Thought Piece: How leaders avoid AI failure

Hi there,

The statistics are sobering: 70% of AI transformations fail to deliver promised results. After working with hundreds of organizations, I've identified the critical factor that separates success from expensive disappointment. But before getting deeper...

If any solution usually brings a new problem, 100s of solutions brought a gigantic problem: making it all actually work together. Most larger organizations have a "One" initiative, often with a multi-million budget. One Microsoft, One Google, One CBRE, One Ford, One P&G, One Sony, One Cisco, One JLL, One Boeing, One HSBC, One GE… and they've had those for years! But it's been so hard to actually connect the blocks together.

How can this be solved?

This is the leverage—and the power—of next-gen digital transformation (NDX). Rapid advances with AI mean two truths reached maturity:

• AI systems can work around how our people work, so change management can be 10x quicker and 10x cheaper

• AI can accelerate coding and maintenance of new systems, In many cases what used to take 100 days to build, can take just 1 day now

Building custom solutions can unlock 100x more value than complex SaaS landscapes. Think of it this way: Traditional transformation is like learning a new language to communicate with technology. Next-gen transformation is like having technology learn your language.

Next-gen transformation doesn't just identify opportunities; it positions your business to seize them automatically, with AI and Machine Learning driving the charge.

This intelligent layer anticipates problems, adapts in real-time, and learns from every interaction. The result is a shift from static integration to dynamic, self-optimizing systems that grow alongside your business.

Next-gen digital transformation has moved beyond adopting and writing software. At its core, it is a strategy that combines AI, automation, and reimagined workflows to deliver better outcomes, solve real business problems, and rapidly accelerate change management. These systems redefine how work gets done, creating smarter workflows that continuously adapt, improve, and drive lasting impact.

The key differentiator is speed to value. Instead of five-year transformation roadmaps, we're looking at continuous cycles of innovation that deliver value in weeks—not years—and let the compounding effect take care of the exponential results. In short: Your technology should adapt to your people, and help them grow —not the other way around.

And if that's not enough, currently every single year AI becomes 10x cheaper (computing power) and 10x more capable (Data and model capability) than the year before. These double exponentials are why progress in AI feels like it's accelerating beyond normal tech cycles—it's not just faster; it's compounding in multiple dimensions at once. 10x has been a tenet of thinking at most successful companies of the last 2 decades. We are now in the 100x era.

I know what you're thinking: Easier said than done. Exactly the reason I wrote this book. It will explore the concept in much more detail and help you start acting now.

But first…

How Not to Start

Want to know the fastest way to fail? Start with technology. I've watched executives spend millions on digital transformation and AI initiatives that looked impressive in boardroom presentations but delivered zero value to the business. They focused on the shiny tools instead of real transformation.

Another dead end? Creating an innovation department and expecting magic. While a dedicated team can be an excellent driver of change, without a larger mandate and increasing involvement from the entire organization, their efforts can't reach their potential. For the past 3 decades traditional digital transformation went roughly this way:

“Everyone's going digital, we must do the same! Oh I see, fine let's. And these Salesforce guys say they can make our sales on the cloud! Splendid. Let's spend 3 years training our teams in the hopes most will become proficient… But wait! This didn't solve everything! We also need this new thing called SAP, and Microsoft says they give us 3 months free if we lock in with them for a few decades.”

I exaggerate but does this sound familiar? Most bigger companies currently have at least 100 different systems that became core to their operations and don't talk to each other. In 2024, the average company used 150-200 SaaS apps [Vena, Better Cloud research]. I've seen application landscape maps with over 500 of them! They all have different interfaces, ways of working, pros and cons and are preferred by different team mates. Roll out of each one took at best a year. Roll out of all? Years…

How to Start

Start with people. Ideally the people you already have. Begin with what Jeff Bezos calls "stubborn vision, flexible execution." Your vision of a future-proof organization should be unwavering, but your path to get there needs to be fluid. Time and time again, this is the biggest predictor of success. Further, we'll explore culture, labor models, incentives, and technology stack fit for truly embracing the Age of AI.

Some of this might read a little like the kind of stuff you'd hear from a consultant—but here's the difference: we're not consultants. We're implementers. We've been doing this for well over a decade, with specific NDX for the last 2 years (it didn't exist before that!). All recommendations here come from the battlefields and are not material for swanky decks and polished presentations—but for real, lasting change.

From both our experience and all research, here's the single most important thing you need to get right: Start by ensuring there's a leader to champion the transformation. This may be you. Without such an executive sponsor, the chances of success aren't low—they're zero-point-zero. With the right leader, everything else falls into place. Studies show that the leader is almost always better selected from within the existing organization, than hired or acqui-hired. A CEO, COO, CTO, CMO—someone already on the inside who's eager to lead change—is the most powerful unlock.

Core quality of that leader? Tenacious enough to start and to not stop.

Much Love,
Matt

At Lighthouse, we love featuring fresh perspectives from our community of AI, tech, and innovation leaders. Got insights to share? Just reply to this email—I’d love to hear from you!