4 900 PLN

per day · net

The day rate usually sits in the 4 000–6 000 PLN net band, and you get a concrete quote after one conversation about the team and the goal. Count the hours your team loses each week fixing what AI produced without verification, and compare that with one training day. You keep the program, the attendance list and the materials.

Let's talk about training for your team

Who this is for, and who it isn't.

We pick the group above; this is about readiness. If the second column describes your case, another provider usually fits better, and we'll say so before the quote.

Ready if

  • You have a real process on the table, not just curiosity
  • Your team is about to reach for AI on the job and needs boundaries first
  • You care about a decision trace and the Art. 4 AI Act literacy duty
  • You want to work on your own data, not generic examples
  • You want working with AI to become a habit for the whole team, not just a couple of enthusiasts

Probably not if

  • You want a trend lecture with no work on your own tasks
  • You only need prompt tricks and a certificate for the wall
  • You want cheap off-the-shelf training with no fit to your process
  • You expect a guaranteed result or an official endorsement from the model vendor
  • You are a technical founder who wants to build software with agents: see the AI-Native course

AI Act · Art. 4

AI skills that stay with your team.

When a client in a tender, or your own legal team, asks how you document AI competence across the team, do you have an answer? Article 4 of the EU AI Act requires companies deploying AI to ensure a sufficient level of AI literacy among the people who work with it. A closed team training is one way to meet that duty. After it you keep the program, the attendance list and the materials, ready for your records. This is not legal advice: confirm the scope with your own counsel.

EU AI Act 2024/1689, Art. 4 (EUR-Lex) →

From a call to a program that stays.

Training starts with one conversation about the team and the goal. We run the rest in a clear order.

  1. 01A call about the team, process and goal: we set the group, format and scope.
  2. 02You get a quote and a program tailored to your tasks and data.
  3. 03We run the training online or on site, on real examples.
  4. 04You keep the materials, attendance list and a working discipline for Art. 4.

No seats sold off the shelf: every training is quoted per scope after a call.

Tell us who we're training and we'll send back a program and a quote.

  • 30 minutes with the engineer who would build it, not a salesperson.
  • A review of the processes that cost you the most time and money.
  • A written summary: what to automate, in what order, with cost ranges.
0 PLN30 minutes · written takeaway within 2 business days
Let's talk about training for your team

No sales deck and no obligations. If automation doesn't make sense, we'll write that too.

Full pricing

from 4 900 PLN/day net · quoted per scope · taught by engineers with systems in production

Training questions

  • How soon can you run it?

    Usually within a few weeks of the first call: that's the time to set the group and scope and tailor the program to your tasks and data. We'll try to fit an urgent date if the calendar allows.

  • How many people can take part?

    A group of up to ~15 people as standard: above that, a workshop stops being hands-on work and becomes a lecture. Larger teams are split into groups and priced in one offer.

  • Online or on site?

    Both. Operator workshops work best on site at your company; board sessions often run online. Logistics and travel are included in the quote.

  • Do we need Claude accounts and licenses?

    We settle participant access to Claude during scoping: if the company doesn't have licenses yet, we'll advise which plan makes sense for the team. We don't sell licenses and don't require a purchase before the training.

  • Why engineers and not trainers?

    Because the program comes from deployments, not presentations. The same people build and run AI systems working in Polish companies: in the training they show the working discipline they use in production themselves.

  • How do you know it sticks, instead of slipping back to old habits by Monday?

    The usual end to AI training is enthusiasm on Friday and old habits on Monday. That's why day two isn't more theory: it's shared rules and a working discipline written down on your own tasks, plus materials and a list of decisions that stay with you. One day narrows the gap, it doesn't make the team experts: what you do with it next is up to you, but you leave with a ready way of working, not just a deck.

  • We're a small, not very technical team. Is this for us?

    Yes. This is training for people who work with documents, emails and data, not for programmers. A small team runs as one group and a shorter format. If someone on your side wants to build software with agents, that's a different course, and we'll say so plainly.