By now, most companies have AI initiatives. And by “initiatives” we mean a money pit with a dashboard. The AI initiatives we have seen often exhibit more activity than achievement. The problem? We’re treating AI like it’s 1995 and we’re rolling out a new ERP system. Spoiler: it’s not working. To often, they are set-up like  expert-driven task forces, disconnected from the business, aiming for “smart” outcomes instead of real impact. It’s as if the lessons from decades of innovation and change management vanished when the AI buzz arrived.

Let’s talk about why and how to fix it.

The Problem: The AI Theatre

Most AI projects are a masterclass in how to burn cash without creating value. Here’s our top six:

Tech-first, use-case-last: You’re building a Ferrari engine when what the business needs is a bicycle. Most often, data quality is not at a level where you can build meaningfull models.

• Too much money, too little business value: You’ve got a team of PhDs fine-tuning models, but no one can explain how it helps the customer or the bottom line.
• IT-led, not business-led: The people driving the project don’t own a P&L. They own a Jira board.
• Staffed by specialists, not teams: It’s all data scientists and zero domain experts. That’s like building a restaurant with only chefs and no customers.
• Change management? What’s that?: You drop a shiny new AI tool into the business and wonder why no one uses it. Because no one was brought along for the ride.
• Big bang launches instead of small bets: AI is exploratory. You don’t know what’ll work until you test it. So why are you betting the farm on a single moonshot?

This isn’t innovation. It’s theatre. And the audience, your shareholders, your customers, your employees, are shaking their heads.

The Fix: AI Ideation Campaigns

Here’s the thing: AI isn’t a tech problem. It’s a business opportunity. And the best way to unlock it? Run an AI ideation campaign.

Think of it as a company-wide hackathon with a purpose. Not just to build cool stuff, but to solve real problems.

Here’s how it works:

1. Ideate

Open the floodgates. Ask everyone, yes, everyone, how AI could improve their work. Yes, You’ll get gold, garbage, and everything in between. That’s the point. You’re not just collecting ideas; you’re building awareness, transparency, and momentum. Asking everyone also floss out smart people from all corners of the business, who might have high levels of capabilities needed for an idea to provide business impact.

Bonus: Tap into AI startups or your prefered tech partner with AI capabilities. They’re hungry, fast, and often very smart. Add them to the mix.

2. Shape

Take the best ideas and build cross-functional teams: business, IT, AI. No silos. No egos. Just people who understand the problem, the tech, and the customer.

Each team answers the big questions: What’s the problem? What’s the solution? Who benefits? What’s the business model? Can we actually build this?

Bonus: add one external AI specialist to the team. Someone that have build models before. Who know’s the need for data quality and can prepare data for the coming model work.

3. Validate

Now test it. Fast. Build a prototype. Run experiments. Kill what doesn’t work. Double down on what does. Meet monthly with sponsors who can greenlight or kill the project. No endless steering committees. Just action and short learning loops. 

Why It Works

Because it flips the script. Instead of AI being a side hustle for the IT department, it becomes a core business capability. You get:

• Real business problems, not tech demos.
• Mixed teams that actually talk to each other.
• Fast feedback loops instead of slow death spirals.
• A culture of experimentation, not perfection.

And maybe most importantly: you find the people in your organisation who get it. The ones who can lead your AI future, not just talk about it. This way, the business owns the solution. The change management issue is half way solved.

Bottom Line

If your AI project is all activity and no achievement, it’s time to stop the madness. Start small. Start smart. Start with the business.

Run an AI ideation campaign. You’ll be surprised where the best ideas and the best people come from.