There’s a specific kind of frustration making the rounds in leadership circles right now. It sounds something like: “We gave everyone access to AI tools six months ago and I’m not seeing the returns. What are we doing wrong?”
It’s a fair question. It’s just not quite the right one.
The Metric Is the Problem
A lot of organizations measured their early AI efforts the way you’d measure a marketing campaign: by adoption. How many employees signed up? How many logged in last week? How many teams are “using AI”?
Those numbers feel like progress. They aren’t, at least not on their own. Counting AI logins as a measure of AI value is a bit like measuring your website’s impact by counting homepage visits. It tells you people showed up. It tells you nothing about whether showing up did anything for your business.
The leaders who are frustrated with their AI returns often invested in tools first and strategy second. They encouraged adoption without asking what, specifically, they wanted AI to do differently in each part of their value stream. AI was treated as an initiative rather than a capability, something to be deployed rather than something to be woven into the fabric of how the organization delivers value. When you measure “AI” as a standalone investment with its own profit and loss, you’ve already framed the problem incorrectly. AI isn’t separate from your products and services. It’s another way of delivering them.
We’ve Been Here Before
Here’s the thing: we’ve seen this pattern play out before, more than once.
When the internet arrived, most organizations didn’t know what to do with it either. The first wave of business websites were essentially digital business cards: the company name, a phone number, maybe a paragraph about who you were and where to find your actual store. We knew something big had arrived. We just didn’t know how to think about it yet.
It took years — in many cases, a decade or more — before organizations began to understand how the internet could be genuinely woven into their value streams. And longer still before the real innovators figured out that it didn’t just change how they delivered existing products and services, it unlocked entirely new ones. That kind of transformative thinking is rare. It doesn’t happen in a single planning cycle, and our brains aren’t really built for it. We tend to map new tools onto old mental models until someone, somewhere, imagines something we hadn’t thought to imagine yet.
AI is at that same early, awkward stage. Most organizations are still building digital business cards.
Where We Are on the Map
If the ROI grumbling feels familiar, there’s a reason. Gartner’s Hype Cycle is a framework that describes the predictable emotional and strategic journey organizations take with almost every major technology innovation. It maps five stages: a technology trigger that captures everyone’s attention, a peak of inflated expectations where possibilities feel limitless and investment floods in, a trough of disillusionment where early implementations disappoint and hard questions get asked, a slope of enlightenment where thoughtful integration starts to pay off, and finally a plateau of productivity where the technology becomes simply part of how good organizations operate.
Gartner Research’s Hype Cycle diagram, Jeremykemp at English Wikipedia
That trough is where a lot of organizations are sitting right now with AI. And while it’s uncomfortable, it’s also clarifying. The hype has burned off. The leaders who bought AI tools because they wanted to look innovative are starting to feel the gap between expectation and reality. That’s not a failure. That’s the moment when serious strategy becomes possible.
After the trough comes the slope of enlightenment, where organizations that did the harder work of integrating the technology thoughtfully start to see real returns. Then the plateau of productivity, where the technology becomes simply part of how good organizations operate. The .com bust wiped out a lot of flashy websites. It didn’t touch Amazon.
But AI Is Moving Faster Than the Internet Did
Here’s where the parallel breaks down a little: the internet changed business over the course of years and decades. AI is changing it in months. Models that were state of the art eighteen months ago are already obsolete. The tools organizations standardized on last year have been updated, repriced, or replaced by something newer and more capable. Strategies built around specific AI capabilities can become outdated before they’re fully implemented.
That pace creates a specific kind of workplace anxiety that leaders need to take seriously. It’s not just that the technology is new. It’s that it keeps being new, over and over, faster than most organizations can adapt. Employees who finally got comfortable with one tool find themselves back at the beginning with another. Leaders who felt like they understood the landscape discover it has shifted again.
This isn’t a reason to stop moving. It’s a reason to build organizational agility into your AI strategy from the start, so that when the tools change, and they will, your team knows how to adapt rather than waiting for new instructions from the top.
The Right Question
The leaders who will get the most out of AI aren’t the ones asking “what is our return on our AI investment?” They’re asking something harder and more useful: how does AI touch each part of our value stream, and are we making intentional decisions about each of those touchpoints?
If your organization is working through these questions and could use a thought partner, Engaged Agility consults on business and product challenges of all shapes and sizes. Reach out to get started.