Why "The Experts" Don't Have the Answers Yet

Here's what's different about AI in 2024 compared to software in 2004: the playing field is level.

When VW decided to outsource software 20 years ago, the technology was mature. Software engineering had been a profession for decades. System integrators like Accenture had deep benches of developers, established methodologies, and proven delivery models. At least, that's how it appeared.

We now know most system integrators relied heavily on waterfall approaches that were already becoming obsolete, favoring rigid contract structures over adaptive development. Even with software, outsourcing was the wrong choice. But at the time, companies could at least justify the decision: the industry was mature, the integrators had decades of apparent expertise, and the playbooks seemed proven.

AI doesn't have that cover story. The consultants selling "AI transformation" today don't even have the illusion of deep experience. ChatGPT launched in November 2022. GPT-4 in March 2023. Claude, Gemini, and other models followed shortly after. Agentic AI frameworks are even newer. The tools, techniques, and best practices are still being figured out.

The system integrators selling "AI transformation" today? They're figuring it out alongside you. They might have hired some machine learning engineers. They might have run a few pilots. But they don't have decades of institutional knowledge. They don't have proprietary methodologies that actually work. They're experimenting - and charging you for the learning process.

When consultants propose AI transformation programs, they're not lying about their experience. They have built some agents. They have run some pilots. They are ahead of most organizations. But "ahead" means months, not years, and definitely not decades. And those months of experience don't translate into lasting competitive advantage for their.

Here's what organizational capability looks like: When your team encounters a new AI tool, they can evaluate it. When a process breaks, they can fix it. When an opportunity emerges, they can act on it. When the technology evolves - and it will evolve rapidly - they can adapt.

This capability cannot be transferred through documentation, playbooks, or train-the-trainer sessions. It requires doing the work, making mistakes, learning what works in your specific context with your specific challenges.

The consulting model optimizes for the wrong outcome. Consultants solve your current AI problems. What you need is the capability to solve your future AI problems.

The difference shows up in how organizations respond to change. When GPT-5 launches, or when new agent frameworks emerge, or when entirely new AI capabilities become available, what happens?

Organizations that built capability internally can evaluate and adopt rapidly. Organizations that bought solutions wait for their consulting partner to update their offerings, then wait for the next engagement, then wait for implementation. That time lag compounds. Fast-moving markets don't wait for procurement cycles and statement-of-work negotiations.

This creates a rare opportunity. Right now, you can build AI capability without playing catch-up. You're not 10 years behind. You're maybe 6 months behind. And that gap closes fast if you start now.

But the window won't stay open. In 18-24 months, the companies that invested in AI literacy will have a structural advantage. They'll have employees who use AI daily. They'll have processes that integrate AI into decision-making. They'll have a culture of experimentation and learning.

And the companies that outsourced? They'll be stuck in vendor relationships, dependent on external partners, and scrambling to rebuild capabilities they should have developed from the start.

Here's the fundamental shift AI creates: the barrier to building software just collapsed. For decades, software required specialized technical knowledge - programming languages, databases, infrastructure. That technical barrier made outsourcing feel rational. You needed people who could code.

AI changes the equation. Natural language is the programming language. English, German, Spanish - whatever language you speak is now how you build. This means your domain experts - the people who understand your customers, your operations, how value gets created - can now build solutions directly. It's easier to teach your supply chain manager AI literacy than to teach a consultant your supply chain. It's faster to train your customer service team on AI agents than to explain your customer problems to an outside firm.

The competitive advantage isn't understanding the technology. It's understanding your business deeply and having the AI literacy to act on that understanding. This is the strategic reversal: for the first time in decades, domain knowledge matters more than technical knowledge. Your institutional expertise about customers, operations, and value creation - the knowledge that lives inside your organization - becomes the differentiator.

Companies that build this capability internally don't just avoid dependency. They create a compounding advantage. Every employee who gains AI literacy can spot opportunities faster, solve problems more effectively, and iterate on solutions independently. That capability multiplies across the organization. It becomes muscle memory. It's how you think, how you work, how you compete.

Companies that outsource this? They rent capability that never becomes theirs. They pay for solutions that don't build internal strength. And five years from now, when AI has evolved three more generations, they'll still be waiting for their consultants to catch up while competitors who built capability are already adapting.

This isn't just about avoiding the mistakes of the past. It's about seizing an advantage that won't be available again.

Don't fall for it. This is your chance to build capability while the playing field is still level.