It’s the leadership stupid

Saturday morning and I am reading an HBR article and can’t help but thinking aloud “No Shit Sherlock”. The article highlights what companies successful with AI do differently and to no surprise it all boils down to a handful capabilities.

The authors describe these behaviors by using the SHAPE Index:
- Strategic Agility (Pivot early in the game)
- Human Centricity (Design change with people, not for them)
- Applied Curiosity (Demonstrate personal engagement in experimentation)
- Performance Driven (Have clear metrics and business outcomes)
- Ethical Stewardship (Embed fairness, traceability, and accountability)

They continue: “Successful organizations drive AI development strategically. Rather than hoping the right leaders emerge organically, they assess current capabilities, identify gaps, and create targeted development plans. The result is that leadership becomes an accelerant to AI transformation rather than a bottleneck”.

Let’s pause for a moment… isn’t this what we have been saying all along? Don’t leave AI to the technologists in your company but embed it enterprise-wide and top down. Think about adoption, use it to create customer value, solve everyday problems, watch out that you don’t get stuck in the experimentation playground and establish governance by design.

And the worse part, we have been there before. I could have told you 🤦🏻‍♂️ that you have a 50-75% chance of complete failure. What happened with ERP and CRM is now also happening with AI. This is not me saying this but the statistics. ERP implementations have a failure rate ranging from 55-75%, with some research showing rates can exceed 75%. Failure rates for CRM implementations can reach as high as 63-70% and around 50% of these fail within 2-3 years. MIT reports that 95% of generative AI pilot programs fail to reach meaningful production while
42% of companies abandoned most of their AI initiatives in 2025, up from just 17% in 2024. Research estimates over 80% of AI projects fail—twice the rate of non-AI technology projects.

Now the good thing is, that is if you are willing to pay your dues of course, you can still course correct your AI program. For sure there will be sunk costs and a dent in your credibility but the French having a saying for this “reculer pour mieux sauter” - think twice and start all over, go back to the drawing board and start writing the new playbook. You have are now at the through of Disillusionment… time for Enlightenment 🚀 - feel free to reach out if you are in need of an AI war room.