AI training on a real workflow of your team

Hands-on courses on generative AI, prompt engineering and AI agents. Before we start, we choose one of your team's tasks with you and agree on the expected output, the checks and the measurement; at the end the team keeps a procedure it can reuse.

  1. 01 A real team task
  2. 02 Expected output and quality
  3. 03 A reusable procedure
Book the first call See the courses

Dates are agreed with each company. The first 30-minute call is free; the cost is set in a written quote after the analysis.

Our AI training courses

Three courses in a classroom or online, adapted to the workflow the team brings to the first meeting, plus a self-paced course in English. Every course includes materials and templates the team keeps.

AI Executive Mastery

Beginner

For managers and department heads. Understand what generative AI does today, choose the processes where it helps and lead its adoption in your own office.

  • 16 hours, 2 consecutive days
  • Up to 20 participants
  • Materials included

What you will learn

  • What generative AI can and cannot do today
  • How to choose candidate processes
  • Starting measurement and result metrics
  • Examples on typical SME processes
  • Hands-on exercises with AI tools

What the course leaves you

  • A prioritised list of candidate processes
  • An adoption plan for your own office
  • Criteria to assess AI projects and vendors
Ask for dates and agenda
Agenda for the two days

Day 1: fundamentals and choosing processes

  • 9:00-10:30: Introduction to generative AI
  • 10:45-12:15: Language models and applications
  • 13:15-14:45: How to choose candidate processes
  • 15:00-16:30: Hands-on exercise with ChatGPT

Day 2: measurement and adoption plan

  • 9:00-10:30: Starting measurement and metrics
  • 10:45-12:15: Examples on business processes
  • 13:15-14:45: Adoption plan
  • 15:00-16:30: Review of the work and questions

Prompt Engineering for Business Workflows

Intermediate

For people who produce content, analysis or code. Prompt design techniques for reliable, comparable outputs on the team's recurring tasks.

  • 24 hours, 3 days, not necessarily consecutive
  • Up to 20 participants
  • Materials included

What you will learn

  • Principles of prompt design
  • Chain-of-thought techniques
  • Differences between AI models
  • Prompt templates for specific use cases
  • Automating recurring prompts

What the course leaves you

  • Shared criteria to judge output quality
  • A test method to compare prompts and iterations
  • A library of ready-to-use prompt templates
Ask for dates and agenda
Prerequisites
  • Basic experience with AI tools (ChatGPT, Claude or similar)
  • Basic machine learning concepts
  • Experience in software development or content production

AI Agent Developer

Advanced

For developers and technical teams. Design, release and run AI agents connected to company systems, with checks, logs and hand-offs to a person.

  • 32 hours, 4 days, not necessarily consecutive
  • Up to 20 participants
  • Materials included

What you will learn

  • Architecture and design patterns for AI agents
  • Integration with existing company systems
  • Monitoring, logging and debugging
  • Release and long-term operation
  • Security, permissions and hand-offs to a person

What the course leaves you

  • A verifiable prototype on an agreed use case
  • A reusable development structure
  • A checklist for checks, exceptions and go-live
Ask for dates and agenda
Prerequisites and technologies

Prerequisites

  • Software development experience (Python or JavaScript)
  • Basic knowledge of APIs and microservices
  • Familiarity with the cloud (AWS or Azure)

Technologies used

  • LangChain, OpenAI API, Anthropic Claude
  • Docker, Kubernetes and monitoring tools
  • Vector databases and embeddings
A participant in a training session works on an AI interface on a laptop while the trainer explains a diagram on the whiteboard

AI Operator Series

Self-paced

Online, at your own pace. Four books by Gregor Maric on MCP, AI agents, no-code automation and one-person AI businesses, each with a free companion page.

  • Self-paced, online
  • Four books, free companion pages
See the AI Operator series

Who takes part, and at what depth

According to the ISTAT report Imprese e ICT, anno 2025 (December 2025), among Italian companies with at least 10 employees the use of at least one AI technology rose from 8.2% in 2024 to 16.4% in 2025, and almost 60% of those that had considered an AI investment without making it cite the lack of adequate skills among the obstacles. Training everyone the same way does not solve this: you need a shared base and a different depth for each role.

  • Everyone who uses AI

    What generative AI does and does not do, how to check an output, which data not to enter, which tools are authorised and whom to report a problem to. It reduces the use of personal tools without rules.

    Shared base, before the other levels

  • Managers and department heads

    Choosing the processes where AI helps, setting the starting measurement, assessing projects and vendors, deciding usage rules and responsibilities. No programming needed.

    Course: AI Executive Mastery

  • Process owners and operational users

    Describing the work to improve (inputs, outputs, exceptions), defining when an output is acceptable, building reusable prompts and templates and knowing when to hand the case to a person.

    Course: Prompt Engineering for Business Workflows

  • Technical leads

    Connecting AI to the systems in use: where the data comes from, which permissions are needed, where the outcome is recorded, how logs and exceptions are handled.

    Course: AI Agent Developer

Training becomes operational capability when, on the same use case, the team connects four skills: process (describing the work), quality (criteria and review thresholds), integration (moving data between tools) and control (owner, logs, exceptions and autonomy in proportion to the impact).

The first use case and how to measure adoption

The course starts from a concrete unit of work and ends with a trial in the workflow, not with a test on the tool.

Eight questions to choose the case

  1. Is the problem frequent and narrow enough?
  2. Is there an owner who can change the process?
  3. Can inputs and outputs be observed and compared?
  4. Is the data needed available and authorised?
  5. Can quality be turned into verifiable criteria?
  6. Are errors reversible or controllable during the trial?
  7. Can the starting point be measured before you begin?
  8. Does the team have the time and responsibility to apply what it learns?

If the owner, the data or the criteria are missing, the case is not ready: sometimes it is better to prepare the process before adding technology.

What to measure after the course

  • Use in the workflow: in how many eligible cases the method is actually applied.
  • Accepted outputs: how many meet the criteria without substantial rework.
  • Cycle time: the change in total time, including waits, reviews and exceptions.
  • Corrections: how many AI proposals are corrected or rejected, and why.
  • Exceptions: how many non-standard cases are recognised and passed to the right person.

What counts is what the team does at 90 days, not how much it liked the course the day after. The number of prompts written measures activity, not capability.

Three training formats

Choose the format that suits your team and the way your organisation works.

A trainer presents an AI dashboard to a group of professionals in a training room overlooking Milan Cathedral

In-person training

At your premises or at our office in Milan.

  • Working directly with the trainers
  • Exercises on the team's workflow
  • Discussion among colleagues from the same office
  • Training materials included

Live online training

Interactive virtual sessions, with work in small groups.

  • Times agreed with the team
  • Participation from several sites
  • No travel
  • Shared materials and templates

Hybrid training

In-person and online sessions, combined around the team's calendar.

  • In-person kickoff, online exercises
  • Calendar adapted to shifts
  • Same course for different sites
  • Final review of the work done

The cost is set in a written quote after the needs analysis; the first 30-minute call is free.

Frequently asked questions

Answers to the most common questions about our AI training courses.

What are the prerequisites for the courses?

They depend on the course:

  • AI Executive Mastery: no specific prerequisites, designed for managers and department heads
  • Prompt Engineering: basic experience with AI tools and with content or code work
  • AI Agent Developer: programming experience in Python or JavaScript and knowledge of APIs

How much does an AI training course cost?

The first 30-minute call is free. The cost depends on the number of participants, duration, format and customisation, and is set in a written quote after the needs analysis.

Is a certificate of attendance issued?

Yes, at the end of the course participants receive a certificate of attendance. It documents the course completed; it is not a professional qualification.

Can the courses be customised for my company?

Yes. Before the course we analyse the team's needs with you and choose a real workflow to work on:

  • preliminary analysis of needs and participants;
  • examples and exercises on your industry and your company's tools;
  • expected output and quality criteria agreed before starting;
  • an agreed check on adoption after the course.

How long does it take to see concrete results?

Results depend on the process chosen, the data available and the team's adoption. That is why we fix a starting measurement before the course and then check:

  • completion of the agreed task;
  • output quality against the agreed criteria;
  • time, errors or rework compared with the starting measurement;
  • the procedure, checks and exceptions to reuse.

Is a prompt engineering course enough to adopt AI?

No. It improves individual use, but adoption in a company also needs a described process, accessible data, quality criteria, integration with your tools, responsibilities and shared checks. That is why the courses start from one of the team's real workflows.

How do you choose an AI training course for your company?

With five questions to ask any provider.

  • Does the content adapt to your industry and your use cases?
  • Have the trainers brought AI into real work, or only studied it?
  • Do you work on real or authorised data, or only on slides?
  • Is there a follow-up after the course?
  • How is the result measured, before and after?

Is AI training mandatory under the AI Act?

The AI Act does not require a specific course. Article 4, applicable since 2 February 2025, asks those who develop or use AI systems to take measures to support the AI skills of the staff who use them; with Regulation (EU) 2026/1744, in force since 27 July 2026, it is an obligation of means, not of result. The AI Act sets no specific penalty for Article 4. It is still worth documenting the measures you take; for your own case you need a legal check.

Tell us about your team's workflow

In 30 minutes, free and without obligation, we work out which task to bring to the course, who takes part and which result to check at the end.

What the first call includes

  • A free 30-minute call
  • Training needs analysis
  • A written quote after the analysis
Book the first call

We reply within 24 working hours to understand the workflow, the participants and the goal of the course.