Some honest thoughts on AI in escape rooms, creativity, and what it actually looks like day to day at Prodigy.
We’ll be upfront: we’ve experimented with AI in escape rooms and beyond. ChatGPT, Claude, Midjourney, Meshy.ai and Suno have all found their way into our workflow at various points. We’re not here to evangelise about it, and we’re not here to pretend it hasn’t raised questions worth sitting with. We just want to share our honest experience as two people who have spent years building immersive worlds out of very little, and who are always looking for tools that help us do that better.
Because that’s really the word. Tools.
Covid pushed almost everything into a screen. Meetings, celebrations, catch-ups, entertainment: for a long stretch, digital was the only option available, and people made the best of it. But something was missing, and most of us felt it. When the world opened back up, we started Prodigy Escapes because we believed in what a shared physical experience could offer. A room full of people, working together, away from their phones, fully present with each other. That belief hasn’t changed. If anything, it shapes every decision we make about how our rooms are built and run.
A brief history of worrying about progress
Every generation has its version of this conversation. The printing press would make memory obsolete. The camera would kill painting. Synthesisers would destroy music. The internet would end books. Some of those fears contained a grain of truth. Mostly, the thing people were afraid of became part of the landscape, and the creative work adapted around it.
We’re not saying AI in escape rooms is the same as a camera. It’s clearly more complex, more far-reaching, and more contested. But we do think the instinct to ask “is this a threat or a tool?” is a reasonable place to start. And from where we’re sitting, in a small creative business in Exeter, it’s mostly felt like a tool.
What running a small escape room actually involves
Before we get to the AI part, it’s worth painting a picture of the job. Between us, we have plastered walls, laid flooring, built websites, troubleshot props at midnight, recorded audio, edited video, written scripts, done the accounts, trained staff and fixed things twenty minutes before a group of twelve people walked through the door.
We didn’t train for any of this. We learned it because we had to, and because the vision in our heads was bigger than the budget in our accounts. Most escape room owners will recognise this picture. You become a jack of all trades not by choice but by necessity, and you get very good at finding ways to stretch what you have.
AI in escape rooms, in that context, is one more thing in the toolkit. It helped us fix a stubborn line of code when we were stuck. Provided suggested spell names and potion combinations when we’d exhausted our own imagination. It offered rhyming words at 11pm when the puzzle needed one more clue and our brains had given up. Small things, mostly. Useful things.
The thing nobody talks about: creative confidence
Before escape rooms, Dan worked in the motorcycle industry and collaborated regularly with professional artists and designers. It should have been straightforward: describe the vision, receive the work, refine from there. In practice, it often felt like a guessing game. You’d describe what you imagined and receive something adjacent to it. Something close, but not quite right. You’d iterate. They’d iterate. The result was fine, sometimes great, but the gap between what was in your head and what ended up on screen was often significant.
With AI image tools, that gap has almost disappeared. Not because the output is always better, it often isn’t, but because you can chase the vision directly. You can try a hundred variations in the time it used to take to write a brief. You can let the image tell you what it wants to be. The brief emerges from the exploration, not the other way around.
Working with AI tools feels different from working with search engines or reference libraries. It feels more like having a tireless collaborator who doesn’t get frustrated when you change your mind. That’s a genuinely new experience, and it’s changed how we approach the early stages of a build.
We use it most heavily in the ideation phase: brainstorming names, testing concepts, exploring visual directions. Less so in execution, where craft and care still matter enormously. The final result in our rooms is still overwhelmingly human. The AI is a starting point, not a destination.
What AI in escape rooms means for small creative businesses
For small businesses like ours, AI in escape rooms and creative work levels a playing field that was previously tilted steeply towards bigger studios with larger budgets. The ability to visualise concepts quickly, test ideas before committing, and explore creative directions without expensive iterations: that’s genuinely democratising. We can punch above our weight in ways that simply weren’t possible before.
Three rooms and one memorable guest
We’d like to tell you about an enthusiast who played all three of our rooms. They were open about having strong feelings on AI use in escape rooms. We were curious to see what they made of the experience, because the three rooms tell quite different stories in that regard.
Room one – The Haunting of Mount Clifton Manor
Our Haunting of Mount Clifton Manor has an introduction video. The original version was a still image of a country house with a voiceover kindly recorded by a family friend, well-intentioned but a little flat. When we discovered Midjourney, we used it to try and visualise Lady Clifton, to give form to a character who until then had only existed in our imaginations. The image felt right. We commissioned someone to animate it. The result is imperfect, you can tell it was generated, and our guest noticed. They had reservations about it, which we completely understand.
Room two – The Lost Temple
No AI anywhere in the Lost Temple. The introduction features Dan’s old school teacher, who is a genuinely engaging performer and brings real warmth to the role of the Professor. With age being a factor, he can occasionally be spotted reading from his script just below the camera. A thoroughly human creation, small imperfections and all. Our guest noticed that too.
Room three – Wizards of Wyvern
The Wizards of Wyvern features over seventy custom voice recordings from our dear friend David McMullan, who brings the whole experience to life. AI played a supporting role throughout the build, helping debug code, generating imagery, suggesting names for potions and spells, finding rhymes for puzzle clues. Not the thing itself, but the scaffolding around it. Our guest, who came in sceptical, had a wonderful time and said they didn’t think about the AI question once they were inside. That felt like the right outcome.
One thing worth mentioning, because it speaks to the broader question of what AI is actually for, is how we approached the hosting of Wizards of Wyvern.
A room like Wizards could, technically, run itself. The timing of hints, the progression of the story, the triggering of effects: all of it could be automated. We had the capability to build it that way. We chose not to.
Every team that plays Wizards of Wyvern has a dedicated member of the Prodigy team watching over their experience. Monitoring what’s been opened, reading the room, choosing the right moment to play a sound effect or nudge the story forward. The game is built so that a human has to be present and engaged for it to work properly. That wasn’t an accident or a limitation. It was a decision made because we believe the human presence in that room matters, and that no amount of clever automation can quite replace someone who is genuinely paying attention.
AI helped us build the experience. It didn’t replace the experience. That distinction feels important to us.
On creativity and the tools we choose
Escape rooms are, at their heart, physical experiences. A chance to bring people together, away from their phones and away from the temptation to Google the answer. What we deeply care about is the experience a group of people has together. The laughter, the frustration, the moment when something clicks.
AI doesn’t create that. People create that. What AI can do is help people with limited time and resources express a creative vision they might not otherwise have been able to realise. That feels worth something to us.
We think about the portrait artist and the photographer. The hand-drawn illustration and the stock image. The electric drill and the hand crank. At some point, the tool fades into the background and the experience, what it feels like, what it makes possible, is all that remains.
There should never be a cap on the amount of creativity in the world. More ways to make things means more things get made, and some of those things will be wonderful. We’d like to keep trying to make some of them.
We’d love to know your thoughts, whether you’re a guest who’s played our rooms, a fellow owner navigating the same questions, or just someone who finds this conversation interesting. It’s one worth having, and we’re genuinely glad it’s happening.