95% of AI pilots fail to deliver impact according to a MIT study. Not because the tech isn’t ready, but because it’s actually not about the tech – it’s about behaviour change. A much trickier beast altogether.
One of the simplest practices we’ve found works wonders for this challenge is a disarmingly simple meeting format called AI Show & Tell. Sounds almost trivial, but it’s turned out to be transformational.
Every week, we have a 45 minute meeting where each person has five minutes to show something they’ve done with AI. It might be a new tool they tried (or built), a prompt that worked out well, a workflow they hacked, or even just a clever idea they spotted elsewhere.
The first rule: it has to be shown, not just talked about. The second rule: we ruthlessly time each person using the timer on a phone.
That’s it. Five minutes. Show us, don’t tell us.
It’s the best meeting of the week – everyone’s favourite bar none. We come away inspired by our colleagues, amazed by what’s possible, and more connected across disciplines. Ideas leap from one domain to another. A challenge one person has been stuck on suddenly sparks a solution from someone in a different team. The result is more than knowledge-sharing. It’s culture-building, capability-building and a genuine accelerant for innovation.
Because it’s not about strategy decks or training modules. It’s about behaviour.
It’s bite-sized. No heavy prep. Just show something you’ve tried.
Momentum. Adoption stops being theoretical. Every week, progress is visible, tangible, and celebrated.
Collective curiosity=cross-pollination. Different perspectives collide. A designer’s hack solves a strategist’s problem. A junior’s experiment creates an idea for the leadership team.
It sparks ideas fast. We often walk out with something we can use straight away, whether that’s for a client brief or an internal experiment.
It’s low-pressure but high-energy. You don’t need to be technical. We all agree to bring something to share and see where it takes the group.
It also shifts the culture. Sharing becomes normal. Curiosity is rewarded. AI isn’t something “the tech team” does, it’s something everyone should be experimenting with. Crucially, this meeting seems to make everyone want to.
In the spirit of show and tell, here’s a few examples from last week’s AI Show and Tell meeting:
AI Governance Blueprint: We shared progress on a framework for privacy, ethics, and responsible AI use across client and internal projects — plus early work on an agent for risk assessment and a tooling audit. The less glamorous side of AI, maybe, but the work that makes everything else possible.
Google Gemini experiments: One colleague turned their holiday photos into Japanese Ukiyo-e prints. Another transformed snapshots from an internal workshop into graphic-novel-style visuals. Playful, yes — but these outputs are already being picked up for brand storytelling and client social campaigns.
Sora 2 video generation: We looked at OpenAI’s latest model upgrade, which now supports much more viable, natural video generation from a prompt or source image. Alongside OpenAI’s new avatar-driven social media app, the conversation quickly shifted from “fun demo” to “what new creative formats does this unlock for brands?” and “what cultural implications does this have?”
We’ve been doing AI Show and Tell for about two years now. When we look back at the transcripts of the things we were sharing then versus now, the level of expertise we’ve built in that time is staggering. For example, where we used to share good prompts, we now show agentic workflows. By analysing our meeting transcripts in preparation for this article, we also noticed a few other crucial things.
Firstly, it eradicates fear. Being forced to show something, anything, without judgement means people learn faster because there’s a heap of psychological safety. Something we know to be foundational to innovation within organisations.
Secondly, people amaze each other. We’ve seen colleagues demo AI turning nightmare to-do lists into parody strategy meetings with the cast of Succession… and others using it to create pocket dermatologist bots.
Knowledge gaps close fast. One person shows how they solved a brand safety monitoring challenge; suddenly the whole team has that capability.
All of these happy side effects make innovation feel natural. It doesn’t need a steering group or a mandate. It just emerges, week by week.
Because of how successful this has been within our business, we always recommend this approach to clients struggling to systemise AI use. However, a common objection we hear is how difficult it would be to find an hour in the calendar where everyone is free every week.” Fair enough. The answer isn’t to not try though, it’s to split it.
Smaller groups. Run multiple Show & Tells and swap participants around each month.
Make teams multidisciplinary. Always mix roles and functions. Innovation happens in the overlap and not in siloes.
Keep it light. No polish, no pressure. Just demos.
Done this way, AI Show & Tell becomes infrastructure: the connective tissue for AI adoption, innovation, and growth.
The point isn’t the meeting itself. It’s what happens afterwards – the habits, the confidence, the ideas that multiply. AI Show & Tell is how adoption stops being an initiative and starts being second nature. The only real question is: what would your team show if you gave them five minutes?
Thanks for reading! Share with a friend or colleague who might find this interesting.

Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.