L&D Works on Tools, Not on Decisions
L&D teams are working more with AI than ever before. Teams create content faster, build courses in hours instead of weeks, and test chatbots, question generators, and translation tools. With many steps of work, things are moving.
So why are so many L&D leaders still fighting for a seat at the AI strategy table?
Because work and impact are two different things, and the gap between them is where L&D’s credibility is lost.
We surveyed over 1700 learning experts to find out where AI in L&D stands today. 78% of L&D teams said they are not in the room when budgets and priorities are decided, and use someone else’s perspective.
It influences composition in organizations as does power. The teams shaping AI strategy now will get credit for the results later, and those left out of those discussions won’t. What is at stake is the ability to have a story in the moment that matters most.
Here’s some data visualizations, and what you can do about them before your next stakeholder interview.
The Number That Should Concern All L&D Leaders
25% of L&D teams say their primary reason for adopting AI is personalization at scale. Less than 4% prioritize business performance.
Now think about your next conversation. When your CFO walks into a room and asks what AI investment for L&D is delivering, what’s the answer? “We’re personalizing the student experience on average,” or “We’ve reduced the productivity of new hires by 30% and here’s the data”?
Personalization without a business case falls flat in the upper echelons. L&D often speaks the language of the student experience, while management speaks of revenue, retention, and productivity, and right now those two languages don’t mix. The cost of that loss is loyalty.
Try this before your next stakeholder interview. Take any AI effort you’re currently working on and ask: What business metric should it go for? Learning metrics won’t help you here, so consider productivity time, sales win rate, compliance incident rate, or customer churn. If you can’t name one, that’s your first problem to solve, and it’s solved before you even enter the room.
Reorder the step around that number and lead with it. Rather than “We improve the student experience,” try “We use personalized learning to close the skills gaps that shorten your sales cycle.”
Same step, but a completely different discussion.
The Resistance Problem Is Not What You Think, And It Doesn’t Come When You Think About It
37% of L&D teams say stakeholder resistance is their biggest challenge to AI adoption. Only 12% say the barrier is lack of internal knowledge.
The resistance that many L&D leaders navigate is rarely one-sided. It appears several times, for different reasons, often at the same time. Treating it as one problem with one solution is why most teams hit the same wall.
Think about who is actually disengaged in your organization right now.
Senior leaders who haven’t seen the business case they believe in are anti-AI. They weigh the risks, and no one has yet shown evidence that the returns justify the investment. That is an issue of credibility, and it is resolved by evidence.
Managers who don’t trust AI-generated content to meet their team’s standards may notice something that misses the mark, or hear enough about AI ideas to be wary. That’s a quality concern, and it’s solved by showing them your review process.
Employees who feel uncomfortable about what AI means for their jobs are resistant to learning. They resist a version of AI discovery that feels forced on them rather than designed for them. That’s a change management problem, and it’s solved by engaging them early, being transparent about what AI will and won’t change, and making the learning experience feel like progress.
Topic experts don’t feel outdone when AI writes content they used to block. They protect something they care about. That is the problem of co-ownership, and it is solved by repositioning it as an expert reviewer and quality filter rather than being sidelined.
IT or legal teams slow things down with governance concerns they don’t resist. They flag a process gap, and it’s solved by bringing them in as a partner before you need their approval.
The point is not that all of these are equally common in your organization. Identifying where the resistance is coming from is the real first step, before you can decide how to respond. Teams that treat all your concerns in the same way often tend to do more communication or more AI training, and end up frustrated because they use the right answer to the wrong question.
Here’s a trick that works no matter where the opposition lives. Go to the most antagonistic person in the room, whoever it is, and ask them one question: “What does success look like to you?”
Skip “What are your concerns,” which invites a list of arguments, and skip “Let me show you what AI can do,” which prompts defensiveness. Stay with the question, and build your next driver to deliver just that. When a skeptic helps explain the path to success, they become accountable, from being the judge to the owners of the other.
That dynamic applies whether the skeptic is a CFO, a company executive, a nervous employee, or a Subject Matter Expert worried about their role.
Resistance often comes back to trust, evidence, and a sense of control. Give people that, in a way that is more relevant to their specific concerns, and resistance tends to move.
Market Segmentation, And The Gap Is Already Wider Than You Think
27% of L&D teams have been using AI for years, 46% have just started, and 27% haven’t started at all.
Reading that as a slow curve, with early adopters, regulars, and contributors, misses what’s really going on. This is a division, and the distance between the teams grows every quarter.
The teams with the longest history are already leading, and they continue to add to it. Every pilot builds institutional knowledge, every win earns more budget and more approval, and every quarter of performance makes the gap harder to close.
The details are where teams invest their AI effort. The most common uses are content creation (30%) and research (21%). The most common are improved reporting (11%) and regular delivery (11%). Teams are focusing AI effort on parts of the job they feel familiar with, such as outlining content and summarizing research, while underinvesting in parts that can change their strategic posture: connecting learning to results, delivering it where and when it’s needed, and proving its impact.
Using AI to do the same things faster is a practical advantage. Using AI to tackle very different problems is a strategic shift, and one wins you time while the other leads you.
If you’re in the 46% who just started, here’s a move. Pick one business problem that is most visible in your organization right now, whether that’s a new product launch, a storage problem, or a compliance deadline, and build an AI-assisted learning intervention around it. Measure it against business metrics from day one. Focused wins in the most visible area do more for your strategic environment than ten efficiency improvements running quietly in the background. Start small, but start where people are watching.
The Cycle of Exclusion, and How to Break It
Only 22% of L&D teams are included in AI strategy discussions.
AI is reshaping the way organizations hire, develop, and retain their people, yet in 78% of organizations the task of building capacity is not included in the discussion.
The cycle goes like this: IL&D is not included in strategic discussions, so it cannot set the direction of AI adoption. Without a seat at that table, it cannot conduct tests that can produce evidence. Without evidence, it cannot make a case to be filed. The cycle continues.
Breaking it means producing evidence before the invitation arrives. Evidence requires access, and access requires a stake, so get yours.
Look for a business leader in your organization who can’t sleep because of a people problem: a skills gap that affects delivery, a new system that no one knows how to use, or a team that’s always off target. Don’t approach them with a learning solution but with the question: “Can I run a six-week pilot to help with this, and can we agree in advance how we’re going to make it work?” Many will say yes. Six weeks later, you have data, and data is how you enter the conversation. Make outsourcing look like a business risk, one outcome at a time.
The Moral Gap Nobody Talks About
15% of learning professionals feel prepared to manage the implications of AI principles in learning.
AI is already informing people’s learning strategies, influencing who gets development opportunities, what learning methods are recommended, and how performance is evaluated. Most L&D professionals, however, do not feel equipped to manage the risks that come with that.
Organizations that have not thought carefully about bias in AI-generated content, transparency in algorithmic decision-making, or data privacy in student analytics are not avoiding ethical risk. They postpone it. Deferred moral hazard does not disappear; it’s quiet until something comes to light that is very hard to go back from.
You don’t need a complete set of ethics on the first day. You need three things. First, a review step in the entire AI content workflow, where someone checks the content before it goes to readers, every time. Second, a clear internal answer to the question “What student data do we use and who has access to it?” Third, a discussion with your legal or compliance team before measuring, not after something has gone wrong. Those three things won’t cover all the behaviors AI creates, but they’ll give you a solid foundation to build on.
What the Data Really Says
Undo all the math in this piece, and the story doesn’t change: IL&D is capable, but it’s not always positioned where the business needs it.
The gap comes down to the distance between improving the student experience and driving business results. It also shows how AI is being used, whether it’s to speed up routine work or take on strategic problems.
Teams closing that gap use one small trial, measuring the right things, building one piece of credibility at a time, and using each win to build on the next.
All the action in this piece is singular: one metric, one question, one pilot, one wedge, one review step. That was done on purpose. Teams that try to solve the AI revolution all at once tend to end up in analysis handicaps, while teams that pick one thing and prove it works are the ones that create significant cumulative advantage.
The whole AI strategy in L&D can wait. All you need right now is one deliberate next move.
