By Sarah Nicastro, Founder and Editor in Chief, Future of Field Service

Some recent statistics paint the picture that companies are struggling to bring AI’s potential to life in tangible ways. MIT reported that 95% of GenAI pilot programs do not show a measurable impact on a company's P&L statement. And according to Gartner, over 40% of agentic AI projects will be canceled by the end of 2027 due to escalating costs, unclear business value, or inadequate risk controls.

Perhaps these stats, at least in part, aren’t representative of any inherent flaws in AI, but rather the outlandish expectations and lackluster effort of the companies investing in it. We’ve all heard the phrase “technology is just a tool,” but there seems to be an even stronger tendency with AI to believe that the tool alone will solve problems and drive value.

Logically, we (should) all know that’s not true. But the stats above lead me to believe that companies are holding fast to hope that AI is magical when what’s needed is a more pragmatic approach. Here are some aspects that come to mind:

• Clarifying the objective. I believe many companies struggling to see ROI from AI made the investment without clarity on why exactly they were doing so, or more specifically what business problem(s) it should solve. Perhaps companies rush because they feel pressure to keep pace with the technology that’s trending, and this is fair, but AI can’t achieve results that haven’t been defined. Being selective and strategic about where AI is best suited for use clarifies the pain point you’re aiming to solve, which increases chances of success and, in turn, improves the likelihood of further investment.  
• Doing the foundational work first. Another thing I see happening, quite frankly, is companies that have done a poor job of implementing foundational technology layering AI on top and hoping it fixes everything. Newsflash: this won’t work. In fact, it will simply compound the technical debt you already have. AI holds true to the same old principle: garbage in, garbage out – whether it’s data, processes, or a combination. There’s no shortcut to the hard work of examining the business needs, processes, data, and existing systems and doing whatever foundational work needs done.
• Leading through change. Change management has been a crucial aspect of digital transformation since digital transformation began. But never has it been more imperative than in the AI era. Resistance to change is human nature, but AI causes a degree of anxiety that earlier generations of technology didn’t because it makes employees fear for their jobs. Furthermore, today’s talent has evolved expectations of company culture and employee experience. This means that the days of “do as you’re told,” while never particularly effective, are over. You simply must communicate early and often, explain the why, be transparent about what you don’t know, get employee feedback early and throughout the process, offer ample and effective training, and reward not only adoption but effort.
• Considering how to future-proof. One of the elements that makes AI truly exciting is the potential it holds to fundamentally change how businesses (the world, really) work. This means there’s a lot to think about, even as you’re climbing the initial mountain of working toward AI ROI. How will AI change your workforce? How will it transform your customer interactions? What elements of accuracy, security, and ethics are paramount for your business to consider now, and in the future? There’s a real responsibility here for companies to take a forward look, even while mastering today’s use cases.
• Create a culture conducive to continual innovation. The pace of change we live in today is truly something else. Gone are the days of investing in a new system, going live, and then maintaining it for a few years before it was time for an upgrade. Today, technology is evolving at lightspeed, but so are customer expectations, the talent landscape, economical and geopolitical conditions. As such, companies who have yet to break down siloes must do so. It’s essential to have the ability to analyze, discuss, decide, and act on business insights in an agile and effective manner.

As I write this, it strikes me how much of this same list could have been written about service management circa 2005 or so. And in many ways, this is the same story, but with a new character. This is because it’s never been the technology that was the “hard” work – it’s all the people and process effort that goes into making any technology work the way it was intended. The difference with AI is that the stakes are even higher. The trick, I believe, is to avoid letting that reality make you feel pressured and instead let it fuel your mission to get it right.