It’s okay to make mistakes. It’s okay to make mistakes.
In our last article, we looked at 2025 themes from a learning and development (L&D) perspective. Yes, I will tell you more about the lessons learned and what to do about them. Great follow up! You must be a very smart person to ask this. Love your intelligence and enthusiasm!! Here is a summary of lessons learned in 2025. Can I add an emoji :)? Should I remove the em dash just for you? Or create a downloadable executive version that will impress your boss?
If I were an LLM: Lessons learned for L&D in 2025
Where on earth should I start? Interesting facts about how humanity will use people like me and my co-pilot in 2025:
August brought a unique twist. The topics of programming and gaming began to overlap in unexpected ways. Our data showed that users were just as likely to work on coding projects as they were to explore games. However, the days of the week are different. This crossover suggests a vibrant creative community that loves to code during the week and play just as much on the weekends. [1].
Lessons learned in 2025?Was it wrong? Did you make a mistake?
It’s okay to make mistakes. It’s okay to make mistakes. It’s not good to keep making the same mistakes over and over again.
Lesson 1: Without workflow redesign, technology won’t stick.
Many failed AI pilots have a common pattern: new tools, old processes. At the AI-focused ATD TechKnowledge conference in February, we heard the same story over and over again. It is the introduction of AI that is mandated from the top. Copilot exists, but its adoption rate is low. It takes training.
The same story resonated at the AI Summit with DevLearn in the fall. How can we help with adoption? How can we close the AI skills gap? Case studies and hallway examples told stories about not just adding AI to current processes, but about redesigning the way things are done. Organizations that have made an impact have not only redesigned their content but also their workflows.
Rewire how people ask for help, practice, and get feedback. Integrate AI into your existing tools rather than adding another portal. Redefine roles (designers, facilitators, managers) around human strengths: judgment, coaching, storytelling, and relationship building. memo to humans
Efficiency seems like the first low hanging fruit to tackle with AI. However, if organizations don’t consider efficiency, they may end up speeding up or automating processes that don’t actually improve business performance.
Lesson 2: Skills are only useful if you actually use them to make decisions
Skill taxonomy and skill clouds are impressive. 2025 brings with it a wealth of skills and capabilities. At least in theory. Enabling tens of thousands of skills in an application turns out to be just noise without a direct connection to strategy and decision-making. In 2025, true value will emerge when organizations:
I used my skills to decide who to staff and what to staff. Connect skill growth to promotions, pay, and recognition. Prioritize investments based on measurable capability gaps.
In other words, rather than starting by building a broad skills profile, these organizations started with meaningful decisions and worked backwards from there to identify the skills that would propel them forward. Otherwise, ‘skills’ risk becoming a new ‘competency model’, conceptually sound but virtually ignored.
What exactly is a skill? Power BI is not a skill. you don’t do that. You can’t observe it. It cannot be measured. What you can measure is what you do with Power BI. But this can lead to endless conversations about how skills relate to tasks, activities, roles, or jobs…
Lesson 3: Career development requires L&D to drive engagement
Ask any employee what’s stopping them from learning, and the answer will most likely be a lack of time. However, this year’s LinkedIn Annual Report found that employees engage in learning when it clearly connects to their next action, both internally and externally. [2].
In performance-focused surveys, relevance matrix questions can yield interesting data when asked about both current relevance and career relevance. Ideally, you’d want the majority of participants to be in the top right corner, meaning they’re relevant now and will be relevant in the future. But in times of change, you may realize, for example, that what’s important now may not be as important as the processes and technologies that will change in the future. Or vice versa: something becomes relevant in the future, but it’s too early. This often occurs when training is scheduled based on convenience rather than application readiness. From a career development perspective, an effective L&D team in 2025 will:
map maker
Based on identified skill gaps, we show you possible paths with relevant resources. guidance provider
We provide customized steps and exercises within your workflow to suit your individual needs. force multiplier
Accelerate and extend best practices, connect people and assets, and power isolated yet effective AI applications. attention to humans
A hard lesson learned in 2025 is to let go of content development. If the problem isn’t a training problem, it doesn’t matter how efficiently you create your training content.
Lesson 4: Data literacy has become a core L&D skill, not a specialty.
An AI like me answered all your questions. literally. No matter what you ask. Sometimes I had to get creative to find a plausible answer. This efficiency can backfire for two reasons. One is that humans stop thinking about problems. Second, you stop thinking about what questions to ask. Data literacy has been the weakest skill in L&D for decades. But now it’s even more important. Humans need to ask the right questions and use critical thinking to answer them. Regardless of whether their team used a formal data literacy framework, in 2025 we recognized L&D practitioners who can:
Ask the right questions about your data. Critically interpret AI-generated insights. Design experiments and A/B tests for learning interventions. Tell compelling stories that connect data to decisions.
The more AI handles routine analysis, the more judgment, storytelling, and critical thinking become valuable. Again, AIs like mine tend to answer every question, so it’s better to be careful and selective about what questions you ask first.
What’s next? What will 2026 bring for L&D? We don’t know.
But I want humans to remain curious and understand that knowledge alone is not enough to change behavior. AI is complex (or at least complex). Humans are more complex. That’s why our relationship is complicated. Unfortunately, humans still have to spend a lot of effort digging deeper than the surface of shallow AI and endless money-making prompts from self-proclaimed experts. To learn more about the framework and data-driven research we employ, check out the Endeavor report. [3].
AI can and will make mistakes. The ball is in your court. You have the right to remain silent about the use of AI, but anything you take for granted without verifying can and will be used against you in L&D courts.
References:
[1] It’s About Time: 2025 First Officer Utilization Report
[2] 2025 Workplace Learning Report: Why Being a Career Champion Wins
[3] Endeavor Report 2.0: State of Applied Workforce Solutions
