Introduction

Hi, everyone. My name is Egle Vinauskaute. I'm a learning and technology consultant, and AI is kind of a big deal, kind of a big technology in learning.

Can I ask, before I go into this, how many of you work in L&D internally? No, not a single one. Interesting, kind of.

I know that we have some founders. Who are everyone else? Are you supporting founders, advisors, that sort of capacity? OK.

The Most Common Questions about AI in L&D

Despite all of these conversations and all the questions that could have been asked in the past year about AI, this is the question that I get most frequently, which is, what are the others doing with AI in L&D closely followed by, am I falling behind? So what I'm going to attempt to do in the next, hopefully, 10 minutes is give you an answer to this particular question.

Recent Research on AI Utilization in L&D

Donald Taylor and I, we now did another piece of research released in April, where we asked various people working in L&D, both internal and external, probably 50% of responses from the US and the UK, and we asked them how they are using AI in L&D. And would anyone venture a guess about the top three uses of AI in L&D?

1Top Three Uses of AI in L&D

We have content. We have using AI to perform learning design tasks, creating learning content, and researching subject matter, all falling within content, which is not necessarily bad. It's obviously a natural use case for generative AI.

The Gap in AI Usage for Skills-Based Organizations

But once we actually looked into, if you've been around in L&D for the last several years, there's been this big conversation about SBO, skills-based organizations, about data-informed decision-making. So where are we at with this? Because AI potentially can give us the technology to power all of that as more targeted learning interventions and L&D supporting the business.

So when we looked into that, we realized that these use cases are at the bottom. where we have qualitative data analysis, skills practice, performance support, and all of these other skills-related management technologies. And then in the middle we have a bit of a mixed bag with some administrative tasks, learning personalization here, I mean adaptive learning and content curation, things like that. using AI for user research and translations, another big use case of AI in L&D.

The Shift from Content Creation to Skills Practice

So the first one is the difference between creating learning content and skills practice. Because historically in L&D, we have used content to pretty much do anything. And the thing is that content is not the optimal medium for a lot of these cases.

If you want a resource at a point of need, here is a course. Here is content. If you want to upskill yourself, here is content.

What skills practice can do, and here I mean conversational bots, mock conversations, coaching bots, things like that. What they can do is that they can potentially take a lot of heavy lifting off of content. So for example,

We would previously create a course about how to give feedback to people, and that would be a course, perhaps a video, here are the tips, and here are some examples of how this is done, and now here we go. The problem is that with these conversational skills, any interpersonal skills really.

It's not just about the skills, it's about how I feel about it, what my confidence level is. So you need to practice, you need deliberate practice. And AI enables that deliberate practice potentially and hopefully we'll see

skills practice catching up to content because right now we are just speeding ahead with content creation for anything, mostly chasing the cost savings off. We have always created courses, now we can create them faster rather than rethinking what else, what kind of capabilities does AI give us that we can do this in a more effective way.

Performance Support as the Goal for L&D

Now, here at the bottom we have this little performance support, which in my opinion is a bit of a holy grail for L&D, because it is about supporting people, giving people the content or whatever support they need, when they need it, in their context, in the format that they need it, and AI potentially in the format of co-pilots, of various assistants. AI assistance, it can finally help us potentially get there.

However, it is quite low on the list and there are a few reasons for that. One is that it's still a technological challenge because if we are replacing something like, for example, compliance training with One of the pilots that is quite common right now among L&D leaders is using some performance support solution, meaning a bot that can give you some contextual answers in the onboarding process, therefore cutting down the time it takes to get people up to performance.

1The problem there is that AI needs to be in these cases 100% reliable. If it is supposed to tell you what procedure, escalation pathway or whatever to use, it cannot say anything that is not exactly what you need to hear because there are various a lot of them legal implications here.

And it is the technological challenge to figure out how to make this thing usable, but also accurate. 1And the other side of the coin is the fact that it is not L&D that makes a lot of these purchase decisions, because if you want to provide people with meaningful performance support, you need to plug AI into the company's knowledge base, and that is not an L&D decision to make.

So generally speaking, looking at all of these use cases, quite low which are potentially some of the more sophisticated, some of the more powerful use cases of AI and L&D, why are they so low?

Understanding the AI and L&D Complexity Scale

So Don and I asked this question, looked at all of the interviews we had, surveys that we carried out, and what we realized is that there is something that I like to call the AI and L&D complexity scale, where on the left hand, we have simple uses of AI where L&D can use AI autonomously and on the right hand we have sophisticated uses of 1AI where L&D needs extra skills, technology and relationships to make them happen.

So if we plot our most common uses of AI AI in L&D, we get something like this, where on the left-hand side, near the simple end of the scale, we have administrative tasks, we have learning design and content creation tasks, and then we go a bit to the right and then we have skills practice, right? So we have role-playing chatbots, we have simulations, we have AI-supported coaching.

But to make that happen, you either, in L&D, need someone who likes tinkering and can start piloting these things and building building them, or if you're buying third-party, then there is a bit of a rollout, which is, okay, so we need now to partner with another part, perhaps some client-facing business, for them to start using this bot and see how it works, and then create this whole feedback mechanism to measure if it is effective, if it needs some fine-tuning, and so on. So it is not rocket science, but it is outside of the let's say regular skill set of L&D where they need a bit of support here.

We have data analysis and there you need to interact with the rest of the business to actually get that data and perhaps, well, for sure a lot of L&D departments do not have people analytics people in their teams and therefore they may have to borrow them or bring them in from the outside to do that analysis. So again, going a bit further outside of L&D.

Then we go to this performance support issue that I already talked about. We have search assistants, we have co-pilots, and here, as I mentioned, it is not L&D that often makes the decision to deploy, say, Microsoft co-pilot and the organization. It is a strategic business-level decision. And a lot of the time, if they want to plug in something more specifically for learning, it's still they need to interact with the business, with AT Legal and InfoSec and various teams within the business to make that happen.

And then once we go to the far right end of the scale, we have personalization and skills management, so skills intelligence, workforce planning, internal mobility platforms. And this is where essentially L&D can be viewed not as the not as the champion of this thing. It is a stakeholder. It is not the function of the business that actually makes that decision that, okay, we're now going to become a skills-based organization. L&D is a stakeholder that helps support that business strategy. And therefore, if L&D wants to go towards the right, it needs those extra skills, technology, and relationships. to achieve these more sophisticated and arguably powerful use cases of AI in L&D.

Key Lessons from Organizations Using AI in L&D

And I want to leave you with a few lessons that we learned from the organizations that we spoke with. The first one is that business goal equals business alignment. What became really clear is that in the organizations that actually had some case studies to share with us, and there were not very many, we actually reached out to as many people as we could, and out of the ones that actually had something, and there is a smaller percentage of the organizations that could get the case studies through PR so we could publish them, one thing they had in common is that they, ahead of that whatever AI implementation they ended up doing, they put together a business case and show the business that this is gonna either save money, make money, make people more productive, creative, whatever. Whatever the business goal is, they manage to make a business case that this is what we're gonna do with this. give us a mandate. And once they got that business goal, that led to business alignment and that led to mandate, which led to resource, meaning money to buy even simple things like licenses, time to experiment, and this general fostering of the culture that allowed people to start using AI in their work.

The second lesson is be creative with what you've got. And it is quite clear that in Because AI use cases are so individual for various companies and so contextual, it is highly likely that you're not gonna find an AI tool. I'm not talking necessarily to IGPT, I'm talking about more tailor-made tools for your specific use case. And therefore, what I often tell organizations is that look at what you have already got and see how you can stitch these things together for them to work for your use case. So specifically, taking stock of what you already have in the organization. There are very many different tools that are implementing various AIs within those tools, such as project management, marketing and things like that. So see what you already have access to.

And the second one is look into whether your vendors can flex and adapt whatever AI tool they're creating for your use case. Because one thing that we have noticed is that because learning tech vendors are creating essentially new products with AI, they need use cases. And if they can get a big organization to pilot their product, they are sometimes willing to actually create something quite custom and essentially work as an extended tech team for that L&D.

And finally, use cases emerge from experimentation, and it is important that L&D creates, not just L&D, but the organization creates an environment for it because it is not a top-down kind of thing with AI. It is about enabling grassroots innovation so that people can create the case studies and they can surface the case studies to make the best out of AI.

Conclusion and Invitation to Read the Full Report

So yeah, this is the report. If you want to read it in full with all the graphs and case studies that we collected with AI and L&D, we have seven case studies in there. So as far as I know, the biggest collection of AI and L&D case studies.

Yep, phone's down. No, still up.

And that's pretty much it. That's me.