In Charter’s recent playbook, “The AI Educator: Using artificial intelligence to transform manager development,” we cover strategies and tools for adopting AI to promote mastery-based learning among workplace leaders, preparing them for the complex set of challenges facing middle managers today.
Egle Vinauskaite is the co-founder of AI learning consultancy Nodes. Recently, Vinauskaite released a white paper with co-author Donald H. Taylor on AI adoption in learning and development (L&D), based on a survey of over 300 respondents and case studies from global companies including Bayer, Ericsson, Leyton, and HSBC.
We reached out to Vinauskaite to learn more about the report’s findings, as well as understand the most promising use cases for AI in workplace learning. Here is an excerpt from our conversation, edited for length and clarity:
In your recent white paper, you talked to workers and leaders about how they’ve been using AI in their jobs to learn on the job. What did you find were the most common use cases?
The most common use cases when it comes to L&D are, quite unsurprisingly, content related. The top three were content creation, learning design, and researching topics. Translation was another use case that wasn’t in the top three, but it was still very popular. None of these use cases are particularly surprising. Usually, when you find use cases for AI and especially generative AI, you map out your process and find the parts of it that are time consuming and resource consuming.
Content creation has historically taken huge chunks of L&D’s time. The thing is, when you think about using AI for learning, we have generally used content in all sorts of learning, from workplace L&D to higher education and so on. In the workplace, we have used content as a proxy for skill development and performance support. If you think about the content that you create in an organization, you want people to either on one hand develop skills. For example, I don’t know how to give feedback or do job interviews, and you give me some content so that I am able to learn to do that.
And on the other hand of the spectrum, you have performance support, meaning that this is not a skill that I need to learn, but something that I just need to be able to do and get on with my work. So in real life, that might be my computer is broken, so I Google ‘why is it making that noise?’ That would be performance support because I don’t need to learn how to fix computers in the workplace. Other examples might be related to some admin tasks, like accessing a database or how to ask for a vacation day. It doesn’t necessarily require learning.
What is the potential role for AI in skill development and performance support?
AI gives us the opportunity to move away from content because it can do certain things that content doesn’t need to do anymore. Before generative AI, you would use content for the entire spectrum, meaning that if you need to learn to do interviews, here’s some content about that. Here’s a video, or here’s an article for you to read. If you need some help with your onboarding as a new hire, here’s a course with everything you need to know.
This has not been great because for skill development, content is just the first step. You need to practice the skill to get feedback on it, to reflect, and to continue this process over and over and over again until you gain mastery and you’re able to perform it. Content has been an out because me giving you some interview questions to be a hiring manager does not make you really good if there is no follow up on that. On the other hand, with performance support, workers usually Google these questions. You don’t go to the company’s learning management system (LMS) to find the course about the thing that you need support with. So content that used to be created just generally wasn’t really serving either need very well.
What AI does is, instead of dumping a lot of content on people, now we can give them actual practice opportunities. For example, creating a little coach buddy that helps me think through certain decisions or helps me do difficult conversation practice. For example, if I’m a salesperson, you might teach the bot to play my difficult customer, and I’m trying to resolve their queries. So AI now can potentially do a lot of heavy lifting here on the skill side.
On the performance support side, with a copilot, you can actually do performance support where, much like with chat GPT, I can ask anything that’s on my mind and it gives me information on the company’s internal information. You can get these answers at the point of need, which is the real performance support.
Are there any particularly exciting case studies that you found in your research and reporting that relate to manager training?
One case study was about capturing and making useful in-person training events from a global events company. They have primarily in-person training and ran a pilot for their senior leaders. Usually what happens with in-person training is that you can have an amazing facilitator, but once the training is over, it’s done. The content is gone, people go back to their jobs, and learning transfer is questionable. It’s a good time while it lasts, and that’s it.
They decided to use an AI to capture the in-person, live event where people sit in the classroom. They configured the AI to, after the event, send each managing director a summary of what was discussed and a list of action items. Their managers also get a summary so that they can pick up some of the themes that would come out of these discussions. That way, themes are not just staying there in the room, but they become something that can actually be brought into the real world.
They also started building a knowledge base with the captured content that is being transmitted in live events, which becomes part of the corpus. You can actually, just like with ChatGPT, start interacting with it. You can start asking it questions, from simple ones such as, ‘What was the sentiment in the room? What were the people’s main concerns?’ But you can also ask it to do more complicated tasks, like an in-depth analysis of the conversation. This is where the case study stopped because it was quite new, but you can extrapolate and imagine where that might lead. You can then reuse the content and perhaps the questions that people asked to create different courses or nudges for people in their jobs. It actually makes that content useful, both searchable and something that lives on, something you can iterate on and continue to create learning from.
Download Charter’s full playbook for more case studies, frameworks, and advice on implementing AI into manager learning and development. Thank you to Valence, the sponsor of this playbook, for making this work possible.