Exploring Next-Generation AI Learning Technologies
Russell Sarder, CEO & Founder of AI CERTs – advancing global AI certification & education.
gettyMost companies are “using AI,” but very few are learning with AI in a structured way. Many still rely on a familiar mix: a learning management system (LMS), slide decks, short videos and one-off “introduction to AI” sessions. That model struggles to keep up with the speed at which skills are changing.
PwC’s AI Jobs Barometer shows that skill requirements in roles heavily exposed to AI are evolving about 66% faster than in other jobs. Degree requirements in those roles have already dropped from roughly two-thirds of postings to under 60% in just a few years, as employers focus more on capabilities than credentials.
BCG’s 2025 research suggests only around a third of employees feel they have been properly trained on AI. AI investment and experimentation are rising, but learning systems are often still built for a slower, more predictable environment.
These pressures are not simply creating a need for more AI training; they are changing what workplace learning looks like. Several shifts are emerging that point toward a more adaptive, continuous model of AI learning.
The first big shift is from content to agents. In simple terms, agentic AI uses AI “agents” that can pursue goals with limited supervision, plan steps, call tools and adapt based on feedback.
McKinsey’s 2025 work on agentic AI points out a paradox: nearly 80% of companies say they now use generative AI, but many still do not see a clear bottom-line impact. One reason is that people are experimenting in isolation, without structured support, feedback and guardrails.
Learning agents are helping offer a different approach. They can maintain a live skills graph for each employee, based on role, work outputs and assessments. They can curate short lessons from internal documents, customer interactions and external sources. They can set practice tasks, review outputs and give targeted feedback, escalating to human mentors when needed.
Walmart is a useful example of this direction. Its internal My Assistant tool helps corporate associates summarize documents, draft content and prepare analyses, acting as a day-to-day partner instead of a separate learning portal. Walmart also announced a collaboration with OpenAI to provide free, customized AI training and certification for its U.S. frontline and office associates through Walmart Academy and the forthcoming OpenAI certification program.
Together, an in-the-flow AI assistant and structured credential pathways show what next-generation AI learning can look like in practice.
The second shift is from flat content to immersive, spatial experiences.
Spatial computing and AR/VR headsets are moving beyond demos into real enterprise use cases. IDC expects training and industrial maintenance to be among the leading drivers of AR/VR spending, and firms are already using headsets for onboarding, safety and complex procedural work.
PwC’s VR training study gives this real weight. In controlled experiments, VR learners completed training up to four times faster than classroom learners, reported far higher confidence and showed better retention of skills. When you are dealing with safety, compliance, or high-value customer interactions, those differences matter.
Accenture’s work with San Diego County is a good illustration. They helped design VR simulations for caseworkers who handle complex eligibility interviews for public benefits. Instead of sitting through slides, staff can practice realistic conversations in a safe environment, learn from mistakes and repeat rare scenarios as often as needed.
When you pair immersive environments with AI that can generate scenarios on the fly, adjust difficulty and provide feedback, you get an embodied way of learning that traditional e-learning isn’t suited to replicate.
The third shift is from standalone courses to copilots and sandboxes that sit inside real work.
The Microsoft and LinkedIn “Work Trend Index” shows that use of generative AI at work has almost doubled in a very short period, with about 75% of knowledge workers already using it. Clutch’s research suggests only one in three has had structured AI training, and many are unclear about their company’s policies.
Multimodal copilots are helping to close some of that gap. These systems work across text, code, images, dashboards and voice. They do not just answer prompts; they walk people through tasks, explain reasoning and suggest improvements. Every interaction can be a learning event, and every mistake becomes a data point.
JPMorgan Chase offers a clear example. In 2024, the bank rolled out its internal large language model (LLM) suite to around 200,000 employees, with plans to reach most of the firm. An in-house coding assistant has lifted software engineer productivity by roughly 10% to 20%, and leadership has talked about potential value in the range of $1 billion to $1.5 billion as use cases scale. More recently, JPMorgan began using internal AI tools to help draft performance reviews, bringing AI into a traditionally judgement-heavy process with strong guardrails.
If you are a CEO or business leader, there are at least five decision areas you need to own.
You need a forward-looking map of the capabilities that matter for your strategy over the next three years, not just a catalog of today’s tools.
Decide what “baseline AI fluency” means in your context, how you will deliver it and how you will measure it.
Agentic AI deserves attention, but not simply because it is the latest technology. Prioritize pilots where AI can materially improve revenue, cost, risk or customer experience, and build learning into those workflows so employees develop capability as they use the technology.
VR and spatial computing will not be the right answer for every learning need. They are most compelling when employees need to practice situations that are difficult, expensive or risky to reproduce in the real world, including operations, safety, compliance, high-value client work and public services.
Make learning outcomes visible in promotion, pay and internal mobility decisions. Digital credentials and verifiable AI certifications can help here if they are tied to real work.
AI is becoming a central driver of productivity and competitiveness. I believe the organizations that win the next decade will treat AI learning systems as core infrastructure, alongside cloud, data and security.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
