The question corporate L&D usually skips
Angelini Academy, the corporate training arm of pharmaceutical and consumer goods group Angelini Industries, has launched a three-year research partnership with University College London that asks a question most organizations deploying AI training tools have quietly avoided: does the AI actually help people learn, or does it just help them finish faster. The study, the first in a wider three-year program, specifically investigates whether AI use strengthens learning and the ability to apply new knowledge independently at work, or whether it allows employees to skip the cognitive effort that deeper, durable understanding typically requires.
That framing matters because it treats ease and speed, the two things most AI training tools are marketed on, as potentially the problem rather than the benefit. Joe Devlin, UCL's Professor of Cognitive Neuroscience on the project, put the underlying premise plainly: learning "always requires a degree of difficulty and effort." He also noted that few organizations open their training programs to this kind of outside academic scrutiny at all, which is itself a telling admission about how little rigorous measurement typically goes into corporate training investment decisions.
What the research will actually measure
The study's methodology goes well beyond the post-training satisfaction surveys that dominate corporate L&D evaluation. Researchers will track groups of managers through Angelini Academy training and into their jobs afterward, using cognitive tests, physiological measurements, and neuroimaging where feasible, to examine how working memory, stress, and sleep affect both learning in the moment and its later application at work. That is a meaningfully more rigorous evidence bar than most vendors selling AI-powered corporate learning platforms are ever asked to meet.
Andrea Prencipe, President of Angelini Academy, described the approach as turning the Academy's own programs into "both the subject of study and a laboratory," an explicit acknowledgment that the company does not yet know whether its training investment is producing the outcomes it assumes. Sergio Marullo di Condojanni, CEO of Angelini Industries, framed the funding as reflecting the group's desire to understand, and help guide, the broader change AI is bringing to workplace learning, rather than simply adopting the technology and hoping for the best.
The gap between AI investment and AI evidence
The research program's timing is notable against the backdrop of Italian corporate AI sentiment it cites. Deloitte figures referenced in the announcement show 82 percent of Italian companies plan to increase AI investment over the next 12 months, with 92 percent expecting productivity gains from that investment. Separately, Cegos International Barometer data cited alongside the study shows Italian HR managers expect roughly one in four jobs could be affected by skills obsolescence within three years, with AI cited as the main driver.
Those two data points sit in direct tension with each other and with the skepticism built into Angelini's own research question. If AI investment is accelerating while the actual learning impact remains unmeasured, and if skills obsolescence from AI is simultaneously expected to hit a quarter of jobs within three years, the pressure to retrain workers faster using the very tools whose learning effectiveness is unproven becomes a real governance problem for any L&D function, not just an academic curiosity. Angelini's research is a direct response to that exact tension, funded by a company with real exposure to the outcome.
Scale of the training operation being studied
Angelini Industries employs about 5,600 people across 21 countries, and more than 500 employees took Academy training in 2025, totaling around 10,000 training hours. That is a modest but real base for the research to draw on, and it gives the study a concrete population to track rather than a lab setting disconnected from actual workplace conditions. Separately, the Academy's Gateway initiative, a 1 million euro effort offering 10 scholarships for IE Business School's International MBA paired with internships at group companies, shows the company is simultaneously investing in more traditional, non-AI-mediated learning pathways.
Planned follow-on projects with academic partners will examine skills gaps in life sciences and robotics specifically, and how training needs to change for employees working closely alongside AI systems day to day. That roadmap suggests Angelini sees this first UCL study as a methodology pilot for a broader, ongoing research function inside its corporate academy, rather than a one-off investigation, which is a more durable commitment to evidence-based L&D than most corporate training functions make.
Why this matters beyond one Italian pharma company
Enterprise learning technology buyers, including the CIOs and CHROs who approve corporate LMS and AI training tool contracts, rarely get independent, peer-reviewed evidence on whether the tools they are buying actually improve learning outcomes versus simply accelerating course completion. Most of the evidence available comes from vendor case studies with an obvious incentive problem. Angelini's willingness to fund outside academic researchers to potentially find that its own AI training investment is not working as assumed is a notably different posture than the typical vendor-funded pilot study.
If the research produces usable findings on when AI assistance helps retention and application versus when it substitutes for the effortful processing that learning requires, that evidence would be directly applicable to any organization's AI-enabled LMS deployment, not just pharmaceutical sector training. L&D and HR technology leaders evaluating AI-powered training platforms should watch for the study's published results over the coming three years as one of the more rigorous independent data points likely to emerge on this question industry-wide.
