Implementing AI? The Technology Is the Easy Part.

Almost every owner and executive I speak with across Durham Region and the Greater Toronto Area is feeling the same pressure right now: everyone says you need to implement AI, and no one quite explains how to make it actually work. So they buy a tool, run a pilot, and a few months later it has quietly fizzled out, used by almost no one. If that sounds familiar, you are not behind, and you are not alone. You are running into the real challenge of AI, which is not the technology at all.

I want to be straight with you about my lane here. I have spent more than 35 years in senior executive roles at organizations including Honeywell, Johnson Controls, and Armstrong Fluid Technology, and I am not going to tell you which AI platform to buy or how to build a model. That is work for technical specialists. What I can tell you, because it is where most AI efforts genuinely succeed or fail, is how to lead the change that makes AI stick.

Why Do So Many AI Projects Fail?

The failure rate is striking, and the reasons are revealing. Gartner reports that by the end of 2025, at least 50% of generative AI projects were abandoned after proof of concept, due to poor data quality, inadequate risk controls, escalating costs, or unclear business value. Notice what is missing from that list: the technology working. As Gartner itself puts it, the biggest obstacle is usually not the technology, but how organizations approach implementation.

Every one of those failure points is a leadership decision, not a technical one. Unclear business value means no one defined the problem before buying the tool. Escalating costs and weak controls mean the effort was not governed well. And even when the technology works perfectly, it delivers nothing if your people do not adopt it. Gallup's 2025 research found that only around 22% of employees say their company has shared a clear plan or strategy for AI, which tells you most organizations are rolling out tools without leading the change around them. That is the gap you have to close.

How to Implement AI: Lead It Like the Change It Is

Implementing AI well follows the same discipline as any major change, and that discipline is a leadership job. Here is the approach I would use, aimed at the people side that determines whether the technology ever pays off.

1. Start with a real business problem, not the tool. Before you buy anything, name the specific problem you are trying to solve and what success would look like. Unclear business value is one of the top reasons AI projects are abandoned, so define the value first and let it drive the technology choice, not the other way around.
2. Communicate a clear plan. Your people cannot get behind what they do not understand. Gallup found that when leadership clearly communicates its AI plan, employees are three times as likely to feel prepared to use it, yet only about 22% of companies have shared such a plan. Be the exception: explain the why, the what, and what it means for people's jobs.
3. Get your managers genuinely behind it. Adoption is driven by managers, not mandates. Gallup found that employees whose managers actively support AI use are far more likely to use it regularly. Equip your frontline and middle managers to model the tools, answer questions, and champion the change day to day.
4. Address the fear honestly. Many people quietly assume AI is there to replace them, and that fear kills adoption. Be direct about how AI is meant to support your people rather than replace them, and involve them in shaping how it is used, so it feels like something done with them, not to them.
5. Build the skills to use it. A tool no one knows how to use sits idle. Invest in practical training so people can apply AI to their actual work, and give them time and permission to learn rather than expecting adoption to happen on its own.
6. Govern it responsibly. Set clear guidelines for how AI can and cannot be used, protect your data, and put sensible risk controls in place. Weak governance is a leading cause of stalled projects, and clear rules give people the confidence to actually use the tools.
7. Measure real value, and adjust. Track whether the AI is moving the business problem you started with, be honest about what is and is not working, and iterate. Treat it as an ongoing change to lead, not a box to check once.

What Should I Be Honest With Myself About?

Let me be straight about the limits, in both directions, because AI is surrounded by enough hype already. Leadership is what makes AI adoption succeed, but leadership is not a substitute for the technical foundations. You still need the right tools, clean and well-managed data, and genuine technical expertise to build and run them, and no amount of change leadership replaces that work. This is a both-and, not an either-or: strong technology with no leadership fails on adoption, and strong leadership with a weak technical foundation fails on capability.

Two more honest points. First, AI will not fix a broken process; it will amplify it. If a workflow is a mess, automating it just produces the mess faster, so fix the process first, which is exactly the kind of operational discipline good leaders bring. Second, leading this well requires a skill many leaders have not built. DDI's Global Leadership Forecast 2025 found that managing change is one of leaders' two greatest skill gaps, which is precisely why implementing something as disruptive as AI so often stalls. Developing your leaders' ability to guide change is one of the highest-leverage investments you can make before, not after, your next rollout. If you want help building leaders who can carry a change like this, the practical next step is a straightforward conversation. It also helps to have strong leaders in place to begin with, which is the focus of our guide on how to develop better leaders. You can also learn more about Malcolm Gurley and why organizations choose Gurley Leadership Solutions.