Everyone is talking about AI adoption in learning and development. Fewer people are talking honestly about what is happening on the ground: the gap between organizational policy and employee behavior, the L&D teams being pulled into transformation work they were never meant to own, and the compliance training employees may already be asking AI agents to complete for them.
That is what OpenSesame’s recent webinar tackled head-on. L&D strategist and EdTech expert Lori Niles-Hofmann joined host Mehdi Tounsi for a fast-moving conversation about intelligent tutors, skills intelligence, global regulation, and the uncomfortable question many teams are already facing: Is L&D being handed a job it was never equipped to do?
Here are the takeaways that matter most (you can watch the highlights below, too!)
Your employees are not waiting for your AI policy
Here is a reality many organizations are still catching up to: Regardless of your internal AI readiness level, your employees are already experimenting.
They are testing tools at home, using personal devices, and building habits and workflows that may have little to do with what IT has approved.
Her own admission that she has not used a Google search in two years was not a flex. It was a signal. The gap between your organization’s official AI stance and your workforce’s actual behavior may be wider than leaders realize.
And the longer that gap goes unaddressed, the harder it becomes to guide AI use in ways that protect the business, support employees, and build the skills people need for what is next.
AI transformation is not L&D’s problem to solve alone
Many L&D teams are caught in a difficult trap right now: They are being asked to take ownership of the entire AI shift.
That can include workflow redesign, technology strategy, change management, governance, and capability building. But those are not all the same job.
The distinction matters. L&D’s role is to help people build the skills and confidence they need to work in a changing environment. It is not to design that environment on its own.
As Lori said, “We do not own the enterprise workflow. That’s something that needs to be owned either by the transformation or your chief AI officer. We’re there to build capability.”
L&D should have a clear seat at the table, but it should not be expected to carry the entire transformation effort.
When that work gets pushed onto L&D without clear ownership, both sides suffer. The transformation loses the right structure and expertise, and capability building gets deprioritized.
If your L&D team is feeling overwhelmed, the issue may be a lack of clarity. The answer is not always more resources. Sometimes, it starts with a better conversation about ownership.
The one-size-fits-all course may have a shorter runway than you think
The traditional corporate course — the same sequence, same content, and same experience for every learner — is already being challenged by a more adaptive model.
Intelligent tutors do not necessarily replace the content organizations already have. Instead, they can make that content more responsive by combining it with organizational context, knowledge management, and individual learning history to personalize the experience in real time.
“Maybe, potentially, the one-size-fits-all course is something that’s no longer going to exist,” Lori said. “Or it will simply become a body of knowledge that will be fed into an intelligent tutor that then brings in context from your organization. It might look at knowledge management. It might look at your CRM. All of these things together, rather than just a standard course. Usually, when I say this, there’s a lot of heart palpitations — because if we don’t build courses, what do we do?”
That is the right question for L&D teams to ask.
A physics study Lori referenced found that human+AI instruction outperformed both AI-only and human-only groups. The takeaway is not that instructors become obsolete. It is that learning design has to evolve to take advantage of what AI makes possible.
For L&D practitioners, the question shifts from “How do we build better courses?” to “What does our content become when it is part of something smarter?”
“AI adoption” is a vanity metric
The adoption numbers circulating right now can sound impressive. One McKinsey survey found that 88 percent of respondents said they were using AI in at least one business function.
But self-reported, single-function use does not tell you whether AI is changing how work actually gets done.
Lori was direct: “Not all adoption numbers are equal.”
She drew a useful parallel to the SaaS era, when teams learned that monthly active users and daily active users were very different metrics, and neither automatically proved business impact.
“Those are very different metrics to somebody actually going in and building an agent and redesigning their workflow,” she said.
Microsoft data puts global generative AI use at one in six people. That number should help recalibrate expectations.
The organizations with a real advantage are not necessarily the ones reporting the highest adoption rates. They are the ones moving beyond logins and surveys to ask better questions: Has behavior changed? Are employees applying AI responsibly? Are workflows improving? Are people building the skills they need to use these tools with confidence?
Governance helps teams scale, not stall
The regulatory landscape is moving quickly, and it is more fragmented than many teams realize.
The EU AI Act, U.S. state-level variation, China’s content labeling requirements, and sector-specific mandates in industries like aviation, financial services, and life sciences can all apply differently. What matters is not just where your L&D team sits. It is where your employees are located.
The practical stakes are real. If AI recommends a learning path for one employee and not another, that decision needs to be explainable.
“It may not seem like a big deal where you say, well, one person was recommended a piece of learning and another person was not,” Lori said. “It actually really, really does. And you need to be able to defend that.”
Governance is not what slows AI adoption down. Done well, it is what makes scaling possible.
The teams moving with confidence are not simply the boldest. They are the ones that set clear expectations early around data visibility, decision traceability, and human oversight. The teams that do not build that structure now may spend more time backtracking later.
The bottom line
The real story of AI in L&D right now is not just the adoption curve. It is the clarity gap.
Who owns AI transformation? What does adoption actually mean? What should learning look like when AI becomes part of the design? And how can organizations move quickly without losing trust, safety, or accountability?
The teams that lead through this moment will be the ones that get specific about their mandate, their metrics, and the learning architecture they are building.
That clarity matters now, before the next wave of tools arrives and the pressure grows.
Ready to see where your organization stands? Take OpenSesame’s AI Readiness Assessment for an honest starting point. You can also watch the full webinar recording here.
