How to Choose AI Training: Online Course, Open Program or Custom
Key takeaway
There are three main AI training formats: self-paced online courses, open programs that bring together people from several companies, and custom training inside your company. An online course lays the foundations, an open program quickly upskills an individual, and custom training gets a whole team working with its own tools, data and rules.
Key points
- An online course lays the foundations but rarely changes how people work.
- An open program is efficient for one or two people with the same profile.
- Custom training works on your own tasks, tools and confidentiality rules.
- The best approach often combines a shared foundation with role-based workshops.
- Keep a record of training: it also helps show your AI literacy measures.
How to choose AI training: the short answer
Choose an online course to lay the foundations, an open program to quickly upskill one person, and custom training to get a whole team working with its own tools and rules. The right choice depends on who you are training, for what purpose, and how many people share the same tools.
This article complements our guide AI training for business: where to start. That guide explains how to assess your team's needs; here we compare the training formats you can choose from.
The three AI training formats
Self-paced online courses
An online course is the fastest and cheapest way to learn the basics of AI. You take it whenever you want, often for free or a small fee.
A well-known example is Elements of AI, a series of free online courses created by MinnaLearn and the University of Helsinki. AI tool vendors also offer explanations and tutorials.
The limit: a general course does not cover your tasks, your tools or your confidentiality rules. Many participants drop out or apply little of it in their daily work.
Open programs
An open program brings together people from several companies around a theme, such as "ChatGPT for marketers" or "AI for HR". It is run by training centers, professional federations or colleges.
Its strength: quickly upskilling one or two people, who also exchange with other companies. Its limit: exercises stay generic, and you cannot use your company's confidential documents.
Custom training in your company
Custom training is built around your tasks, tools and rules, for your own team. It usually starts from a short assessment and combines a shared foundation with role-based workshops.
Its strength: your team practices on your own documents, learns the rules specific to your company and leaves with reusable prompts. Its limit: it takes more preparation and pays off only when several people share the same tools or tasks.
The comparison in one table
Each format has its place: the question is what your team needs today.
| Criterion | Online course | Open program | Custom, in-house |
|---|---|---|---|
| Best for | Everyone, as a foundation | 1-2 people with the same profile | A team or the whole company |
| Content | General | Thematic | Your tasks and tools |
| Exercises | Standard | Fictional cases | Your own documents |
| Usage rules and GDPR | General | General | Tailored to your company |
| Language | Depends on the offer | Depends on the session | French, Dutch or English, your choice |
| Follow-up | None | Limited | Possible, with adoption tracking |
| Cost | Free to low | Per participant | Quoted, for the group |
Which combination works best?
For most small and mid-sized businesses, a combination works best: a shared foundation for everyone, then hands-on role-based workshops. Here is a sample path:
- Foundation for everyone: what AI does, when it gets things wrong, which data never to share.
- Usage rules: one page with approved tools, prohibited data and the duty to check.
- Role-based workshops: sales, marketing, admin, customer service, on their own cases.
- Champions: a few motivated colleagues who answer questions and share good prompts.
- Follow-up after a few weeks: what the team really uses, what can improve, what can be automated.
8 questions to choose AI training
Ask every provider these questions: they quickly show whether the training will really move your team forward.
- Does the training start from an analysis of our tasks and tools?
- How much time goes to practice, and on which documents?
- Are confidentiality, GDPR and fact-checking covered?
- Does the trainer work on AI projects, or only teach?
- Does the training cover the tools we already have, such as Copilot or ChatGPT?
- Can it be delivered in French, Dutch or English, depending on the team?
- Is there follow-up after the sessions?
- Do participants receive a certificate, and how can we document the training?
AI training and the AI Act
Training also helps you meet your legal obligations. Article 4 of the AI Act requires companies using AI to take measures for their staff's AI literacy. No certificate is required, but document what you do. We explain this in AI Act Article 4: AI literacy, and our article ChatGPT at work includes a policy to get started.
On funding, regional support exists in Belgium, such as VLAIO in Flanders or chèques-entreprises in Wallonia. Budgets and conditions change regularly, so check them before deciding. Our guide to funding an AI project in Belgium covers each region.
Our approach at Ozapp
Ozapp is an AI-first agency based in Belgium that builds AI agents, chatbots and automations. Our custom training draws on that hands-on experience: a short assessment, exercises on your own documents and clear usage rules. Every participant receives a certificate of attendance. Explore our AI consulting & training service, or start with our free AI maturity diagnostic.
- Explore our service: AI consulting & training
- Get a free quote: Contact
FAQ — How to choose AI training
Are there free AI training courses?
Yes. Elements of AI, a series of free online courses created by MinnaLearn and the University of Helsinki, explains AI basics for non-specialists. AI tool vendors also offer free tutorials. These courses are a good starting point, but they do not teach your team to work with your own documents and rules.
What is the difference between AI training and ChatGPT training?
ChatGPT training focuses on one specific tool. AI training covers more broadly how generative AI works, its limits, fact-checking, confidentiality and role-based uses, whatever the tool. Ideally, you train on the tool already available in your company, on top of that general foundation.
How many people do you need for custom training?
There is no fixed minimum. Custom training becomes worthwhile as soon as several colleagues share the same tools or tasks, because you then work on your own documents and rules. For a single person, an open program or online course is often more efficient.
How much does custom AI training cost?
The price depends on the number of participants, the number of sessions, preparation on your own cases and whether an assessment or AI policy is included. There is no single rate. At Ozapp, you receive a detailed quote within 24-48 hours after a free 30-minute consultation.
Is AI training mandatory for companies?
Article 4 of the AI Act requires companies using AI to take measures to support their staff's AI literacy. No specific training or certificate is mandatory, but training is the most obvious measure. Keep a record of who completed what: according to the European Commission, an internal record is enough.
