What AI can and cannot do
Strengths and limits of generative models, and the difference between a chat, an assistant that knows your context and an agent that works in several steps.
AI training for healthcare staff
Workshops and training to build and refresh AI skills for the people who work in a healthcare facility. We start from the processes of your offices and departments, with clear rules on data and privacy.
AI is already in your facility, but at what risk?
In many healthcare facilities generative AI has already come in through the back door: some people use it with a personal account, some do not trust it, and hardly anyone has been told what is allowed and with which data. Meanwhile a large share of staff work is still stitching things together across Excel, email and systems that do not talk to each other.
Training brings order: where AI saves time, where it needs supervision and where a person has to decide. It is also a concrete AI literacy measure, which the AI Act (Article 4) asks of organisations whose staff use these systems.
-
Repetitive tasks that eat up hours every week
Minutes, email drafts, reformatted procedures, comparison tables: tasks where a well-generated draft can be reviewed in a few minutes. -
Health data and personal accounts do not mix
Without a shared rule, patient or staff data can end up in tools the facility never approved. -
Plausible answers that still need checking
AI proposes, and it can be confidently wrong: numbers, regulations and quotations must be checked by whoever signs the document.
What your staff learns
Six blocks, from context to practice. Content is adapted to the functions in the room and to the tools the facility already has.
The tools compared
Microsoft Copilot, ChatGPT and Claude: what they see of company data, when each one fits, and why work is done only with company accounts.
The rules of the game
GDPR, the AI Act and Italian Law 132/2025 in plain words: when to disclose the use of AI, and what changes for healthcare, employment and staff management.
The data traffic light
A practical green, amber and red classification: which information can be used with AI, which must be pseudonymised and which always stays out of chats.
Writing a good prompt
The six-building-block method (role, context, task, format, constraints, examples), comparing a poor request with one that produces a usable draft.
Group lab
Participants work on realistic cases with dummy data, improve the prompt version after version and present it to the room against a scoring grid.
Use cases for every function in the facility
The examples used in the room come from the daily work of the people attending. For each area we also point out what to watch.
Management and Quality
Minutes with decisions and owners, briefings before meetings, comparison between internal procedures and accreditation requirements.
Nursing coordination and Operations
Constraint checks on a draft rota using staff IDs instead of names, work instructions and checklists. AI proposes, the coordinator decides.
Human resources
Clear and inclusive job adverts, onboarding kits, interview outlines. No automated CV screening.
Administration, Controlling and Procurement
Commentary on budget variances, formulas explained step by step, comparison tables for supplier offers, contract summaries.
Fundraising and Communication
From the call for proposals to a draft application, one piece of news adapted to several channels, tone of voice and person-first language as standing instructions.
Reception and front office
Answers to recurring requests, rewritten patient communications, internal FAQs. Always without clinical data in general-purpose chats.
Two formats, depending on how you work
Both formats are delivered on site at the facility and include a focus on data security and privacy.
Cross-functional workshop
A single session where the different staff functions work together, starting from the repetitive tasks of every day.
- Length
- half a day, about 4 hours
- Participants
- up to 25-30 people, in mixed teams
- Best for
- getting separate offices to talk and surfacing needs from the ground up
Departmental masterclasses
Short, focused sessions built on the processes and documents of a single function, with hands-on practice at the computer.
- Length
- 2-hour modules per function
- Participants
- small groups of 5-8 people
- Best for
- leaving the room with prompts ready for your own work
How a workshop runs
The typical outline of a half-day session. Timing and cases are agreed with the facility beforehand.
-
01
Context and warm-up
Each participant names the repetitive task that takes most of their time. Then an overview of what AI can do today and how work changes: less drafting, more preparation and review.
-
02
Tools and rules
A comparison of the available tools and the essential regulatory picture: GDPR, the AI Act, Italian Law 132/2025 and the data traffic light.
-
03
Prompts and live demos
The method for writing effective requests and a few demonstrations on concrete cases: from meeting to minutes, from offers to a comparison table, from research to a one-page brief.
-
04
Lab and next steps
Group work on realistic cases, presentation to the room, practical tips and mistakes to avoid. It closes with a method to keep experimenting independently.
What the facility keeps
The Prompt Book
A digital collection of tested prompts, organised by function, to reuse and extend after the course.
The course slides
The full material as a PDF, with the data traffic light and the regulatory overview to keep at hand.
Who has already taken part
In October 2026 we ran an AI workshop for the central offices and management of Centro Clinico NeMO in Milan. This is the review published on Trustpilot after the session. All the others are on the reviews page.
Published on Trustpilot
Competence, confidence and empathy
RUN2AI is a young, dynamic and highly competent company. Through its CEO Andrea Giannuzzi, who had already been a valuable support in recent months on Fondazione SERENA's Digital Transformation journey, we were able to explore the effects and opportunities of AI in depth, relying on the CEO's competence, confidence and professionalism. An experience worth repeating!
Frequently asked questions
Audience, prerequisites, tools and how data is handled during the training.
Staff of healthcare facilities: management, administrative offices, quality, human resources, nursing coordination, communication, fundraising, procurement and front office.
The course covers the use of AI in internal and organisational processes. It is not a course on AI for diagnosis or clinical decisions, which always remain with the healthcare professional.
No. We start from the basics and from the daily work of the participants: being comfortable with the usual office tools is enough.
People who already use AI still gain a method to write better requests, check results and share what works with colleagues.
The tools the facility has or plans to adopt. We usually compare Microsoft Copilot, ChatGPT and Claude, explaining what changes between free and business versions.
The working method, from prompts to checking results, applies to any tool.
No. Exercises use dummy data or neutralised documents, and part of the course is dedicated to understanding which data can be used with AI.
The data traffic light separates information that can be used with company tools, information to pseudonymise and information, such as health data, that stays out of general-purpose chats.
Yes: Article 4 of the AI Act asks organisations that use AI systems to ensure an adequate level of AI literacy among their staff, and a documented training course is one of the measures to do so.
Training does not replace the assessments of the facility's DPO or legal adviser, with whom the internal policy and obligations should be defined.
The cross-functional workshop lasts half a day, about four hours; departmental masterclasses are two-hour modules per function. Both are delivered on site at the facility.
Content, use cases and schedule are agreed together before the session. Sessions are currently delivered in Italy.
Bring AI training to your facility
Tell us how many people you want to involve and from which functions: we will suggest the most suitable format and send a quote.
See all Run2AI solutions → Browse the healthcare AI glossary →