SERVICES — ARTIFICIAL INTELLIGENCE AXIS
Operational AI in the company, from the team to annual oversight
Adopting AI in a company means three things: training people so they actually use the tools, understanding AI Act exposure before it becomes a problem, and overseeing the regulatory obligations that don't end with a single project. This page covers the full path.
THE METHOD
Learn · Practice · Build
Each module lasts 4 hours, in person or remote, up to 12 participants. The principle is one: every participant leaves with a working deliverable built during the course on their own real case — not on generic examples.
TRAINING — L1
A common foundation, then the platform you use
AI training is built on three levels: one mandatory base module for everyone, one or more deep-dive modules on the platform in use at the company, and applied modules by function. Each level adds operational competence — not theory to forget.
LEVEL 1 — BASE · AI-01
Introduction to Artificial Intelligence
Target: Mixed teams, management, any company function. No technical prerequisites.
- →How generative AI works: essential concepts without the math
- →Overview of available tools — Copilot, ChatGPT, Claude, Gemini
- →Concrete use cases by company function: operations, HR, finance, sales, quality
- →Limits and risks: hallucinations, sensitive data, AI Act
- →Fundamental prompt engineering: how to ask the right questions
Workshop — what you build: A personal prompt library (5–10 prompts) calibrated to your role and daily activities. Tested and working by the end of the session.
LEVEL 2 — PLATFORM · ONE TO CHOOSE
After the base, you deep-dive into the tool the company uses or intends to adopt. Each track is designed for that specific platform — the advanced features, the high-impact use cases, the operational limits.
PLATFORM TRACK — MICROSOFT COPILOT
For whom: Organizations with active Microsoft 365 Copilot licenses.
Copilot in Office (Word, Outlook, Teams, Excel, PowerPoint)
How to use Copilot in the main M365 applications to cut time on repetitive tasks. Each application has its highest-impact use cases.
Workshop: each participant builds a personal Copilot workflow that spans at least two applications — for example analyzing data in Excel, turning it into PowerPoint slides, communicating it via Outlook.
Power Automate + Copilot
How to design business-flow automations with Power Automate, using Copilot to speed up flow creation. For those who want to go beyond individual productivity.
Workshop: each participant (or group) builds a working automation flow on a real process — an automated approval, a Teams notification, a data collection.
Copilot Studio: Custom Agents and Workflows
For those who want to build conversational agents connected to company documents, processes or systems. Use cases: internal FAQs, onboarding, operational support.
Workshop: each participant (or group) builds a working conversational agent connected to a real company document or process.
PLATFORM TRACK — CLAUDE
For whom: Organizations adopting Claude as their main AI tool for professional work.
Fundamentals: Working with Claude
Anatomy of an effective prompt, Projects to organize contexts and reuse instructions, file and document handling. Focus on output quality and reducing iterations.
Workshop: each participant builds a Claude Project with a structured system prompt for their role — context, tone and format instructions calibrated to daily work, with a set of pre-tested reusable prompts.
Advanced: Complex Cases and Automations
Long-document analysis, Artifacts for structured reusable outputs, Cowork for multi-step tasks on local files, wireframe generation, code review. For those already familiar with Claude.
Workshop: each participant builds an advanced workflow tied to a real use case — document processing, structured-output automation, an assistant with complex context handling.
PLATFORM TRACK — CHATGPT
For whom: Organizations adopting ChatGPT as their main AI tool.
Fundamentals: Working with ChatGPT
Anatomy of an effective prompt with GPT-4o, memory and context management, PDF and Excel file analysis, best practices for output quality and verification.
Workshop: each participant builds a library of reusable prompts — at least 3 structured prompts, tested on real cases, ready to use.
Advanced: Custom GPTs and Integrations
Building no-code custom assistants with a knowledge base and actions. ChatGPT with web search. Advanced data analysis with the built-in Python interpreter. Canvas for collaborative editing.
Workshop: each participant builds a working Custom GPT for a real business use case — document analysis, FAQs on internal procedures, support for drafting reports.
CROSS-CUTTING TRACK — COMPARISON
For whom: Organizations that need to choose which platform to adopt before investing in licenses.
AI platform comparison
Comparative overview of the main providers (Copilot, Claude, ChatGPT, Gemini), selection criteria by use case, integration with existing systems, data security, GDPR and AI Act implications. How to build a minimal company AI policy.
Workshop: the group builds an evaluation scorecard of the relevant AI tools — with weighted criteria and a preliminary recommendation to bring to management.
LEVEL 3 — APPLIED BY FUNCTION
The applied modules bring AI into the specific activities of each business area. Content and workshop are calibrated to the department's real use cases during the pre-course briefing.
- Operations and Production — AI for technical documentation, KPI analysis, non-conformity management, production data. Workshop: an operating procedure or analysis template built on a real department dataset.
- HR and People Management — Recruiting, onboarding, internal training, communication. On request.
- Quality and Compliance — Documentation, audit, regulation, complaint management. On request.
- Finance and Administration — Reporting, data analysis, document automation. On request.
- Sales and Marketing — Content, CRM, proposal support. On request.
The applied modules (except Operations) are developed on request with a dedicated pre-course briefing to gather the department's specific use cases.
PATHS
Three starting combinations, customizable
The pre-packaged paths cover the most frequent use cases. Every path is adjustable: the pre-course briefing exists precisely to calibrate it to the company's context.
Starter
Ideal for: First approach, small teams, companies that want to test before scaling.
Composition: AI-01 + 1 platform module of choice.
What you take home: A solid understanding of AI and operational skills on the platform in use. At least one working deliverable built during the course.
Practitioner
Ideal for: Functional teams, structured training on a specific tool.
Composition: AI-01 + 2 platform modules + 1 applied module.
What you take home: Operational command of the platform and practical tools applied to your business function.
Full
Ideal for: Teams that want a complete path, companies in structured adoption.
Composition: AI-01 + 3 platform modules + 2 applied modules.
What you take home: Complete skills — conceptual, operational and applied — and a set of working deliverables for each department involved.
Custom — from 2 modules. For those with specific needs or who want to combine modules freely.
RESULTS
What you find in the company the next day
Every participant has at least one working deliverable built during the course — prompt library, Copilot workflow, Custom GPT, Claude agent — immediately applicable to their work.
On targeted documentary and repetitive tasks, the observed time reduction is between 20% and 40%. With one less risk: the unwitting use of AI on sensitive data.
Management can read the regulatory risks of AI adoption — AI Act, liability, audit trail — without depending on vendor slides.
The company has an internal base of competence to evaluate subsequent AI investments: governance, tools, automations.
ASSESSMENT — L2
Capturing AI Act exposure before it becomes a problem
The AI tools already in use — Copilot, automations, decision systems — fall within the AI Act's scope sooner than many companies expect. The AI assessment exists to understand where the exposure is, which obligations apply and what to do in the next 90 days — before finding out during an inspection.
AI Radar
AI TRACK · 1 DAY · DECLARED OUTPUT
AI tools in use, processes involved, data handled, vendors. Interviews with the CEO, IT contact and 1–2 process owners.
- →AI Act exposure map: risk level, applicable obligations, deadlines
- →The 3–5 priority gaps to address
- →First concrete actions for the next 90 days
- →A concise document (4–6 pages) to share internally
AI Compass
AI TRACK · 2 DAYS · 15–20 PAGE REPORT · ROADMAP
Extended interviews with the CEO, CTO, IT, HR, operations; processes, data, systems and AI vendors; existing governance.
- →Complete AI assessment report (15–20 pages)
- →Classification of AI systems by AI Act risk level
- →Gap analysis against applicable obligations
- →Compliance and adoption roadmap with milestones and indicative costs
- →Executive slides for the board or investors, on request
The Radar doesn't oblige you to the Compass. From the Compass, the recurring services open up with no additional onboarding cost — knowledge of the company is already acquired.
ANNUAL OVERSIGHT — EXTERNAL AI OFFICER
The AI Act requires an accountable figure. It can be outsourced.
ANNUAL RENEWAL · FORMAL LETTER OF ENGAGEMENT
External AI Officer
TRIGGER → AI SYSTEMS CLASSIFIABLE AS HIGH-RISK UNDER THE AI ACT
The AI Act requires organizations using high-risk AI systems to have a figure accountable for compliance — with technical, regulatory and governance skills. Few SMEs have one in-house. The role can be outsourced and formalized with a letter of engagement.
- →Formal designation as the organization's external AI Officer
- →Maintenance and update of the AI systems register
- →Oversight of transparency, documentation and AI risk assessment obligations
- →Management of ongoing compliance with applicable AI Act requirements
- →Continuous updates on supervisory authority guidelines and interpretations
- →Interface with the competent authorities in case of inspection or request for information
- →Monthly session with the CEO/CTO (45 minutes) for updates and decisions
- →Annual training of the internal team on AI Act requirements (1 four-hour session included)
ISO/IEC 42001 LEAD IMPLEMENTER (PECB) · MEMBER UNI/CT 533 — ARTIFICIAL INTELLIGENCE
Those who have already completed an AI Compass access the service with no additional onboarding cost. Those arriving without an assessment start with a half-day of initial alignment.
AI GOVERNANCE — ISO 42001
Structuring a formal AI Management System
ISO 42001 is the international standard for artificial intelligence management systems. For those who have already started AI adoption and want structured governance — compliant with the AI Act and demonstrable to customers, partners and authorities — it's the natural next step after training or assessment.
ISO42-01 · 1 MODULE · 4 HOURS
Introduction to ISO 42001
Structure of the standard, relationship with ISO 9001 and ISO 27001, application to those who develop or use AI systems, relationship with the AI Act. Sellable on its own as an AI governance awareness session.
Workshop: each participant builds an AI organizational context map — identifying existing or planned AI uses, main risks, governance areas to oversee.
ISO42-02 · 1 MODULE · 4 HOURS
Implementing an AI Management System
How to concretely build an AIMS in a manufacturing SME: gap analysis, AI policy, risk management, mandatory documentation, the path toward certification.
Workshop: each participant starts an ISO 42001 gap analysis — mapping requirements against the current situation, with priorities and a 90-day action plan.
Complete ISO 42001 path: ISO42-01 + ISO42-02 · 1 day · for quality managers, innovation managers, CTOs, CISOs.
The ISO 42001 modules are taught by Alberto Scarpa, certified ISO/IEC 42001 Lead Implementer (PECB), member of the UNI CT 533 Technical Committee — Artificial Intelligence.
Where do you want to start?
If you're not sure which module or path is right for your situation, the Regulatory Spark is exactly for this: 45 minutes to map your exposure and choose the next step on concrete ground. If you already have a clear idea, write to me directly for the detailed program.
Book the Regulatory Spark — free