Key takeaways
✓Remote and self-paced AI training suits individuals building foundational knowledge, but it rarely produces the shared understanding that makes AI adoption stick across a team.
✓In-person delivery creates the peer discussion, live problem-solving and social accountability that accelerate behaviour change, not just knowledge transfer.
✓The format question is not binary. The strongest enterprise AI programs often combine a live workshop to build shared context with follow-up resources for individual reinforcement.
✓Certain teams get an outsized return from in-person sessions: those adopting AI into collaborative workflows, those where trust and psychological safety matter, and those where hands-on practice with real work scenarios is the fastest path to competence.
✓Making the case to leadership comes down to one question: are you trying to inform people, or change how they work?
Why does delivery format matter for AI training?
AI training is not like learning a new expense system or completing a compliance module. The tools change every few months, the right answer depends heavily on context, and the skill you are actually building, the ability to think alongside an AI, is harder to demonstrate on a slide than to discover through practice.
That gap between knowing and doing is where delivery format starts to matter.
Most enterprise L&D content transfers information. You watch, you read, you tick the box. AI fluency is different. It requires people to form new mental models about how work gets done, which means they need space to experiment, make mistakes, ask questions that sound foolish, and observe how colleagues in the same role think through the same problems.
The skill you're building isn't knowledge retention
AI fluency is a behaviour change, not a knowledge transfer. That distinction should drive every format decision you make for AI training.
Self-paced e-learning can introduce concepts, but it rarely changes behaviour. A short video on how to write a prompt does not prepare a finance analyst for the moment they need to use Copilot to draft a stakeholder briefing under time pressure. The gap between understanding a concept and applying it under real conditions is where most AI training loses people.
Delivery format also shapes how teams adopt AI together. Prompting is a team skill, not an individual one. When everyone in a team learns separately, they develop inconsistent habits, miss the chance to build shared language, and end up with low collective adoption even when individual completion rates look fine. The format you choose either supports that collective shift or works against it.
None of this means in-person is always the answer. But it does mean the stakes of getting format right are higher for AI than for almost any other learning investment your organisation will make this year.
What does in-person AI training actually do differently?
The short answer: it changes the social dynamics of learning, and for AI adoption, those dynamics matter more than the content delivery itself.
Peer learning happens in the room, not on Slack
When a group of people works through the same AI task together, something useful occurs. One person finds a shortcut. Another spots a failure mode. A third asks the question everyone was too self-conscious to type into a chat window. That exchange, unscripted and immediate, is often where the most durable learning lands.
Remote sessions can replicate this in part. Breakout rooms work. Collaborative documents work. But there is a threshold of spontaneity that asynchronous tools consistently fall short of, and it matters most at the beginning of an AI capability journey, when people are still forming mental models of what these tools actually do.
Shared context produces shared practice
In-person training works best when the cohort has something in common, a department, a workflow, a shared tool set. When that condition is met, the room becomes a place to work through real problems together, not generic exercises designed for no one in particular.
A finance team learning to use Microsoft Copilot, for instance, will surface questions specific to how their organisation structures reporting, approves spending, or handles month-end close. A facilitator who can take that thread and run with it, adapting examples on the fly, covers more practical ground in four hours than a polished e-learning module covers in four weeks.
This is not a dig at e-learning. It is a structural point. Generic content cannot anticipate your organisation's specific vocabulary, edge cases, or anxieties about AI. In-person delivery, in the right hands, can respond to what is actually in the room.
The facilitator effect is real
A skilled facilitator does not just present slides. They read resistance, answer the awkward questions, and adjust the pace when a concept is not landing. That responsiveness is difficult to replicate in recorded or self-paced formats, and it is especially valuable when the topic is new enough that participants do not yet know what questions to ask.
Real-time practice with immediate feedback
AI fluency is a practical skill. Understanding what a language model does conceptually is useful background. Actually constructing a prompt, reviewing the output critically, and iterating, that is where the skill develops.
In-person sessions make it easier to structure that practice at the right pace. A facilitator can pause the room, look over shoulders (literally or figuratively), spot a common mistake across multiple screens, and address it once for everyone. That density of feedback loops is hard to achieve when participants are distributed across time zones and working asynchronously.
There is also the question of accountability. Sitting in a room with colleagues and a facilitator creates a kind of productive pressure that solo e-learning rarely generates. People stay engaged because opting out is conspicuous. That is not a trivial point for training that asks people to change habitual ways of working. As the article on why enterprise AI training fails makes clear, disengagement is one of the most consistent predictors of poor adoption outcomes.
When is remote or self-paced the right call?
In-person training is not the right answer every time. There are genuine scenarios where async or virtual delivery serves teams better, and pretending otherwise would be dishonest.
Geographic spread is the obvious one. If your workforce is distributed across five states or three time zones, getting everyone into the same room is expensive and logistically painful. For foundational AI literacy, where the goal is broad awareness rather than deep behaviour change, a well-designed self-paced module can reach more people at a fraction of the travel and venue cost. The L&D leader's playbook for a global AI training rollout covers this in detail, but the short version is: wide baseline coverage and deep cohort work are two different jobs, and they often call for different formats.
Refresher and reference content suits async well. Once a team has completed foundational training, the follow-up work, prompt refinement, feature updates, edge-case guidance, tends to be better served by short on-demand modules they can pull up when they need them. Asking busy professionals to attend a live session to learn about a minor product update is not a good use of anyone's time.
Individual pacing matters for some learners. A skilled analyst who already uses AI tools daily does not need to sit through two hours of context-setting aimed at someone who has never opened Copilot. Self-paced paths let people move through material at a speed that matches where they actually are, rather than where the slowest person in the room is. If you have not yet established where your workforce sits on the capability curve, assessing AI readiness before you commit to a format is time well spent.
The format should follow the goal
In-person works best when the goal is behaviour change, team alignment, or building shared practice. Remote and async work best when the goal is coverage, individual pacing, or reinforcing skills already introduced. Choosing the format before you have defined the goal is where most L&D decisions go wrong.
Budget constraints are real. Facilitated in-person workshops cost more to design, deliver, and host. If your training budget is limited and you need to reach two hundred people this quarter, a blended approach, remote for foundation, in-person for the teams where adoption is most critical, is often the most defensible answer.
The point is not that one format beats the other. It is that they solve different problems. The mistake most organisations make is defaulting to virtual because it is cheaper and easier to schedule, not because it is actually the right fit for what they are trying to change.
Which teams benefit most from in-person AI training?
In-person delivery pays off most clearly when the work is collaborative, the stakes are high, or the team is starting from a low baseline of AI confidence. The return on the room drops when individuals already have the fundamentals and just need to build habits over time.
A few team types where the case is particularly strong:
Senior leadership and executive cohorts. Executives rarely finish an e-learning module. They will, however, show up and engage for a half-day session with peers when the agenda is framed around decisions, not tools. In-person format lets the facilitator read the room, adjust the depth, and handle the inevitable "but what about our situation" questions that no pre-recorded content can anticipate. The peer dynamic also matters: when the CFO hears the COO working through a prompt live, it normalises the learning in a way a dashboard of completion rates never will.
Teams adopting a new AI tool organisation-wide. A Microsoft Copilot or Google Gemini rollout is not just a software deployment. People need to rebuild habits around how they draft, summarise, search and communicate. In-room workshops let participants practise on their actual files and workflows, ask questions in the moment, and leave with something they built themselves rather than a slide deck they watched. That distinction matters for adoption. The research on why Copilot adoption stalls points consistently to a gap between initial training and real-world use, and in-person delivery closes that gap faster than most alternatives.
Cross-functional project teams. When multiple roles need to collaborate using AI, training them together is more efficient than training them separately. A procurement manager and a finance analyst who learn prompting in the same room can start building shared prompt libraries and common language from day one. Separating them into individual e-learning tracks and hoping the coordination happens later is optimistic.
Teams with low AI confidence or high scepticism. Psychological safety is a real factor in AI adoption. People who are worried about looking foolish, being replaced, or making a costly mistake are not going to experiment freely in a self-paced online environment where no one is watching. A skilled facilitator in a room can name those concerns directly, demonstrate that confusion is universal, and create enough safety for people to try things and fail without consequence. That shift in mindset is often worth more than any individual skill taught in the session. It is also exactly the kind of thing you cannot replicate in a recorded video.
Customer-facing or high-compliance roles. Teams in financial services, legal, healthcare, or government often need to understand not just how to use AI tools but where the boundaries are: what they can put into a prompt, what they cannot, what the organisation's policy is and why. Those conversations benefit from a facilitator who can handle edge cases and push back on assumptions in real time. A compliance officer asking "what happens if a client name ends up in a prompt?" deserves a live, considered answer, not a link to a FAQ.
The pattern worth noticing
The teams that gain most from in-person delivery are not necessarily the least technically capable. They are the ones where shared understanding, trust, or real-time judgement matters most to the outcome.
How do you make the case to leadership?
The honest answer is that you probably won't win this argument on principle. "People learn better face-to-face" sounds like intuition, not evidence, and a CFO reviewing a travel and venue budget will want more than that. The case needs to be practical, specific to your organisation, and tied to something leadership already cares about.
Frame it as a deployment cost, not a training cost
The most common objection to in-person AI training is the line-item: flights, accommodation, a facilitator, half a day of lost productivity per attendee. But that framing treats training as an expense rather than part of a larger investment.
Your organisation has already spent money on AI licences, implementation, and change management. If low adoption is the risk, the question isn't "why does training cost this much?" It's "what does low adoption cost us?" A team of twenty people underusing a Copilot licence for twelve months is a material number. Put that number in the conversation.
The question that changes the framing
Don't ask whether in-person training is worth the cost. Ask what underperforming AI adoption costs the business over the next twelve months, and whether that number is larger than the training budget.
Anchor to a specific outcome, not a general capability
Leadership are more likely to approve a programme with a defined scope than one sold as general AI literacy. "We want our finance team to reduce the time spent on monthly reporting" is an easier yes than "we want to improve AI fluency across the organisation."
This also makes measurement easier. If you can point to a workflow before and after, you have something to report back on. That's useful when you're building the case for the next cohort.
For guidance on how to structure that kind of baseline, the article on how to build an AI literacy baseline across your organisation is a practical place to start.
Address the logistics objection early
If you have a distributed team, leadership will often raise logistics before they raise cost. The response isn't to minimise the complexity. Acknowledge it, and then show you've thought it through.
A few options that tend to work in practice:
Hub-and-spoke delivery: Train one cohort per city rather than flying everyone to one location. A facilitator travels; the participants don't.
Anchor to existing gatherings: If your team has a quarterly offsite, an all-hands, or a planning week, attach the training to it. The travel cost is already absorbed.
Start with a pilot cohort: Rather than committing to a national rollout, propose one city, one team, one day. A small, well-run pilot produces the internal evidence you need to justify the rest.
Be honest about what you're not claiming
One thing that tends to build credibility with sceptical leaders is naming what in-person training won't fix. It won't compensate for unclear AI policy, a tool that hasn't been properly integrated into existing workflows, or a team that hasn't been told why they're being trained. Those are separate problems.
What in-person delivery does well is accelerate the shift from passive awareness to active, habitual use, particularly in teams that work closely together and need shared language and shared confidence to move forward. Keep the claim tight, and it's easier to defend.
Not sure how to frame the business case for your organisation?
We can help you work through the numbers, scope a pilot, and build a proposal that holds up to scrutiny. Bring the context; we'll help you shape the argument.
Book a 30-minute discovery call →
Frequently asked questions
Is in-person AI training worth the cost when remote options are cheaper?
In-person AI training tends to deliver better adoption outcomes for teams where the goal is behaviour change, not just awareness. The higher upfront cost includes facilitated practice, live problem-solving on real workflows, and peer learning that self-paced modules cannot replicate. For a team that already struggles to apply AI tools after completing e-learning, the question is not whether in-person costs more; it is whether the cheaper option is actually delivering anything.
How long does a typical in-person AI training session run?
Most enterprise AI workshops run between half a day and two full days, depending on scope. A half-day session works well for awareness and tool introduction. A full day allows time for hands-on practice with realistic scenarios. Two-day formats suit teams building more complex workflows or working across multiple tools. The right length depends on what participants need to be able to do when they leave the room, not on what fits neatly into a calendar.
Can in-person and remote formats be combined effectively?
Yes, and for distributed teams this is often the most practical structure. A common approach is to run one in-person session with a core cohort or team leads, then follow up with shorter virtual sessions to reinforce skills and answer questions as they arise in practice. This is particularly relevant for Australian organisations with staff across multiple states, where flying everyone to a single location is not always feasible. The L&D leader's playbook for a global AI training rollout covers how to sequence delivery across geographies.
What if our team is already comfortable with the AI tools we use?
Comfort with a tool and proficiency in applying it to real work are different things. Teams that use Microsoft Copilot or another AI assistant regularly often develop individual habits that may not reflect best practice and are rarely shared across the group. In-person training for experienced users tends to focus less on the basics and more on prompting as a shared team skill, workflow integration, and spotting gaps that individuals do not know they have.
How do we measure whether in-person AI training actually worked?
The most reliable indicators are behavioural: are people using the tools differently in the weeks after training, and are they doing so consistently across the team? Pre- and post-training benchmarks, manager observation, and tracking tool usage data where available all contribute to a useful picture. Completion rates and satisfaction scores tell you whether people showed up and enjoyed themselves, not whether anything changed. The how to measure ROI on enterprise AI training article outlines a practical framework for building this into your program from the start.
Ready to explore in-person AI training for your team?
If your organisation is at the point where AI tools are deployed but adoption is patchy, or where a rollout is planned and you want to get it right from the start, a facilitated workshop is often the fastest way to move from awareness to actual use.
Better People's workshops are designed for exactly this situation: working teams, real tools, practical outcomes. Sessions can be delivered onsite across Australia, built around your specific stack and workflows, and calibrated for the mix of technical confidence in your cohort.
Explore our workshop options or get in touch to talk through what delivery format makes most sense for your teams.
