Today’s businesses are racing to adopt AI. McKinsey’s recent report on AI in the workplace makes the situation very clear: 92% of companies plan to increase their investments in AI over the next three years, but just 1% of business leaders say their company has achieved a “mature” deployment of AI.
Leaders know what they want from AI - they just don’t know how to get there.
Adopting AI and succeeding with AI are two very, very different things. The commonly held wisdom is that 70% of business transformations fail, and AI represents one of the most high-stakes transformations in the history of technology. Most will adopt AI tools that ultimately don’t deliver value.
The pressure is even more intense for AI because of its role as a business accelerator. Companies that adopt new AI tools aren’t just changing direction: integrating AI means they’re able to run in that new direction further and faster than ever before. The successes and failures of the AI era will be magnified significantly.
But that doesn’t mean business leaders should be hesitant or pessimistic - standing still is another guaranteed path to failure. While the majority of transformations fail, organizations can take steps to improve their odds for success.
1. Change management and bureaucracy
The elephant in the room on that business transformation statistic is that most of those transformations don’t fail because of technology - they fail because of people. Changing your technology stack can make a difference, but no company can overcome a bureaucracy problem by buying new software.
Change management requires leadership. For companies to succeed with AI, they need an executive to lay out a clear vision - what these tools are going to do, who they’re going to help, how and why. If everyone’s aligned on goals, tools, and data, the odds of success will increase dramatically. Likewise if different business units are pulling in different directions or if employees are experimenting with a fragmented set of tools, the transformation becomes much more challenging.
In today’s AI market, laying out that vision isn’t easy. There are dozens, if not hundreds, of vendors competing to bring new AI products to market, and there’s little clarity on which solutions will become the eventual winners. Experimentation is unavoidable to some extent, but leaders should be ready to make decisions quickly - moving on from tools that don’t provide value and doubling down on those that do.
2. Defining and measuring success
While businesses are still in the experimental phase with AI, different departments will naturally go down different paths in terms of tooling and goals. That’s not necessarily a bad thing: a tool that works for marketing may not work for sales, and a tool that helps the product team may be completely irrelevant for the go-to-market organization. As long as everyone is making decisions quickly and aligning on a narrow set of tools, the company as a whole doesn’t have much to worry about.
However, businesses can go astray if they haven’t decided how to define and measure success. Who makes the ultimate decision on which tools make it into the budget and which don’t? What does that decision-maker prioritize? If one business area isn’t seeing value, the organization may give up too early on a tool that’s proving useful in other areas.
In this situation, businesses will benefit from clear priorities and open lines of communication. Give teams the opportunity to advocate for the solutions that are delivering measurable results, then allocate your resources accordingly.
3. Siloed, unintelligent data layer
When they’re set up properly and given the right information, a lot of these new AI tools can feel like magic. But that magic doesn’t just emerge from nothing. For an AI tool to deliver precise, valuable outputs for your business, it needs a deep, comprehensive understanding of your business.
This is where AI breaks - or breaks through. If a company has expansive, clean, well organized data, that data will be able to flow freely throughout the company’s AI systems. The better your inputs, the better your outputs, and the faster you’ll be able to run in new, AI-powered directions. But if your data is siloed, messy, or incomplete, there’s not much the AI tools can do to overcome the problem.
If you want to build something new, the first thing you need is a clean, strong foundation. For today’s businesses - especially those in customer-facing industries - that foundation is your data layer. If you ignore it, you’ll be setting yourself up for diminished returns and disappointing results.
AI will make business faster, smarter, more efficient - but getting there won’t be easy. The work starts now: untangle the bureaucracy, align on outcomes, and clean up your data. Do that, and you’ll be ready to ride the next wave - not get swept under it.