Every few months a study lands showing that people who learn something in an immersive environment — VR, 360° projection, walk-through simulation — score the same as, or worse than, people who learned the same content from a slide deck. They report enjoying it more. They report feeling like they learned more. The test does not agree.
This finding is then usually framed as a debunk. See, immersive is hype. The room is theater. The medium is the message and the message is vibes. Vendors rebut with their own studies, mostly funded by themselves, showing decision confidence and recall up. Everyone goes back to their corner.
I design immersive AI demonstration environments for senior audiences. I think both sides are measuring the wrong thing.
The literature keeps asking whether immersive rooms produce learning. They mostly don't, in the way the literature defines learning — declarative knowledge, recallable later, transferable to a test. Engagement and learning come apart in nearly every controlled study, and the room's sensory richness frequently makes the cognitive load worse, not better. If your goal is to teach a senior executive a concept they will retain six weeks from now, the room is not the best tool. A short paper would do more. A conversation with a sharp peer would do more still.
But the room was never built to teach. It was built to do something the literature has no good instrument for, which is why the literature keeps missing it.
The standard format for selling a serious idea to a senior person is a deck and a meeting. The deck is broadcast. It says: here is the problem, here is our solution, here is the ask. It moves in one direction, controlled by the presenter, optimized for narrative tightness. The senior's job in this format is to react to what's being delivered. They can ask questions, but the questions are downstream of the deck's framing. The format decides what's salient. The buyer reacts.
The immersive room inverts that. The room offers a set of entry points and lets the senior pick. The first thing they walk toward is information you cannot get any other way. It tells you what they actually came in worried about, before they have edited themselves into the question they think they're supposed to ask. The second thing they walk toward tells you whether the first was real curiosity or politeness. By the time they've moved through three or four choices, you have a high-fidelity read on what they care about that no discovery call would have produced, because a discovery call invites them to narrate themselves, and senior people narrate themselves the way they want to be seen.
The room produces something a deck cannot: an unguarded preference order.
Once you see that, the rest of the design follows. The visuals are not the product. The visuals are the permission structure — they make a senior person willing to spend forty-five minutes in a room with you instead of eight minutes at a booth. They lower the social cost of curiosity, because in an immersive environment looking at something is not the same as committing to it. You can walk toward an idea and walk away, and nobody has to interpret the walking-away as a no. That's a feature most formal sales formats lack. The room buys you permission to be wandered through, and wandering is where the real signal lives.
The work, once they're in the room, is not delivery. It's reading. It's noticing what they paused on, what they walked past, what they doubled back to. It's adjusting the conversation in the next thirty seconds based on the last five. The content is, at this point, somewhat interchangeable — there are five things you could talk about, and the buyer, by their feet, has told you which two matter. The skill the format rewards is the willingness to drop the other three and follow.
This is harder than it sounds, and it's the reason most organizations that invest in immersive setups don't get this kind of value out of them. Most groups that demo for a living have built their muscle around controlling the conversation. Scripted decks, rehearsed Q&A, a defined ask at the end. That discipline is not a flaw — it's what makes those teams good at the formats they were built for. But the immersive room punishes it. The moment a senior person picks a path the script didn't anticipate, the script holder has two options: herd them back to the planned path, or abandon the script and follow them. Almost everyone, on instinct, herds. Once you herd, the room collapses into a worse version of the meeting it was supposed to replace, with more pixels.
This isn't a consulting problem or a vendor problem. It's a problem for any function whose default mode is deliver instead of listen — sales engineering, investor relations, board prep, internal program rollouts. The room rewards a different muscle than the one those functions were optimized for.
The corollary, which is the part that matters for AI specifically: when the demonstration is about an AI capability — a synthetic data generator, an agent, a model in production — the temptation to script is even higher, because the technology is unfamiliar and the presenter is anxious about handling questions they haven't rehearsed. So the AI demo, in most rooms, becomes the most script-locked format of all. The buyer ends up with less information about what they actually wanted to know, because the presenter spent the time defending a predetermined narrative about what the AI does. The room was set up to surface their real question. The script made sure it didn't.
The way to use the room well, for AI in particular, is to assume the buyer's real question is not the one in your deck and design for that. Build entry points around the questions you keep being asked at the end of meetings, the ones that come after the formal Q&A is over and people are walking out. Those are the questions worth surfacing earlier. Let the buyer find them. Then talk about those.
So when someone asks me whether immersive demonstrations produce real learning, I have stopped arguing about the studies. The studies are right. They aren't learning environments. They are listening environments dressed as learning environments — and the dress is load-bearing, because nobody senior would sit still for forty-five minutes of being listened to if you called it that. The visuals buy the time. The choice architecture surfaces the preference. The speaker, who is the only part of the room that can actually adjust, does the work.
The room is a diagnostic. The speaker is the instrument. The content is a menu. None of those three things show up in the engagement metrics, which is why the engagement metrics, in both directions, are mostly noise.