We had five minutes, two datasets, and an audience waiting. Could AI turn raw survey data into a professional presentation before time ran out?
That was the challenge we set ourselves for a live demo at the end of a recent bootcamp. Not a polished, pre-prepared showcase, but a genuine test: take real data, feed it through AI, and produce something presentable while people watched. It was a bit tongue-in-cheek, deliberately rough around the edges, but the concept we wanted to prove was serious. If this worked, it would demonstrate that the gap between raw information and professional output has fundamentally changed.
The traditional route for something like this is familiar to anyone who’s built a presentation under pressure. You collect your data, spend time analysing it, pull out the themes, structure your narrative, design your slides, refine the wording, check for typos, rehearse, refine again. Even for a five-minute presentation, you’re looking at a few hours of work. It’s not that any individual step is difficult, it’s that the cumulative effort adds up, and most of it isn’t the high-value thinking. It’s the grunt work of getting ideas into a format that looks professional.
We wanted to see if AI could compress that.
The Setup
We had two datasets to work with. The first captured the skills and capabilities developed over the course of the bootcamp, a record of what participants had learned and could now do. The second was market research we’d gathered by asking a simple question: if you could wave a magic wand and fix something with AI, what would it be? This gave us a picture of real pain points, the problems people actually wanted solved, not what we assumed they needed.
The goal was to bring these together, show our capabilities, show the market demand, and demonstrate where the two aligned. Three slides, clear narrative, professional standard.
The method was straightforward. We fed both datasets into ChatGPT alongside a prompt I’d prepared in advance. ChatGPT analysed the information and generated a prompt for Gamma, a presentation platform. When we ran that prompt in Gamma, it produced a three-slide deck ready to present.
Total time from data input to slides on screen: under five minutes.
The Friction
It wasn’t seamless. There were moments along the way that reminded me this is still a new way of working, and new ways of working come with friction.
The first challenge was confidence. For some people involved, the concept felt unfamiliar. Trusting that the process would work, committing to a live demo without a safety net, required a leap. As the person behind the scenes, I found myself working to establish trust and buy-in, supporting others through moments of doubt. That’s not a technical problem, it’s a human one, and it took more effort than the AI part.
We also hit a technical limitation mid-demo. The free model we were using couldn’t handle the volume of data we’d fed it in one go. Ironically, this was only because the data collection had been so successful, we’d captured rich, detailed responses that exceeded what the model could process at once. A problem caused by abundance, but a problem nonetheless.
And of course, the demo itself wasn’t perfect. It could have used more rehearsal. The physical environment, the flow, the delivery, all of that still required human preparation. AI didn’t replace the need to be ready for the room.
What It Revealed
Despite the rough edges, the concept held. We demonstrated that AI can get you 80-90% of the way to a professional standard in minutes rather than hours. That’s not a small thing.
But here’s what I’d reflect on: that remaining 10-20% is where the real value lives. The refinement, the judgment calls, the human context that makes a presentation land with a specific audience. AI compresses the grunt work, it doesn’t replace the thinking. What it does is free up your time to focus on the bits that actually matter.
If I’ve only got an hour to produce a presentation now, I know I can spend the first few minutes generating material and the rest of the time making it right for the room. That’s a different way of working, and it changes what’s possible under pressure.
There’s a broader application here too. For managers and teams, this approach offers a rapid prototyping capability. Get to a minimum viable standard quickly, then focus collective effort on refinement. The last 10-20% is where teams add value, so getting to 80% in rapid time accelerates everything that follows.
The Takeaway
The demo wasn’t flawless, but that’s partly the point. AI capability is real, and it’s accessible now, even with free tools and a bit of preparation. The harder work isn’t the technology. It’s the confidence to try it, the trust to commit, and the human judgment to refine what it produces.
If you’ve been curious about AI for content creation but unsure where to start, this kind of exercise is low-risk and high-learning. You don’t need perfect conditions. You need a clear question, some data, and the willingness to see what happens.
If you’d like to explore how this approach might work in your context, get in touch for a chat. I’m always happy to think it through with you.