If AI works, why are 53.3% of L&D still struggling?
If your engine light keeps blinking but your car’s dashboard says everything is fine, the problem isn’t just under the hood. The problem is also in the dashboard. This is more or less what the enterprise L&D landscape will be like in 2026. In Scheer IMC’s 2026 State of Learning Technology report, 53.3% of decision makers say effectively integrating AI or new learning technologies is their biggest challenge. At the same time, four in five say current learning technologies are at least somewhat effective. Both statements could be true. This is exactly why it’s important.
This is not a contradiction that arises from confusion, but a contradiction that arises from measurement. Many organizations still evaluate learning technologies according to one standard while expecting outcomes relative to another. Dashboards report on stability, access, and basic usability. Businesses demand competency, productivity, and certification. These are not the same conversations.
AI moves from demo stage to delivery pressure
For a while, AI in corporate learning existed like a concept car in a showroom. It looked impressive in bright lights, garnered confident comments, and held up very well on Monday morning’s wet highways. That stage is over.
The report makes that clear. AI is now actively used throughout the learning process for 43.1% of organizations, and another 14.8% say it is fully integrated across L&D operations. Investments are moving in the same direction. Approximately 61.4% plan to invest in AI-powered authoring tools and 60.5% in AI-powered coaching tools within the next 12 months.
The appetite is real. The difficulty is also real.
Once AI enters the enterprise environment, it stops being a shiny feature and starts acting like a new railroad line in an old city. Suddenly, the questions were no longer about speed, but about signals, safety, routes, and who was responsible if something went wrong. Integration into existing systems, technical complexity, and data security are current friction points. It’s no coincidence that 92.9% of organizations say they have concerns about data security and privacy in AI-based solutions.
The challenge isn’t deciding whether AI belongs in L&D. The challenge is making it work in a way that businesses can trust.
The word “effective” carries too much weight.
When four out of five organizations say their learning technology is effective, it’s tempting to hear that as a verdict of strategic success. That would be a very generous read.
Most likely, many respondents answered more specific questions.
Will the system work reliably? Will people have access to learning? Will it support constant, frictionless childbirth? If that is the norm, then yes, many systems are likely effective.
But AI has raised the bar.
Just as a kitchen can be well-organized but still make the wrong meal, a learning platform can be stable but strategically inadequate. Having a good shelf, a sharp knife, and decent ingredients doesn’t necessarily mean that dinner will help you achieve your business goals. Similarly, a platform may offer healthy logins, strong learner feedback, and frictionless management, but pass more difficult tests of whether it improves performance, closes skills gaps, or supports transformation in measurable ways.
This report reveals more. The most common way organizations measure learning is employee feedback, used by 55.5%. However, 44% say the biggest barrier to measuring L&D ROI is tying learning outcomes to tangible business impact. The problem isn’t that organizations don’t have data. That means much of the data still acts like a weather forecast when executives want to know what to expect for their business.
Success stories from the past still feel good.
For many years, learning technology was often valued like infrastructure. If it’s safe, compliant, easy to use, and widely adopted, it’s doing its job. This logic makes sense when the central challenge is digital distribution at scale.
The briefs became even heavier. L&D is required to support workforce adaptability, skills visibility, and AI readiness. In this report, 86% of organizations say systematic skills management is a strategic priority for 2026. This is a serious change. The shape of the learning stack itself is similar. 73.1% currently rely on one central LMS as the backbone of their L&D ecosystem, rather than aggregating platforms such as kitchen items purchased during a late-night shopping spree. It’s a maturity, but it’s also an architectural maturity. Measurement maturity has not yet caught up.
Completion rates still matter. Compliance remains important. Satisfaction remains important. But if these remain headlines, with AI integration struggling and business impact ambiguous, L&D risks being presented with a beautifully wrapped package with no clear evidence of what’s inside. There are no ambitious issues with this function. There is a problem with the translation.
If learning is alive in work, measurement must follow as well.
One of the most revealing findings in the report is that engagement works best when learning is embedded in daily work. In fact, 85.5% of decision makers say embedding learning into their daily workflow is the most effective way to drive engagement. That should also tell you something about measurements.
When learning happens more in the flow of work, evidence of impact doesn’t remain locked inside the LMS like luggage left on an airport carousel. It has to show up where work shows up, whether it’s faster time to competency, better decision-making, increased internal mobility, or reduced delays in change efforts. Not all results require complete numbers. Senior leaders know this. What they expect is a reliable line of sight between learning efforts and business moves.
Many teams still lack that line. The report shows that L&D is moving away from activity metrics and focusing on outcomes such as productivity gains, skills development, and skills gap analysis. The direction is correct. Execution is more difficult than intention.
This begs the question, “What is L&D measuring?” It’s very important. It’s not a provocative headline. It’s a strategic test. When integrating AI remains difficult, proving skills remains difficult, and connecting business impact remains difficult, the old definition of “effective” is no longer sufficient.
So what is enough?
The full State of Learning Technologies 2026 report dives deeper into where this gap is most pronounced, where investments should go next, and why trust, governance, and connected data will be the real differentiators. It leverages the perspectives of more than 420 corporate L&D decision makers around the world and is further shaped by Shea IMC’s more than 25 years of experience helping organizations address complex learning challenges at scale.
Founded from a pioneering university initiative by IT visionary Professor Scheer, the company has supported more than 1,300 organizations and 10 million learners through its learning platform, content, and strategic expertise. The combination of market perspective and real-world experience makes this finding even more relevant at a time when L&D is under increasing pressure to prove not just its activity, but its impact. If your learning dashboard looks healthy, but the engine still sounds uncertain, it’s worth taking a closer look at the broader findings.
