AI in Learning and Development with Beyond the Sky: Custom Learning Solutions

About my guest:

Danielle Wallace is the Chief Learning Strategist at Beyond the Sky Custom Learning, an organization that ​​provides custom corporate training solutions designed to be truly effective. Beyond the Sky has over a decade of experience delivering quality, on-time, and on-budget learning solutions, and their services have been provided to many of the top companies across the globe.

More about Danielle Wallace

More about Beyond the Sky: Custom Learning

Learn more about the Three Levels of AI Use in L&D (beyondthesky.ca)

Other related articles 3 Levels of AI Integration | ATD

AI + L&D Community of Practice AI + Learning & Development Community of Practice | Groups | LinkedIn

Navigating the Maze of Learning Management Systems Free Report
Navigating the Maze of LMS – White Paper

Uncover the perfect Learning Management System (LMS) with our comprehensive guide. Simplify the process and make an informed choice by downloading our FREE whitepaper, “Navigating the Maze of Learning Management Systems: A Comprehensive Guide to Choosing the Right Platform.”

Transcript of the Interview

{Edited for Ease of Reading}

Petra Mayer 0:06

Hello and welcome to our discussion on Artificial Intelligence in Learning and Development. I’m Petra Mayer, the founder of Petra Mayer & Associates Consulting, a Learning and Development consultancy in British Columbia, Canada. And today we’re joined by Danielle Wallace. Danielle is the Chief Learning Strategist at Beyond the Sky: Custom Learning, an organization that provides custom corporate training solutions designed to be truly effective. Beyond the Sky has over a decade of experience delivering quality on time and on budget, learning solutions, and their services have been provided to many of the top companies across the globe. Welcome, Danielle. So glad you’re here.

Danielle Wallace  0:52 

So lovely to be here. Thank you.

Petra Mayer 2  0:55 

Now, Danielle, could you start by sharing a little bit about your journey? How did you get started in learning and development? And how have you integrated AI into your work?

Danielle Wallace  1:04 

Yeah, so I started my journey within the world of marketing. So, I was a marketer, getting people to change their behaviour. And we had at our disposal a great marketing budget and were always at the forefront of using new technology. And I saw there was an eight-year gap it felt between the marketing world and then learning and development. And I really am committed to ensuring that perhaps learning development is not eight years behind from an AI standpoint. So that’s my foray. And I’m just very passionate about this topic.

Petra Mayer  1:44 

And so how have you started integrating it? I mean, that’s kind of a great lofty goal. But when it comes to the actual day-to-day stuff, how are you integrating AI in L&D?

Danielle Wallace  1:56 

Yeah, it’s pretty fun. Because again, if we borrow from the world of marketing and look at learning and development, there are things that we can do on a base level set. So, I have a framework that actually goes through this. The first part of that framework is really just thinking about looking into what marketers doing, which is, okay, how do we speed up an automated workflow? Okay, how in learning and development do we speed up and automate our workflow? I think we have, as an industry in learning development, a true opportunity to really do something that’s actually effective and breakthrough and really changes learners behaviour. We collectively as an industry, we’re acquiring new skills. I think we now have the ability to use technology for that.

Petra Mayer  2:49 

Okay. And so when you’re saying, How can we speed up and automate our processes? So specifically for learning and development? What processes come to mind for you first where we can actually utilize AI?

Danielle Wallace  3:05 

Yeah, so I’ll give you the framework. At a glance, the idea of the framework is that this first level one is all about doing what we currently do in learning and development, but doing it faster. Level two is about actually doing things, different things that we can’t do right now at scale. So, using AI in new and novel ways. And then the third level is just a seamless integration. It’s a bit harder to imagine it and this is why we’ll start off at the first level, the level one. It’s all about doing what we do now, faster. For example, perhaps we write quiz questions now. Okay, now we have an opportunity to write quiz questions faster. Perhaps now we write scripts, okay, now we have the opportunity to write scripts faster. Or we actually are creating a video. So we can use generative AI from image creation or you can actually use stock image creation, we can create images quickly, we can create videos quickly from a prompt. A caveat, they should still be effective. That’s a whole other realm. But these are things we can do now. What are, Petra, what are some of the things that you might do within a learning piece or within a consulting piece?

Petra Mayer  4:30 

Yeah, so as I’m listening, I’m thinking about a client we have at the moment where we’re looking at a learning management system and a change of learning management system. And what you’re talking about really is something that would impact this client tremendously. They have a lot of manual processes, manual processes of enrolment manual course creation processes, quiz questions, you hit the spot there, they were talking about that. You know, even understanding what are the differences of their client’s needs with the customization of content. So, in all those areas, we are actually also looking at how can AI support them. And ideally, how can that be integrated into their learning management system going forward, rather than, you know, stitching together multiple solutions? So that’s an exciting project that we’re doing right now in this area.

Danielle Wallace  5:26 

It’s so exciting that it actually enables a more efficient and effective way. I’ll share my screen to illustrate those levels. So, it’s all within this first level, things that we might do faster. Even building upon that, it could be generating questions to ask a client within a consulting engagement, I can use generative AI for that, as well as creating a persona faster, great, I can use AI for brainstorming, and creating slides. There are a number of tools that do that as well. I do want to caveat that we need to make sure these are all effective. But there’s so much rich realm that we can be looking at from using, especially generative AI, which many of these examples are. We can use this now. Let alone, tapping into some more examples that you have, but using AI. There’s so much richness right now that I do encourage everybody listening to just try. Play, test, try. Try to use custom images from Dall-E or Imagery. It’s faster, it can be more effective when used appropriately. So why not?

Petra Mayer  6:42 

And I think you’ve mentioned a few times, we have to be sure that the effectiveness isn’t going to be impacted in a negative way. And I think this may be speaking to some of the big fears of AI. Oh, my God, AI is coming in and taking my job. So, talk to that one. For those who may be listening and like, oh, I really, I really don’t like what I’m hearing because right now I’m the one writing questions. I’m the one who is creating content. I’m the one who is maybe working with a designer to create the images. And now a system comes in and does it all for me. And what about my job? Talk to that.

Danielle Wallace  7:20 

So, there’s this huge fallacy that’s happening right now in the industry that I’m seeing where businesses are like great, I use AI and need to do that faster. Great. The vendor said it only takes me six minutes to create eLearning, I can create a simulation, I can create a quiz, I can create a video in 30 seconds. So, this yardstick is time. Who cares? Who cares if it’s quick if it’s not effective? And that’s where our role as learning development professionals comes into play. Because we are the ones that need to validate to make sure it’s effective, to actually do heavy manipulation to make sure that whatever that solution is effective, it’s not just creating slides written, we’ve got training. It’s dreadful, dreadful. We all as learning and development professionals, we know that it isn’t about just creating a slide or creating a talking head video that makes for effective learning. So, we have a lot of richness in our role as learning and development professionals to guide the business. So, hand in hand, we are actually creating something that can be more efficient. But it’s not sacrificing any of that effectiveness.

Petra Mayer  8:31 

Even potentially driving more of it. And having more ability to do more personalized learning, more customized learning, and be much more specific to the learner as well. And to the learners responses, something that would be very often not possible, budget-wise, time-wise, resource-wise, not possible in organizations to bring in that sort of learning experience that now maybe comes into the area of possibilities. So, as we move to level two, what are some of the learning experiences that AI enables, which were previously impossible with the older technology? Give us some examples.

Danielle Wallace  9:32 

Yeah, exactly. And again, you bridge that so nicely from what we could be doing within the world of AI now, but actually diving into what’s not possible with technology today. So on this level two and this framework’s available on my website, it’s having the learner interact with AI themselves. So it could be things such as a personalized chat bot. It’s really the idea of creating at scale what we can’t create right now. So, a good example of that would be an eLearning course. Imagine, Petra, an eLearning course where instead of having a scenario with three quiz responses, great, we’ll choose this, these responses to a role play. What if it was 100% personalized, what if whatever you wrote in, you got actual feedback on that, or actual coaching advice like for real, like, based on you. That’s what’s possible. Now with AI. To help illustrate that, I can share an example that might bring this to life a bit more. So this is an eLearning course, that I’ll be sharing my screen. And in particular, what it is, is a role play, but in a very safe a safe format. So imagine that you have a direct, you’re a manager, you have a direct report, who wants to take time off over the holidays, but you’ve already set down the policy in the policies you need to have submitted like three months ago. But what do you say when you say in this situation, you need to respond with consistency to the policy, and you need to still show empathy and flexibility? So Petra, what might you say?

Petra Mayer  11:14 

I hear that you have found out about a wonderful deal for the holidays. As you’re aware, our policy is that you have to pre-book these vacation times. I want to be responsive to your desire. So let’s check in with the team if this will be acceptable for everybody else.

Danielle Wallace  12:00 

Nice, and what’s key here is whatever you’re typing in is what it’s going to be. Yeah. And with generative AI, this allows us now to have truly personalized feedback. So it says your response is empathetic and shows that you’re considering the team’s needs as well. However, it may be more effective to clarify, to clearly communicate the company policy and the reasons behind it to the team member. That will help set expectations and avoid misunderstanding. Yeah, I wondered if you might actually continue to do this? Do you want to try another one?

Petra Mayer  12:37 

Sure, why not?

Danielle Wallace  12:39 

So now imagine that your employee comes to you saying my brother recently had a stroke and needs full-time care as he recovers. His wife’s out of town for two weeks. She’s asked me to care for him while she’s away. But lives in another city. Can I please get the time I need to look after him. So you do need to consider responding with empathy, flexibility here.

Petra Mayer  13:01 

So I’ll make up a policy. I am so sorry to hear about your brother and the impact it has on the family. I fully understand that you want to support him. Let’s say the policy is that they she’s out of vacation just to set the scene here. Our policy is that we cannot grant more vacation time then you then is assigned. However, we may be able to provide you this with unpaid time off. So you can support your brother and your sister in law during that time in need. Now again, didn’t say why we have this policy. The policy is to set expectations. Okay, we’re now I’m already being coached on this.

Danielle Wallace  14:31 

See, look, it’s fast.

Petra Mayer 14:34

Super fast.

Danielle Wallace 14:38

Your response effectively communicated the company policy while also showing empathy towards your team members’ situation, by offering unpaid time as an alternative you demonstrate flexibility and a willingness to come in to meet needs, which strikes a good balance of being understanding and adhering to the company policy. That’s actually really good. And it allows us a way that people are actually learning by doing so while I could have just taught these, and had a quiz question, which is irrelevant. Who cares? Or I just taught these and then had a multiple choice. Great. You know, like, who cares? This, Petra, got you to think as a learner, it got you to feel, it got you to practice difficult conversations. It’s fun.

Petra Mayer  15:17 

This is this is, I’m assuming, this is something that you’ve built in your organization, and that is something that you incorporate into your clients learning activities. Is that right?

Danielle Wallace 15:31

Yeah. When it when it’s the right discretion.

Petra Mayer 15:41

Yeah, very impressive. Okay. This goes to level two, because that wasn’t necessarily possible with previous technology. An organization may have substituted that with coaching sessions, etc. But that’s very pricey. Not everybody can go that way, or not have as much coaching as people could possibly use. So now let’s look at your level three. And when we’re coming to level three, AI promises seamless and predictive integration in learning and development. What does this look like in practice? And how does it fundamentally change the way organizations approach L&D?

Danielle Wallace  16:12 

Yeah, so the idea of level three, which is a seamless integration into the flow of work, this is exciting, scary, interesting, all at the same time, it is yet to be imagined. But if you can think about it purely from an objective point of view, imagine if your friend Eric, a learner said, let’s not put ourselves in, let’s just imagine someone else, Eric, comes to work. He knows he needs to have a difficult conversation with somebody. And imagine if he’s getting like real-time, before he needs it, prompts on what he should be prepping before the meeting. That would be, I’m just thinking about that last example, that’d be helpful if he knew, Oh, I’ve got a meeting coming up, I better prepare this. And consider maybe Eric, you should review the company policy, you should respond with empathy. Here’s some good ways to respond with empathy. Eric, in the past, you’ve not responded with empathy. Some people just aren’t empathetic. Here’s some things you could say to sound more empathetic. And remember, this is some of the guidelines. Like what if that was fed automatically to you before the meeting started? What if, when you’re using new software, that there was no training needed? Because there’s contextual help in there? Before you came to it. Before you needed it? What if you had this digital and virtual companion, this digital virtual system, it was always there feeding you what you need it exactly when you need it. So, imagine you’re going for a run. You know, I could hire myself a coach, I run, I could hire myself a coach, or if I had this virtual, visual coach. So it feels like I’m running with somebody, because I like companionship, and they were encouraging me, personally, that’s what I need. I need encouragement. So it would automatically do that for me. And it would it would tell me to go faster and keep pace, which is like what I need like it just what I mean, needs very different from somebody else. Imagine that imagine everything’s proactively and seamlessly there. What this leads to, though, is learning and development is just really redefined. So the role of a learning development professional is still there, just in a different capacity.

Petra Mayer  18:34 

Now, while the potential of AI is obviously vast, you have a lot of kind of great ideas and imaginations and what could this look like in the future? What are some of the key challenges and ethical considerations that organizations need to be aware of when integrating AI across these three levels?

Danielle Wallace  18:56 

Yeah, I’d say two big considerations, especially one right now, which is governance policies, many organizations are still grappling with creating an effective governance policy, keeping up to date is going to be a challenge. Like if organizations aren’t thinking about that regularity of that that’s a huge challenge. Even just getting one aligned is difficult. Equal Access is difficult. And then even once there’s governance policies and equal access, things like Co-Pilot rolled out, the employees aren’t even using it anyhow. You got this very uneven use of a AI, which both prohibits use, it stifles use, and people aren’t motivated to use it anyhow. So that’s an issue that will affect learning and development. It affects organizations at large but I think has a particular impact on learning and development going forward, which does help mitigate some of the other things but be mindful of the risks. So when we’re using generative AI, I’ll just speak to learning and development, when using generative AI to create scenarios, for example, the scenarios get actually really generic, which is an issue like overly generic. And you need to be mindful of bias. I was creating – there was a like a little hackathon at the ATD TK conference. And I really, really, really wanted to create an attractive, ethnically ambiguous person. For one of the dating challenges. I really wanted to be ethically ambiguous that was actually important. Couldn’t do it couldn’t do. It always came up with a Caucasian male that always had the beard. You try not doing a beard, you still got the beard. This was supposed to an older person, and it was mutually exclusive to have attractive and older, like, really?

Petra Mayer  20:52 

Yeah. So those are challenges that maybe are because they’re on the path? We’re still in development. Yeah, this isn’t, we’re not at the end game of what technology can do. And I think the learning, talking about machine learning, there’s the learning is still happening. Now, I’m curious about what you said about kind of more the ethical challenges. I mean, and understand even here, we’re talking about accessibility, we’re talking about inclusiveness of the solutions. I hear you there with your challenges, but even you know, setting up the right policies, what can what can they not do in in the workplace? How do you suggest organizations should approach this? Get started with it?

Danielle Wallace  21:43 

Yeah, so in general, the governance policies is you can be championed by learning development is amazing to have learning development, creating your own table, bringing others to the table, it does need other players. So you have to ensure that legal, finance, operations, all the business units are represented. Marketing, communications, that’s, that’s table stakes for this, this table we’re creating. And inherent in that it becomes a process. So I fully recognize the the time requirements for that. But it is imperative that all the different groups do align. And there is consensus at that. And then secondly, so that’s the overall governance policy, which often is maintained, held, kept through compliance, meet useful to get it out of compliance as well to make sure everybody has a stake in the in the policy, then typically, the second thing that happens is organizations and that tool, so they would say yes, we can use this, no, we can’t use that. So this is the process to get approved, a new system approved, which everybody knows, takes like a year. It’s a long process. So within this amazing governance process that is so essential. I do have this caveat, I say of this theory, that it’s too late. So both too late, because your employees are getting around and they’re using generative AI, those who want to be using it, and then it’s too late because those who should be using it, once you roll it to policies, will not. It’s too much effort. It’s not they don’t they don’t see the reason why, and especially within learning and development, I definitely see that they were not.

Petra Mayer  23:36 

So it could be a generational shift. You know, I’m assuming that younger generations are coming in, they’re more open to using it, or am I biased?

Danielle Wallace  23:47 

I don’t think it’s a generational, I think it’s just simply a matter of great governance at the outset. So like, you know, a year and a half ago, two years ago, being stopped. Stop using AI, because it’s not secure. And yeah, you can’t use AI for your work like, of course you can’t. But so many organizations have not adequately rolled out a governance policy in time. And I just mentioned it takes a lot of time. That’s like a catch 22. So I don’t think people are using it in their workplace as much as they could be, even when it’s sanctioned to allow the efficiency. So even that level one I spoke of, and that’s all I’m speaking about. From an organization standpoint, I think we have we have a both an impetus to help ensure our employees are actually using AI effectively, right.

Petra Mayer 24:16

And stay within the boundaries that we’re setting. And I think that’s also an interesting one where I feel larger organizations and smaller organizations are very different and very different situations of putting the resource behind it, putting policies behind it, but also maybe creating a close circle of information that smaller organizations are probably less likely able to do. So having access to kind of a system where we’re really using our own internal data. And we’re not spreading things out into the bigger, wider web universe. But lots of challenges still ahead. And also a lot of opportunities. Danielle, thank you so much for being with us today, and sharing your insights, sharing your examples, that was very impressive, you really did put me on the spot here. I started sweating a little bit.

Danielle 26:17

We’re mimicking real life.

Petra Mayer 26:30

I loved it. So thank you so much for sharing that with us sharing your your three layers, because I think that really helps people to structure and think about AI in very practical terms, and then start having those internal discussions if they’re not even if they’re not having if they haven’t started yet. So I really appreciate you coming to the microphone today for us. And I look forward to having another conversation with you because obviously there is so much more to dig into.

About the author 

Vraya Forrest

>