Forrester analyst Kate Leggett used a phrase that has stuck with many of us: the idea that this year won't be defined by dazzling, headline-grabbing transformation, but by gritty, foundational work. This unglamorous, behind-the-scenes effort truly drives progress. We think she's right, and frankly, it's a relief to hear someone articulate it so clearly. The relentless hype cycle has made everyone feel like they're falling behind if they aren't already levitating on a cloud of AI-powered innovation.
The reality, however, is far more ordinary and practical. The businesses getting real, tangible value from AI are not the ones with the flashiest, most complex tools. They identify a painful, specific, and measurable problem (like a bottleneck in their supply chain or a gap in their customer support) and point a targeted AI solution directly at it.
We see this pattern constantly in customer service, which is the corner of the world we know best. The business owner, who vaguely proclaims, "I want AI to transform my business," usually ends up with a digital junk drawer full of half-used subscriptions and little to show for it. Conversely, the owner who says, "I am losing too many customer calls during the lunch rush, and I want that one specific thing fixed," tends to actually solve the problem. More importantly, they can tell you how much that fix improved their bottom line. The narrowness of the goal is the entire point. A problem you can name is a problem you can measure, and a problem you can measure is a problem you can prove you solved. This distinction is the critical difference between the AI projects that quietly work and the far larger number that quietly fade away.
We won't pretend every deployment lands the same way. Some of the businesses we've worked with saw incredible gains almost immediately. Others took longer to find their footing. In nearly every case, the variable that determined success wasn't the sophistication of the technology; it was the clarity of the goal going in. We've now made it a practice to ask people, before we discuss any solution, what number on their dashboard they hope to change. The ones who can answer that question crisply and without hesitation are (almost) always glad they bothered. The ones who can't, who are reaching for "AI" the way you might reach for a lucky charm in a casino, usually drift away disappointed, regardless of how powerful the tool is.
So, where does this leave us? Our guess is that the current wave of disillusionment is both temporary and profoundly useful, it's the natural hangover after a few years of treating AI like a magic wand. The hype promised that simply having AI would change everything, and that turned out to be false in the most expensive way possible for many. What we believe will replace this is something slower, more deliberate, and more grown-up. It's a return to the boring but essential questions that have always separated good operators from hopeful ones: What exactly is broken? How will we know when it's fixed? And is this tool we're considering aimed squarely at that problem, or is it just aimed at the general direction of the future?
If there's a comfort in any of this it's that the small business actually has a significant advantage here that the giant enterprise often lacks. You don't have a thousand competing internal initiatives, a labyrinthine, year-long procurement process, or a data swamp built over two decades of shifting priorities. You have a handful of core problems you could probably list on a cocktail napkin. You can point a tool at one of them this week and know by next month, with real data, whether it worked. That ability to be nimble and focused isn't a small thing. In this new era, it might be the whole game.
We'll admit we find this period strangely hopeful, even though it's often framed as a failure story. The dismal failure rates in the headlines sound like a definitive verdict on AI, but read them again and they're really a verdict on how people went about implementing it. That is something any of us can change tomorrow without waiting for a better algorithm or a more advanced model. The question worth sitting with isn't whether AI works, because in the right spot, it plainly does. The real question is whether we're willing to be specific about what we actually want from it. Specificity is a discipline, not a gift, and it's the one we've built our whole practice around: it's the first question we ask a new customer, and it's the reason our flows are deterministic instead of hopeful. From experience, the businesses that name the number first are the ones that get to keep it.
