The cycle of tech disappointment
You know how it goes. A company gets enthusiastic about AI, machine learning, or whatever the latest tech revolution is. Everyone's excited, pilots pop up everywhere, and there's buzz in the air. Sound familiar?
Those pilots start multiplying across departments. Marketing is doing one thing, operations another, and IT something else entirely. Before you know it, you've got disconnected pilots.
The real pain hits in what I call the "stagnation phase" - that's when everyone realises these cool pilots can't scale. The pilots are all over the place, nobody knows who's responsible for what, resources are stretched thin, and systems don't talk to each other.
And the final blow? After pouring serious money into all these pilots, the big enterprise-wide impact everyone hoped for is nowhere to be seen. Confidence tanks, and executives start questioning if AI was all just hype.
It's not just disappointing – it's expensive.
Two paths: one-off vs. the COE journey
Why one-off projects keep letting us down
Let's be honest about what happens with the traditional project-by-project approach:
- Teams end up solving the same problems over and over because nobody's sharing solutions
- The knowledge gained in one project walks out the door with the team when they're done
- Success gets measured in these tiny bubbles that don't reflect real business impact
- When a cool pilot works, nobody knows how to scale it across the organisation
- Security and quality standards are all over the place
- The person who built something becomes irreplaceable (until they leave!)
- Teams are constantly playing catch-up with problems instead of preventing them
What makes centres of excellence different
A centre of excellence (COE) isn't just a name for a team - it's a completely different philosophy. Think of it as creating an internal powerhouse that becomes your engine for consistent, scalable innovation.
When you build a COE, you're saying: "We're going to be smart about this. We're going to build capabilities, not just projects."
This means:
- You create standards everyone follows, so solutions work together
- The lessons learned from every project get captured and shared across the organisation
- Teams can grab pre-built components from previous work instead of starting from scratch
- You measure success based on company-wide impact, not project-specific metrics
- There's a clear playbook for taking successful ideas from small pilots to full-scale deployment
- Governance and security are baked in from the start
- Expertise gets distributed across the organisation rather than sitting in silos
- You're constantly improving based on what you learn
The difference? One approach gives you a collection of projects. The other builds an organisational capability that gets stronger over time.
The real challenge isn't technology - it's how we organise around it
Here's what I've come to realise: The technology itself is rarely the biggest hurdle. The real challenge is building an organisation that can consistently identifies opportunities, implements solutions, and scales successes.
It's like the difference between learning a few recipes and becoming a chef. One gives you some meals; the other gives you the ability to create endless meals in any situation. Companies need to become AI chefs, not just follow AI recipes!
How we approach this at Advania
When I work with clients at Advania, we break this journey into two phases that make it manageable and practical.
First, we do the groundwork:
- We assess where you are today and where you want to be.
- We immerse ourselves in your business to understand your real challenges.
- We identify the use cases that will deliver real value (not just shiny toys)
- We evaluate your organisation's maturity for implementing AI.
- We create a vision that connects technology to actual business outcomes.
Then we build the COE structure around three key pillars:
- Platform enablement and support: This is where we make sure you have the technical foundation to build and run AI applications - everything from architecture to security to governance.
- Engagement and adoption: The best technology in the world is useless if people don't use it. We focus on change management, building communities, and making sure the value materialises.
- Build factory: This is your engine for consistently delivering solutions, using both professional development and low-code tools.
The beauty of this approach is that it's comprehensive without being overwhelming. You don't have to do everything at once - you can grow your capabilities over time.
Making the shift: practical steps you can take
If you're nodding along and thinking, "Yes, we need to move from project-based to capability-based thinking," here are some practical next steps:
- Start with an honest assessment: Where are you on the maturity curve? Are you still in the enthusiasm phase, or already hitting stagnation?
- Look for redundancy: Are multiple teams solving similar problems? That's a clear sign you need more coordination.
- Count your dependencies: How many critical systems or processes depend on specific individuals? That's a risk factor that COEs help mitigate.
- Review your scaling successes: How many pilots have successfully scaled to enterprise deployment? A low number suggests a structural problem, not a technology problem.
- Consider your governance: Is security, compliance, and quality consistent across initiatives, or does it vary dramatically? Inconsistency points to the need for standardisation.
Let's chat about your journey
I'd love to hear where you are in your AI implementation journey. Are you seeing signs of the fragmentation or stagnation phases? Or you have already started building COE-like capabilities and have insights to share?
The path from one-off projects to a centre of excellence isn't always easy, but it's one of the most valuable transitions an organisation can make.
Coffee's on me if you want to discuss how this might work in your specific company!