One prompt can leak sensitive data
A contract, client data or an internal document can end up in a public AI service before the user realises the consequences.
[ Using artificial intelligence safely ]
AI speeds up work. It can send company data to the wrong place just as quickly.
Artificial intelligence is already an everyday work tool, and a new security risk. This course explains how AI really works, what you can share with it, how to verify its output and how to set rules that protect data, employees and your organisation.
[ AI in everyday work ]
You're probably already using AI at work. The question is what people are putting into it.
Without shared rules, each employee decides which tool to use, what data to share and whether its output can be trusted.
Convincing output isn't always correct
Hallucinations, bias and missing context can produce confident answers that lead to poor decisions.
Without shared rules, everyone draws their own line
The difference between safe and risky use cannot be left to individual intuition alone.
Attackers use the same tools
AI accelerates phishing, social engineering, deepfakes and the creation of convincing scams.
[ Training outcomes ]
Participants will know when to use AI, what not to share and how to verify its output.
01
Choose the right tool
Click to learn more →01
Distinguish between public and enterprise AI services and understand why their terms and data handling differ.
← Back02
Protect company information
Click to learn more →02
Learn to classify and anonymise data and decide what should never go into an AI tool.
← Back03
Verify output quality
Click to learn more →03
Look beyond confident wording and use checks proportionate to the consequences of the result.
← Back04
Recognise AI-assisted attacks
Click to learn more →04
Understand deepfakes, voice cloning, prompt injection and how AI changes phishing and fraud.
← Back[ Audience ]
For anyone who uses AI, approves its adoption or sets the rules
The course connects practical AI use with security, accountability and company policy. No technical background is required.
[ Suitable for ]
- Employees using AI tools
- Management and team leaders
- Marketing, HR, sales and customer support
- Legal, compliance and risk teams
- IT and cybersecurity teams
- Internal AI policy authors
[ Course content ]
From your first prompt to company policy and defence against attacks
We adjust the emphasis to your tools, participants' roles, types of data and whether your organisation already has an AI policy.
01What AI is and how it works
- Generative models, neural networks and tokens
- Hallucinations, bias and AI limitations
- Tools such as ChatGPT, Copilot, Gemini and Midjourney
02Using AI effectively
- Writing good prompts and practising prompt hygiene
- Practical workplace scenarios
- Best practices for safe use
03Risks of using AI
- Data leaks and inappropriate inputs
- Public versus enterprise AI
- Anonymising and classifying information
04Company AI policy
- Defining rules and roles
- Data classification and access
- Prohibited practices and recommended tools
05Legal framework
- Copyright and the use of AI output
- The AI Act, risk categories and organisational obligations
- Local models and on-premises AI
06Current threats
- Jailbreaks, prompt injection and poisoning
- Attackers and their AI tools
- AI in phishing, social engineering and ransomware 2.0
07Deepfakes and voice cloning
- How to spot a deepfake
- Real-world examples
- Protecting reputation and preventing misuse
[ Training format ]
From your first prompt to rules the whole company can use
Explanations are supported by practical scenarios, examples of safe and risky inputs, and current attacks using generative AI.
01
Technical foundations without jargon
Click to learn more →01
Participants learn AI principles and limitations without needing programming or data science knowledge.
← Back02
Practical workplace scenarios
Click to learn more →02
We show where AI genuinely helps and where its use creates disproportionate risk.
← Back03
Safe and risky prompts
Click to learn more →03
Concrete examples show effective prompts, inappropriate inputs and ways to anonymise data.
← Back04
Threats and defence in practice
Click to learn more →04
We connect deepfakes, prompt injection and AI phishing to measures an organisation can actually implement.
← Back[ Practical information ]
A practical format tailored to your organisation's tools and rules
We adapt the content to your AI services, participants' roles, internal data classification and the maturity of your AI policy.
- Format
- In person
- Duration
- 3–4 hours
- Number of participants
- 5–30 people
- Prerequisites
- None
- Price
- From CZK 45,000
- Customisation
- Based on the organisation's AI tools, roles, data types and internal rules
- Wrap-up
- Practical recommendations, discussion and Q&A
[ Your trainer ]
Pavel „SPAJK“ Matějíček
Pavel “Spajk” Matějíček is a cybersecurity trainer and consultant at BOIT. He has worked in cybersecurity for over 10 years, focusing on user security awareness, social engineering prevention and practical protection of company data in everyday operations. Previously, he worked at ESET, helping educate people about cyber threats.
In his training, he explains things clearly, without unnecessary technical jargon, using real-world situations. He shows participants how to use technology effectively, where risks arise and how to put defences into practice at work.
[ Why BOIT ]
We explain AI in the context of data, people and real attacks
We combine productivity and security
We don't ban a useful tool. We show how to benefit from it without unnecessary risk.
We use real-world situations
We translate general principles into the prompts, data and decisions employees encounter.
Rules must be usable
An AI policy only works when people understand its boundaries and can apply them in everyday work.
We tailor the content
We account for your tools, data types, internal rules and participants' specific roles.
[ Client experiences ]
BOIT training combines expertise with practical, clear delivery
Pavel's cybersecurity training has the pace of an ice hockey game: fast, practical and full of involvement. Participants aren't spectators: they're active players. No boredom, just clear insights they can use straight away.
Accolade has worked with Pavel for several years. His proactive approach makes his training one of our most popular and best-attended sessions, engaging people across generations.
We've worked with BOIT on information security talks for a long time. Pavel combines excellent expertise with an entertaining style. In over 10 years of organising specialist events, BOIT has been one of our best partners. Highly recommended.
[ FAQ ]
What you need to know before booking
01Is the training only for technical teams?
No. It is designed for employees, management and teams that set company policies. We explain the technical principles without unnecessary jargon.
02Do participants need experience with AI?
No. We start with how AI works, then move on to practical use, risks and protection.
03Does the training cover specific AI tools?
Yes. We explain the differences between public and enterprise services and tailor the examples to the tools used in your organisation.
04Will the training help us with our AI policy?
The training explains the principles, roles and boundaries of an effective AI policy. It does not replace legal services or the separate preparation of internal documentation.
05Does the training cover the AI Act and copyright?
Yes. Participants gain a practical understanding of the relevant legal framework and its impact on AI use within the organisation.
06Is the training available online?
We offer this training in person. The face-to-face format allows for live demonstrations, group scenarios and discussion.
07How long does the training last and how much does it cost?
The training lasts 3 to 4 hours, with prices starting at CZK 45,000. The final quote depends on the scope and level of customisation.