AI skills won’t scale until we put humans back in the loop

<h2><span class="h6">In-person workshops demonstrate the value of learning the technology in a group setting.</span></h2>

<p><i>Originally published by the </i><a href="https://www.weforum.org/stories/2026/06/ai-skills-scale-people-in-the-loop/" target="_blank"><i>World Economic Forum</i></a><i> in June 2026.</i></p> <p>Something unexpected happens when you put people in a room to learn AI together. They stop being passive recipients of content and start solving real problems. This pattern has emerged consistently across industries, geographies and seniority levels, and I think it points to something missing in the global AI skilling conversation.</p> <p>As an industry, we have invested heavily in platforms, certifications and self-directed learning tools. These have a role, but they’re not closing the gap. What’s holding most people back from AI is simply the confidence to get started. That’s just one of the lessons we’ve learned after presenting Cognizant’s AI for Impact Community Labs program across ASEAN.</p> <p>The AI for Impact Community Labs brings together nonprofit staff, mid-career professionals and senior leadership teams for structured, in-person AI workshops. Trained facilitators guide participants through progressive prompt engineering frameworks, while volunteer mentors sit at every table to help attendees apply new skills to their own real work challenges in real time. The format is deliberately tool-agnostic: Participants use whatever AI tools they already have access to.</p> <p>Sessions close with a structured ideation activity in which each group identifies a specific challenge, opportunity or workflow in their organization and collaborates live to design a solution that brings out the best of human and AI working together.</p> <p>Since the launch of this program several years ago, the results have been consistent enough to draw some clear conclusions.</p> <h3><span class="h4">ONE: The barrier is skills and the confidence to apply it</span></h3> <p>Across every cohort we have worked with, we see a consistent pattern, no matter their seniority, sector or background: Most participants arrive believing they cannot prompt well. They have access to the tools. What they lack is the ability to structure their thinking to get the best out of AI.</p> <p>These are directors, heads of function, transformation leads and senior leaders with decades of expertise. Their challenge with AI begins with confidence. A platform can present content, track completion and issue a certificate. What it cannot do—but what is needed—is something more human: someone looking at you and saying, you already know how to think, and here is how you apply that to AI.</p> <h3><span class="h4">TWO: A single session produces a measurable skills lift</span></h3> <p>The data from structured, facilitator-led workshops is unambiguous: Across every cohort measured, average prompting ability rises by more than one full point on a five-point scale in a single session. In one senior leadership cohort at a global pharmaceutical company, prompting ability rose from 2.67 to 4.07, and 100% of participants rated themselves at 4 or above afterwards. In a broader nonprofit cohort of 103 participants, 95% said they expected to save 30 or more minutes a day with generative AI.</p> <h3><span class="h4">THREE: Groups unlock use cases that individuals can’t find alone</span></h3> <p>My favorite moments in AI workshops come when people at a table start talking to each other. A fundraising manager hears how a colleague uses AI to draft reports and immediately sees an application for grant writing. Or a clinical educator hears a peer describe AI-assisted scenario planning and recognizes a fit for simulation training.</p> <p>In one session with 37 nonprofit organizations, groups generated use cases spanning donor fundraising strategy; multilingual chatbots for seniors; board paper preparation; volunteer event management; and clinical training materials—all within a few hours. Nearly every participant agreed that group work helped them recognize the importance of human skills like ideation, collaboration and critical thinking. I cannot stress enough that connections like these don’t happen in isolation.</p> <h3><span class="h4">FOUR: Demand is telling us something important</span></h3> <p>It didn’t take us long to learn that human-led AI workshops fill faster than they can be run. One session designed for 73 participants received more than 200 registrations from 93 organizations before the list had to be closed.</p> <p>What’s equally striking is who is registering: senior leadership teams from global organizations—people who already have enterprise licenses for every major AI platform on the market—are choosing to attend these volunteer-facilitated workshops. When asked why, they consistently say something along these lines: Platforms tell me what AI can do, but they don’t help me figure out what I should do with it in my specific role, in my specific organization, right now. Across a sample of 40 senior leadership registrations, lack of skills and talent was cited 28 times as a barrier to AI adoption—outranking unclear use cases, data security concerns, compliance and cost.</p> <h3><span class="h4">FIVE: Learners become teachers</span></h3> <p>It’s been gratifying to see participants say their first planned action after a workshop will be to bring their teams along. When senior leaders leave the room, many are already planning internal sessions for their functions. Frontline workers almost universally say they will share what they learned with colleagues who could not attend.</p> <p>Our volunteer mentors themselves illustrate this dynamic. Across 43 volunteer responses, average AI confidence rose from 3.45 to 4.40 (again, out of 5) through the act of volunteering alone. And 91% said mentoring increased their own understanding of generative AI. Teaching accelerates learning; organizations that provide mentors should view this as a capability-building investment.</p> <h3><span class="h4">The path forward</span></h3> <p>The world has no shortage of AI content. The real work is building the human infrastructure to turn that content into capability. This requires confidence, and confidence scales through people. This can happen through the colleague who sits with you and works through a real problem. Through the mentor who recognizes your hesitation and names it. Through the room that reminds you that everyone else is figuring this out too.</p> <p>The organizations that will move fastest are those investing in the humans who can bring other humans closer to AI—closer to understanding it, applying it and eventually shaping how it works in their specific context. That means creating the conditions for peer learning, recognizing mentors as capability builders and trusting that when humans and AI work together well, the combination produces something neither could reach alone.</p> <p>The most important moment in any AI skilling journey is the conversation that makes someone believe they belong in this future.</p>