Why AI Activity Isn’t the Same as AI Progress
Walk into almost any learning organization right now, and you’ll find AI everywhere. A pilot here, draft guidelines there, a working group that meets monthly, a slide about the future of learning in every deck.
Then ask a harder question: what has actually changed about how learning gets built and delivered? Too often, the honest answer is not much, yet. The pilot is a year old. The guidelines are still in draft. The working group is still working.
Activity and progress are easy to confuse, and right now much of the industry is confusing them. An organization can look busy with AI, even impressive with AI, and still be standing exactly where it stood a year ago.
Activity and progress are easy to confuse, and right now much of the industry is confusing them.”
What Is AI Maturity?
AI maturity isn’t a measure of how many people use AI. It’s an organization’s ability to adopt AI beyond a few early enthusiasts, scale what works across teams, govern its use, and keep learning as the technology changes.
Those capabilities depend on more than the technology. AI accelerates change, but people drive performance. A mature Human+AI organization knows where AI extends human capability, where judgment and expertise still lead, and how to build both into the way work gets done.
Our Five Stages of AI Maturity
Matt Donovan, Chief Learning and Innovation Officer at GP Strategies®, maps AI adoption in learning teams as a journey through five stages, and most people can locate their own team within seconds.
Ad Hoc | A few individuals experimenting quietly on their own.
Exploratory | Colleagues starting to compare notes and swap prompts.
Structured | The first shared processes, usage guidelines, and deliberate tool choices.
Integrated | AI woven into everyday workflows, beyond L&D and into HR, IT, and compliance.
Transformational | The stage nearly everyone can picture but few have reached, where AI becomes a catalyst for reinventing what learning does.
Donovan unpacks the full model in Chief Learning Officer. He also notes where most learning teams sit today: in Exploratory, or just entering Structured. Which sets up the question the map alone can’t answer: when did your team last move?
A Stage Is a Position, Not a Trajectory
Maturity models are useful, and quietly dangerous. Useful because they give leaders a shared language. Dangerous because the label starts to feel like momentum: find your stage, name the next one, put the journey on a slide.
But a stage is a snapshot. It describes your position, not your motion.
Two organizations can occupy the same stage with opposite futures. One is weeks away from its next transition. The other stalled a while ago and hasn’t noticed, because the activity looks the same either way. The question that separates them isn’t “where are we?” It’s “what’s slowing us down?”
Four Transitions, Four Kinds of Drag
Every transition between stages has its own characteristic drag: the specific friction that keeps a team circling inside a stage instead of leaving it. Which of these describes your organization?
From Ad Hoc to Exploratory: Invisibility
In Ad Hoc, people are experimenting, but nobody can see it. Wins stay private, prompts live in personal notebooks, and the same discovery gets made five separate times in five corners of the organization. When someone leaves, their learning leaves with them.
Getting to Exploratory doesn’t take a program. It takes visibility: early wins circulating widely enough that isolated experiments turn into shared practice. A quick test: could someone on one team find out what another team tried last month? If not, all that experimenting is evaporating rather than compounding.
From Exploratory to Structured: Gatekeeping
This is where momentum most often falters, because the instinct is to answer risk with restriction: approval queues, locked-down tools, review processes that mostly say no. Governance built that way doesn’t remove risk. It hides it, because people keep using AI, just quietly.
Structure earns its keep when it makes good use easier, not harder. Shared prompt libraries, version control for prompts, and clear usage guidelines move a team into Structured faster than any policy that only sets limits (Training Industry has a practical starter list). The test here: does your governance mostly say “no,” or mostly say “here’s how”?
From Structured to Integrated: Silos
A learning team can standardize everything within its own walls and still sit at this threshold for years, because integration is the one move L&D can’t make alone. AI becomes part of how work gets done only when L&D, HR, IT, and compliance move together, and in most organizations each function is running its own AI agenda on its own timeline.
The test: when did those functions last plan an AI decision together, rather than inform each other after the fact?
From Integrated to Transformational: Measurement
The final transition runs on sustained investment, and investment follows proof. That’s where the industry hits a paradox: 98% of learning leaders want to measure impact, yet only 24% have budget allocated for it. Teams arrive within sight of Transformational and stop, not because the value isn’t there, but because no one can show it.
The test: can you demonstrate what improved in performance, not just what was completed in training?
How Learning Velocity Accelerates AI Maturity
Look at the four drags together and a pattern emerges: none of them is about the technology. Every one is about how capability moves through an organization. From one person’s experiment to a team’s shared practice. From one function’s process to an enterprise’s way of working. From activity data to proof of performance.
That movement is what learning velocity means. Not speed for its own sake, but speed with relevance and quality attached: the right capability reaching the right people while it still matters. The maturity model hands you a map of position. Velocity is the discipline of motion.
The teams that keep moving work on two tracks at once. They embed AI into today’s workflows and prove near-term value, while repositioning learning for tomorrow, from delivering content to enabling performance. Progress on one track without the other is how an organization ends up looking mature and standing still.
Progress on one track without the other is how an organization ends up looking mature and standing still.”
How to Tell If Your Team Has Stalled
Ask yourself these two questions:
- What stage are we in? Every learning leader is already asking some version of it, and the model gives you a shared language for the answer.
- How long have we been there? Almost nobody asks this one, and it’s the one that predicts the next two years, because it’s the difference between a position and a trajectory. If the honest answer is “longer than we’d like,” the next step is naming your drag.
The 2026 Learning Velocity Guide breaks down the seven blockers slowing organizations down, including several that show up exactly when a team looks most structured.
About the Author
