Why Most AI Projects Fail Before They Even Start

Dr. Niklas Richter · 2026-09-08

Digital transformation is on every company's agenda. Yet so many projects stall, miss deadlines, or simply don't deliver the results everyone hoped for. In a recent conversation, Virginie Claes sat down with Nick Allaert, co-founder of SANDT, to dig into why technology projects so often go sideways. They didn't talk about shiny tools or the latest AI breakthrough. Instead, they focused on the messy, human side of implementation. The stuff that actually determines whether your investment pays off or becomes another expensive lesson. ### The Execution Gap: Where Strategies Go to Die You've probably seen it. A beautiful strategy deck gets approved. Everyone nods. Six months later, nothing has changed. Nick points out that the problem usually isn't the strategy itself. It's the bridge between planning and doing. Companies underestimate how much change management is required. They treat technology as a plug-and-play solution when it's really a people project. One big culprit? The gap between business and IT teams. They often speak different languages, have different priorities, and rarely sit at the same table early enough. > "Choosing the right technology is almost always more important than choosing the newest technology." — Nick Allaert That quote stuck with me. Because in the rush to adopt AI, many businesses skip the boring but critical question: does this actually solve a real problem we have? ### Why AI Only Works When It's Woven In AI is everywhere right now. Every vendor claims their tool is AI-powered. But Nick makes a sharp distinction between AI hype and AI value. AI becomes valuable when it's integrated into daily workflows. Not when it's a standalone experiment that lives in a sandbox. If your team has to remember to use it, it's already failing. The real win happens when AI quietly handles repetitive tasks, surfaces insights, or speeds up decisions without anyone thinking about it. That's when you see return on investment. So how do you get there? Here's what stood out from the conversation: - Start with a specific business problem, not a technology looking for a home - Involve the people who will actually use the tool from day one - Expect resistance and plan for it, don't be surprised by it - Measure outcomes, not just adoption metrics - Be willing to kill projects that aren't working ### Change Management Isn't Optional You can have the best AI in the world. If your team doesn't embrace it, it's worthless. Nick shared how SANDT guides companies from strategic advice all the way through full implementation. That end-to-end approach matters because most failures happen in the middle. The strategy is fine. The technology is fine. But the handoff to everyday use is where things fall apart. Convincing employees to embrace change isn't about slick presentations. It's about showing them what's in it for them. Will this make their job easier? Less frustrating? Will it free them up for more interesting work? Answer those questions honestly, and you're halfway there. ### The Takeaway for Business Leaders If you're investing in digital innovation, don't get distracted by the hype cycle. The newest tool isn't always the right tool. The flashiest AI demo isn't always the one that solves your actual bottleneck. Focus on integration. Focus on your people. And focus on choosing technology that fits your business, not the other way around. That's how you turn a digital transformation project from a costly experiment into a genuine competitive advantage. And that's the kind of return worth chasing.