How do you get staff using AI well, without leaking data?
In a 2024 survey of 31,000 people by Microsoft and LinkedIn, 75% of knowledge workers used AI at work, and 78% of those users brought their own AI tools.
The risks sit in everyday tasks. People paste in client files, accept answers that sound right, and send text nobody checked. The result can be leaked data on one side and low-quality output on the other.
This course gives every participant one working method. Understand what the tool does, brief it properly, decide what data it may see, and check what comes back. By the end of the day each person has used the method on their own tasks, and the team has a prompt library it can use on Monday.
After this course
- Can you explain in plain words how a language model produces an answer and why it can be confidently wrong?
- Can you write prompts that state the task, the context, the format and the length?
- Can you decide which data may go into which AI tool, using your organisation's rules?
- Can you check AI output against an independent source before it leaves your hands?
- Can you spot hidden instructions in documents and requests that try to steer an AI tool?
- Can you use a tested prompt library for 10 common tasks in your own role?
Skills you practise
- Prompt writing
- Output checking
- Data classification
- Safe tool use
- Context management
- Deepfake awareness
Course structure
The course is 1 day of 6 teaching hours, in 2 blocks of 3 hours with a lunch break. Online, it runs as 2 live sessions of 3 hours on the same day, not recorded.
The first block covers how the tools work and where they fail. Participants see the same question produce a fluent wrong answer and a correct one, then learn why. They ask for sources and open them. They then rewrite 3 prompts from their own work and compare the results side by side.
The second block covers data and checking. Each participant labels realistic text snippets with the organisation’s data rules, then practises checking AI output against a second source on a task from their role. The last hour goes to hidden instructions in documents, fake voices on calls, and finishing the team prompt library.
About two thirds of the time is hands-on. Every exercise uses the tool your organisation has approved, so nothing has to be translated back to work afterwards.
Every exercise follows the same method: understand what the tool does, brief it properly, decide what data it may see, and check what comes back.
1How a language model produces an answer60 min
- Next-word prediction
- Training data and knowledge cutoff
- Why answers sound certain
- What the tool can look up and what it cannot
A language model writes by predicting the next piece of text, one piece at a time, from patterns in its training data. It knows nothing after its training date unless it can search.
Caution. A fluent answer is not evidence that it is current or correct.
You leave with: A one-line explanation of how the tool works, written in the participant's own words.
2Where AI is strong and where it fails60 min
- Strong and weak answers from one tool
- Made-up facts and citations
- Agreeing with the user
- Tasks to keep away from AI
A model can draft a difficult letter well and then invent a court case in its next answer. Judge each output on its own, never on the last good one.
Caution. In June 2023 a US court fined 2 lawyers for filing 6 cases that ChatGPT had made up; every citation needs a check against the original.
You leave with: A personal list of 5 tasks in the participant's role where AI helps and 3 where it should not be used.
3Prompts that state the job60 min
- Scope: outputs, order, limits
- Length and format caps
- Instructions phrased as what to do
- Action verbs and a stated goal
Current models follow instructions literally. The text you type is the whole brief, so what you leave out gets guessed.
Caution. Remove names, account numbers and other personal data from anything you paste unless the tool is approved for it.
You leave with: 3 rewritten prompts from the participant's own work, tested in the session.
4Context and memory60 min
- What the model sees in one chat
- One topic per chat
- Putting key facts first
- Long documents and what gets missed
Apart from its training, a model works from what is in the current chat and any memory the tool says it keeps. A long chat that mixes topics tends to get less accurate answers.
Caution. Research shows models miss information placed in the middle of long inputs; check that a long document was actually read.
You leave with: A short checklist for starting a new chat on a new task.
5What may go into which tool60 min
- Shadow AI: personal accounts used for work
- Personal and special-category data
- Confidential and inside information
- Your organisation's approved tools and rules
Treat the prompt box as a message sent outside the company, unless your organisation has a contract with the provider that says otherwise.
Caution. The Dutch data protection authority has received breach reports after staff entered patient data and customer addresses into chatbots; such uploads can be reportable data breaches.
You leave with: A one-page AI use card: approved tools, the data colour each may take, and who to ask.
6Checking the work, and hidden instructions60 min
- Checking every answer before you use it
- Never let the model grade its own answer
- Hidden instructions in files, emails and web pages
- Fake voices and video calls
Every AI answer you use becomes your answer, so check it before it goes anywhere. Instructions hidden in a file, email or web page the model reads can change what it does.
Caution. In 2024 an Arup employee in Hong Kong transferred HK$200 million after a video call with deepfaked colleagues; confirm unusual payment requests through a known, separate channel.
You leave with: The finished team prompt library: 10 prompts, each tested and checked in the session.
How a language model produces an answer
Ask for the date of the model's knowledge before asking about recent events.
What are the latest rules on AI in Europe?
What is the most recent date your training data covers? Then list what you know about EU AI rules up to that date, mark anything that may have changed since, and tell me which official source I should check.
Where AI is strong and where it fails
Ask for sources you can open, then open them.
Give me 3 court cases that support my argument.
List 3 court cases on this point with full citation and a link to the official court or database page for each. If you cannot give a working link, say 'not found' for that case.
Prompts that state the job
Name every output, the order, and the limits.
Can you help me with this email to the supplier?
Write a reply to the supplier email below. Goal: move the delivery date to 14 November without changing the price. Tone: polite and firm. Length: under 120 words. End with one question that asks for written confirmation by Friday.
Context and memory
Give all requirements in the first prompt, key facts at the top.
Summarise this report. [pastes 40 pages] Also make it shorter. Actually focus on costs.
Summarise the cost findings in the report below for our finance director. Format: 5 bullets, each under 20 words, then 1 line on the biggest open risk. Use only figures from the report and give the page number for each. [report]
What may go into which tool
Classify the data before you paste it.
Here is the client file, please summarise the problems.
Using our data rules (public: any tool; internal: the approved tool only; personal: no AI tool), label each of these 10 text snippets public, internal or personal and explain the label in 1 line.
Checking the work, and hidden instructions
Check against a second source, never the same chat.
Are you sure this is correct?
List each factual claim in your answer as a numbered line with the source you used. I will check each one against the original document.
The team leaves with 2 documents. The first is a prompt library of 10 tested prompts for the tasks the team does most, each with the check that goes with it. The second is a one-page AI use card: the approved tools, which data each may take, and who to ask when in doubt.
Teams use the card in onboarding. The prompt library lives wherever the team keeps shared documents, and the person who owns it adds new prompts as they are tested. Each participant also receives an attendance record for your AI literacy file.
- Teaching time
- 6 hours on 1 day, in 2 blocks of 3 hours with a lunch break.
- Before the day
- None. Participants need access to the AI tool their organisation has approved, or a practice account we arrange.
- After the day
- A team prompt library of 10 tested prompts and a one-page AI use card listing approved tools and the data rule for each.
Completion
You complete the course by working through each exercise on your own tasks. There is no exam and no grade. Each participant receives an attendance record for your AI literacy file (not a certificate).
Who is this course for?
Anyone who uses, or is about to use, an AI assistant at work: office staff, team leads, specialists, new joiners. No technical background is needed. Participants who have never opened an AI tool and participants who use one daily both fit, because the exercises run on each person’s own tasks.
It works best when a whole team attends together, because the prompt library and the data rules are shared. A mixed group from several departments also works when the organisation is rolling out one approved tool to everyone.
Choose AI for Daily Work: Prompts, Assistants and Automations if your team already uses AI every day and wants to build assistants and simple automations. Choose Shadow AI and Safe AI Use if your main concern is staff using unapproved tools.
In the Staff path, this course is step 1 of 4. Next: AI Literacy and the AI Act.
- Level
- Beginner
- Before you start
- None. Participants need access to the AI tool their organisation has approved, or a practice account we arrange.
- Group
- 1 to 15 people, online, live or at your office
This course is right for you if you
- use, or are about to use, an AI assistant at work
- want one method to brief the tool and check what comes back
- handle client, customer or company data in your daily work
Unsure whether it fits? Ask us. We reply within 1 working day.
Practical information
- 1 day, 6 teaching hours, with a lunch break
- Online: live in 2 sessions of 3 hours on one day, not recorded; times agreed in the intake call
- At your office: 09:30 to 16:30; you provide a room, a screen and internet access
- In English, in the AI tool your organisation has approved
- Taught live by a practitioner with more than 15 years of commercial and technical leadership in software and data companies across Europe and Asia-Pacific, including maritime, energy and fintech.
- Attendance record for your AI literacy file
- Online, live: 1 day in 2 live sessions of 3 hours, not recorded. Book it as a private day for your team of 1 to 15 people, on your own Teams, Zoom or Meet.
- At your office: 1 day (6 teaching hours) for 1 to 15 people, travel within the Netherlands included.
Dates and booking
Choose the format, an open weekday in the calendar and the number of participants, then reserve on this page.
The calendar and booking form load with JavaScript. Without it, ask us for a date on WhatsApp, by phone on +31 6 20 51 73 96 or at advisory@ai-business.school.
Investment
| Item | Online, live | At your office |
|---|---|---|
| Price | €399 per person | €2,900 for the day, for your group |
| Including 21% Dutch VAT | €482.79 per person | €3,509 |
| The training day | Included | Included |
| A 30-minute intake call on your tools and data rules | Included | Included |
| Course materials and templates | Included | Included |
| Attendance record for your AI literacy file | Included | Included |
| Travel within the Netherlands | Not needed | Included |
| Group | 1 to 15 people | 1 to 15 people |
Prices exclude VAT. Online, live, you pay per participant. At your office, one price covers your whole group, from 1 to 15 people; for a larger group, book a second day. You reserve on this page; the payment details appear when you do.
Questions
Do my staff need technical knowledge for an AI basics course?
No. The course starts from how the tools behave in daily work. Participants need a laptop and access to the AI tool your organisation approved, or a practice account we set up.
Does this course meet the EU AI Act AI literacy requirement?
It supports your Article 4 AI literacy duty. The AI Act does not require a certificate; you keep internal records, and each participant gets an attendance record for your AI literacy file.
Which AI tools do you use in the training?
The tool your organisation has approved, such as Microsoft Copilot, ChatGPT Enterprise, Gemini or Claude. If you have none yet, we use a practice account and show how the rules transfer.
Can we run AI basics training online for a team in several countries?
Yes. The online version runs live in 2 sessions of 3 hours on one day, in English, and is not recorded. Book it as a private day for your team, from 1 person, on your own Teams, Zoom or Meet.
Why not use a free online AI course instead?
A free course explains AI in general. This day trains your team in the tool your organisation has approved, with your data rules, on their own tasks. It ends with a tested prompt library and an attendance record for your AI literacy file. More on free courses
What is the difference between this course and AI for Daily Work?
AI Basics at Work gives everyone one method, whether they are new to AI or already use it daily. AI for Daily Work starts from regular AI use or this course, and adds assistants and simple automations.
Last reviewed 21 October 2026.
Related courses
AI Literacy and the AI Act
Explain what the EU AI Act asks of your organisation and keep the records for your AI literacy file.
€399 per person online · €2,900 a day at your office, up to 15 people
Shadow AI and Safe AI Use
Use AI at work without leaking data or falling for deepfake scams, and leave with the input for your AI policy.
€399 per person online · €2,900 a day at your office, up to 15 people
AI for Daily Work: Prompts, Assistants and Automations
Build a tested assistant or simple automation for a recurring task, with a check step and an owner.
€399 per person online · €2,900 a day at your office, up to 15 people