The numbers are stark. According to research from the Adecco Group, McKinsey, and the World Economic Forum, 90 percent of CEOs don’t believe they have the workforce needed for an artificial intelligence-powered business. Nearly half of leaders cite skill gaps as a barrier to AI adoption. And 39 percent of today’s skills are expected to be obsolete within three years. Yet only 1 percent of companies investing in AI are investing in workforce skills to match.
For Learning and Development (L&D) leaders, that gap isn’t abstract. It shows up in the distance between what leadership is asking for and what your team has the time and resources to deliver: more courses, faster deployment, clearer outcomes—all while the content you need to build keeps evolving.
The good news: Workforce reinvention doesn’t have to mean starting over. Research and practitioners working in this space have identified a set of practical moves that L&D teams can make right now, without blowing up existing programs or waiting for a top-down mandate to get started. Here are five of them:
1. Know what skills you have before deciding what to build.
Most L&D teams have a general sense of where skill gaps exist. Fewer have a systematic view of it, one that maps current capabilities to where the business needs to go. That gap matters more now because AI is shifting role boundaries faster than traditional skills assessments can keep up.
A practical starting point: Run a lightweight skills audit focused specifically on human+AI collaboration, the skills that help people work alongside AI tools effectively, rather than just the technical skills to use them. This includes things such as knowing when to trust AI output, how to structure work that involves both human and machine input, and how to stay oriented when automated processes change the nature of a role.
You don’t need a full competency framework to do this. Start with three to five roles where AI is already being introduced and map what’s changing about the work itself. That gives you a baseline to measure against and a defensible story to tell leadership about why certain training investments are being prioritized.
2. Treat trust and safety as a training priority, not an afterthought.
One of the most consistent findings across organizations navigating AI adoption is that governance, ethics, and compliance aren’t just legal considerations. They’re workforce competency issues. When employees don’t understand the boundaries of how AI should be used in their role, they either over-rely on it or avoid it entirely. Both create risk.
For L&D, this means building AI ethics and responsible use into the fabric of training rather than relegating it to a one-time compliance module. Practically, that looks like incorporating “when and how to use AI” guidance into role-specific onboarding, adding scenario-based practice that surfaces the limits of AI tools, and giving managers a framework to coach their teams on appropriate use, not just whether they’re using the tools at all.
Organizations that get this right tend to move faster on AI adoption overall because their people feel more confident making judgment calls, rather than waiting for explicit permission at every step.
3. Make the business case to leadership—and make it specific.
Workforce reinvention stalls when it stays in the L&D lane. The organizations that move fastest are the ones where Chief Human Resources Officers (CHROs) and Chief Learning Officers (CLOs) have made skills development a boardroom-level conversation, tied to business outcomes rather than training completion rates.
If you’re trying to elevate the conversation internally, the most effective approach is to connect skill gaps to specific business risks. Rather than framing it as “our people need AI training,” the framing that tends to resonate with executive teams is more concrete: “Here are three processes where we’re losing speed or accuracy because our team doesn’t yet have the skills to work effectively with the tools we’ve already deployed.”
That specificity is what turns workforce development from a cost center into a strategic argument. It also makes it easier to get budget and cross-functional support, because the problem is no longer abstract.
4. Focus on collaboration skills, not just technical ones.
There’s a tendency to equate AI readiness with technical skills: prompt engineering, tool proficiency, data literacy. Those matter. But research from McKinsey, HBR, and Deloitte consistently points to a different category as the differentiator: the human capabilities that become more valuable, not less, as AI handles more routine work.
These include the ability to redefine task boundaries as AI takes on more of a workflow, cross-functional communication about how AI is changing shared processes, and the judgment to know when a human decision is needed versus when to trust an automated output. These aren’t soft skills in the traditional sense. They’re collaboration skills specific to human+AI work environments, and they require deliberate development.
A practical tip: When building or sourcing content for AI readiness, look for scenarios that put people in situations where the AI output is ambiguous or wrong, then ask them to decide what to do next. That’s where the real skill-building happens, and it’s far more valuable than training that only shows people how to use a tool correctly when everything goes as expected.
5. Embed development in the flow of work, not separate from it.
One of the most consistent failure modes in L&D programs is the assumption that learners will carve out dedicated time for training separate from their work. In practice, they don’t, especially in an environment where everyone is already being asked to do more.
The shift toward “learning in the flow of work” isn’t new, but it’s become more urgent as the pace of change accelerates. In practice, this means designing content that’s short enough to consume in context, ideally 5 to 10 minutes per module, and specific enough to apply to what someone is working on that week. It means delivering content through channels people already use, whether that’s a Slack integration, a manager-led team conversation, or a two-minute pre-meeting primer.
It also means shifting how you measure success, away from completion rates and toward indicators that something changed in how people work. That’s a harder measurement problem, but it’s the one that matters to business stakeholders.
The Shift from Episodic Training to Continuous Development
Taken together, these five moves represent a broader shift that organizations navigating AI adoption successfully tend to share: from treating training as an episodic event to treating skills development as a continuous, strategic function. That doesn’t mean doing everything at once. It means finding a manageable starting point, whether that’s a focused pilot, a high-priority role, or a specific skill gap, and using it to demonstrate what’s possible before scaling.
The organizations adapting well aren’t the ones with the most elaborate plans. They’re the ones that found a place to start and built from there.
OpenSesame’s Workforce Reinvention Series is a free collection of eight short courses grounded in research from McKinsey, HBR, and Deloitte, designed to help L&D teams give their learners a practical starting point on human and AI skills without requiring a full program overhaul to deploy it.