You've decided on corporate AI training. The budget is approved and the participant list is ready. The next question often sparks more debate than expected: in-person or online?
One side says "nothing replaces the classroom"; the other objects that "teams are in different cities, and bringing everyone together is expensive and hard." Both can be right; the answer depends on the content and purpose of the training.
What does the research say?
One of the most cited studies is a 2010 meta-analysis prepared for the US Department of Education. The researchers screened more than a thousand studies from 1996–2008 and compared 50 independent effects from those meeting quality criteria. Most came from adult settings such as higher education, medical training and professional development.
Findings:
- Learners in online conditions performed modestly better on average than those taught face-to-face.
- The difference was larger in blended designs combining online and face-to-face elements.
- But the researchers add an important caveat: blended conditions often included extra learning time and additional instructional elements. So the difference should be attributed to design, not the medium.
In short, the real message is: format matters less than design. A well-designed online course can beat a poorly designed classroom course, and vice versa.
One more finding: spacing improves retention
One of the most robust findings in cognitive psychology is that spaced practice produces more durable learning than massed practice. Cepeda and colleagues' 2006 meta-analysis reviewed 317 experiments from 184 articles, confirmed the effect, and showed that the ideal gap between sessions grows as the required retention period lengthens.
The implication for corporate training is clear: instead of filling two consecutive days, splitting the same content into short sessions spread over several weeks increases the chance that learning carries over to work. The online format makes this kind of spacing much easier logistically.
Comparison
| Criterion | In-person | Online | Hybrid |
|---|---|---|---|
| Interaction and discussion | Very strong | Depends on design | Strong |
| Hands-on workshops | Very suitable | Suitable (needs good design) | Very suitable |
| Teams in different cities | Hard and costly | Very suitable | Suitable |
| Spacing over time | Logistically hard | Very easy | Easy |
| Attention and focus | High | Prone to distraction | Medium–high |
| Team bonding and shared language | Very strong | Limited | Strong |
| Recording and replay | No | Possible | Partly |
Which course suits which format?
Awareness and safe use (2–3 hours): Works very well online. Short, focused content that must reach large groups at once. Our AI Awareness Training fits this format.
Strategy and roadmap workshops: Should be in person. Executives discuss and prioritise their own use cases and reach shared decisions. That's why our AI Strategy & Roadmap Workshop works best face-to-face.
Hands-on training for business teams: Hybrid is ideal. An opening session, then tasks participants try in their own work, and short online follow-ups. Generative AI at Work delivers very good results in this set-up.
Technical training (Python, machine learning, LLMs, MLOps): Both formats work, but a shared practice environment is essential. Prepare it in advance so participants don't lose time on installation problems.
7 principles for effective online training
- Short sessions: 90–120 minute blocks instead of a full day.
- Practice tasks in between: A concrete task for participants to try in their own work after each session.
- Small groups: Interaction drops as numbers grow; keep hands-on sessions small.
- Cameras and active participation: Questions, polls, breakout group work.
- Practise with the organisation's own tools: Work with the approved tools participants will actually use.
- Recordings and materials: Session recordings and short summaries for review.
- Follow-up session: A "What did you try, where did you get stuck?" session a few weeks later.
5 questions to decide
- Is the goal knowledge transfer or a shared decision? (Decision → in person)
- Are participants in the same city? (Dispersed → online or hybrid)
- How many participants? (Large awareness groups → online)
- Is a shared practice environment needed? (Yes → preparation is essential)
- How will you measure transfer to work? (Plan measurement from the start, whatever the format)
You can use the four levels of the Kirkpatrick model (reaction, learning, behaviour, results) to measure impact; see how to measure the impact of AI training for details.
Conclusion
There is no single right answer to "in-person or online?". The right question is: which design best serves this training's purpose? For most organisations, the best results come from running strategic topics in person, broad awareness online, and hands-on training as a hybrid programme spread over time.
All MLTeam courses can be delivered in person or online, and participants receive a certificate of attendance. Use the Training Finder to see which courses suit you, or browse all courses.
Sources