Here is what nobody tells you about corporate AI training: the staff who need it most are often not the ones showing up, and the ones who do show up often already know more than the curriculum assumes. In a class I ran in 2025, a group of Millennials and Gen Zs sat through my dataset analysis demo — bar charts, pie charts, line charts generated in ChatGPT — and politely informed me afterwards that they were already generating receiver operating characteristic (ROC) curves and scatter plots as part of their daily work. Their boss had sent them anyway. That moment captures a problem playing out across enterprises right now. Companies are rushing to upskill their workforces in AI, and many of those efforts are quietly misfiring. Not because the trainers are poor or the technology is overhyped, but because of avoidable mismatches between what is being taught, who is in the room, and what those people actually need. Irrelevance of material One of the pieces of feedback that stood out was the irrelevance of the material taught. Do not get me wrong. It is not that the generative AI content is lousy. It is the combination of two main factors: generative AI content is wide in scope and not every segment is relevant to every attendee; and enterprises, by virtue of the way they operate, are not usually able to send entire departments for training at the same time. Take content analysis, which is comprised of the summarisation and analysis of written content, its sentiment and bias, the implications of the development to the staff or team, and the suggested ways forward. Beneficial as this generative AI feature may sound, it may not be relevant to everybody. It might be relevant to someone doing active research on the internet for the purpose of developing a new line of products or services, but it may not be immediately relevant to someone with oversight of the company’s finances who needs to go over the general ledger on a daily basis. As companies cannot afford to have a single department go offline while training happens, a typical class often consists of a mix of people from different departments. This makes it difficult for the trainer to customise the material to suit the attendees. The result? The trainer often uses a consumer example for demonstration or chooses to focus only on a specific topic that may alienate staff from other departments. Also Read: The accordion effect: How AI follows the rhythm of expansion and compression Younger staff already know generative AI There is a widespread assumption by paymasters that employees are starting from near zero when it comes to AI. For older cohorts or less technical functions, that assumption may hold. For Millennials and Gen Zs who have been using ChatGPT, Copilot, and similar tools in their personal and professional lives for years, it does not. In a corporate class I conducted in 2025, the staff was made up mostly of Millennials and Gen Zs. As I went about delivering the lessons, I realised that these young learners were already aware of the content. Being tech-savvy, they were very much at ease with technology. This resulted in a challenge for the trainer far greater than educating from zero: learners were somewhat bored. When a company signs up for the syllabus in advance without assessing baseline competency, this outcome is almost guaranteed. This is not an argument against training younger staff. It is an argument for knowing what they already know before deciding what to teach them. The expectation gap: Agentic AI vs generative AI In another company, there was a misunderstanding about the content that would be delivered. Most of the staff were already technically savvy and were, in fact, already using ChatGPT and Copilot every day. Some of the senior staff were questioning if I would be covering agentic AI in detail. The truth was that agentic AI — the next big thing in the AI universe that has already arrived in a small way — exists as a small segment in the generative AI course and only serves to introduce learners to this whole new world. While a deeper dive workshop is available, that is a separate programme entirely. What I am sharing here is that there was a disconnect between what I was to deliver and the learners’ expectations. Thankfully, the senior staff in question accepted my explanation and were happy to participate in the class activities. In addition, there were some staff who were new to some of the concepts I delivered, so my lessons went ahead without a hitch. Also Read: AI slop is a strategy problem, not a content problem How to address these issues Although the issues described above were not huge issues, they are common and do show up across different corporate clients. Here are some practical solutions: Understanding what the company needs: Leaders might wish to understand the profile and technical competencies of their staff before deciding whether to send them for training. Having a better understanding of these would ensure a better fit between the needs of the company and staff and the technical complexity of the courses. Understanding whether training matches the needs: Before rushing into the latest corporate hype, company representatives could seek clarification about the course content. This can be done through a visit by the trainer to better understand the needs and clarify any misunderstandings or concerns. Having the trainer at the company’s premises before course commencement allows greater understanding and, in turn, a better, tailored fit to the trainees. Time to set it up right In most of my corporate engagements, the majority of attendees are genuinely motivated. People take notes. Some arrive early. The appetite for learning is there. What is wasted, when things go wrong, is that appetite. A staff member who sits through two days of content they already know does not come back with a negative view of AI — they come back with a negative view of their company’s ability to invest in them thoughtfully. The fix is not expensive or complicated. It requires leaders to slow down slightly before committing — to ask who they are sending, what those people already know, and whether the course they are booking is actually the course those people need. That pause, taken seriously, is the difference between training that sticks and training that is forgotten by the following Monday. — Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic. The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27. Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected. The post How to fix corporate AI training: A practical guide appeared first on e27.
Author: Jack Yu
Source: e27