Transcript

AI in Workplace Learning

Watch Video

Liam 

Hello, good afternoon and welcome to our webinar here today. My name Liam O’Meara and I am the chair for today’s session. So we’ve got a very interesting session here for you today and one warning, it is going to be interactive. So, I see a number of you have already inputted your name, your location, and the weather. And we’ve got a very international group which is fantastic to see, quite a mixer of weather. What we would like for you to do is to be able to use that chat functionality throughout the whole session, so as we are making comment if you agree or disagree that would be fantastic to know or if you have questions or comments about anything that is being discussed. Please do feel free to use the chatbox and that is what we are going to do to help tailor out session, best we can.

Liam 

So, the session here today is all about tackling where we are in regards to AI in workplace learning. What is real and what is hype with regards to workplace AI in workplace learning. So I have a couple of experts here to help me with the session. Firstly we have Donald Taylor from the LPI institute. Donald perhaps you would could say a few words.

Donald Taylor 

Thank you very much, Liam. Great to be here and yep I am Donald Taylor. I am the chairman of the Learning and Performance Institute right here in the learning technologies conference in the UK, also chaired at Singapore in Berlin.

Donald Taylor 

Very interesting topic of IA, I am not going to spoil the answer to the question is it real, is it hype. [crosstalk 00:01:52]. At the end, over to you JD.

JD Dillon

Hey everybody I am JD. I am the Chief Learning Architect at Axonify. I have never been the chairman of anything, so Don has that covered. I have spent the last 15, 20 years or so in learning operations type roles with large enterprises most notably the Walt Disney company and Kaplan Higher Education, so that is where my perspective I today’s conversation comes from.

Liam 

Fantastic, so two absolute industry experts with us. It is why I promoted myself today’s staging, that is why I call myself the chairman. To say that I can call myself a chairman for a particular event. Just a quick note, it does always come up in questions, the slides will available after the session and we will send out records as well. So do not worry you do not have to scribe away with the notes, we will send these out afterwards also.

Liam 

So I thought the first place to start is with some terminology, there is lots of different words and acronyms flying around when it comes to sort of IA. So, JD perhaps you could help us out by setting down a sort of foundation that we can use for today.

JD Dillon

So I think when we start talking out things like IA and how it relates to the work that we do within organizations it is important to make sure we have a common language and understanding of some of the basics because I do not think that I am the first to say that, we as learning professionals or HR professionals are not great at defining things. Because if you, I still challenge everyone to tell me what eLearning is. I cannot find a standard definition of that. Donald if you agree with me on this.

JD Dillon

AI for me is different, when we look at in terms of application to learning and development, because this is a discipline with decades of background and expertise and people who are very good and very smart at what they do. So it is not something that we are creating for ourselves, by any means, this is something that we are looking how it is influences our workplaces and how could we adopt some of the ideas.

JD Dillon     

So when I threw these two definitions on the slide, I think these are kind of textbook definitions of terms that I think are unfortunately used interchangeably in a lot of chases when machine learning is actually a subset of the larger AI conversation, but I think the overall meaning here is that is important that we do our homework and we understand what this topic really is before we start leaping into how it impacts us.

Donald Taylor 

I think it is also important, JD, to bear in mind that there is an awful lot of excitement naturally around AI, but there is also a lot of confusion and misconceptions and much of it steeped in history, my daughter is studying Frankenstein for exams at the moment and this idea that you could create an intelligent form that has human like intelligence is of course very ancient but it does not mean that it is actually realistic.

Donald Taylor 

Artificial intelligence, the idea of creating a general human intelligence, I am not sure we are there yet. I think when we talk about AI, very often yes that is where we would like to be, but very often what we are talking about is the specific very, very narrow application of algorithms which is more than machine learning.

JD Dillon 

Yes, I think especially focusing on I think two big consideration, for me. One, the idea that we are replicating our inherit capability to solve problems, reason, and learn and doing it in ways that is not automatically pre programmed, so I think there is, we will talk a little bit about, obviously hype in this conversation, but there is a lot of people saying AI but do they really mean it? Two, I think we would be in a much better place if anyone of everyone whoever an article on this topic never used a picture of the Terminator ever again. [crosstalk 00:05:33]. Played out, that is not what this is. But the good thing I think is that we are seeing the conversation around is it going to take my job, I think that one is nicely fading down. Again coming to be to, for us, in our profession I think it is critical that we our homework and not necessarily be experts in this but really understand what this means and what these terms are before we leap into the conversation around what it could mean to us.

Liam 

Fantastic, why don’t we find out where people are today? As we sort of said this really is something that would welcome your firsthand comments, so please do respond in the chat box. How engaged are workplace learning professions in the AI conversation today? I am sure given the audience we have got here, we have got people from all around the world, from lots of different professions. So, it will be really interesting to understand actually what is happening in your workplace today in regards to AI. So I will give you a couple of seconds to just type away and then we will open up to Don and JD for their comments.

Donald Taylor 

If you just throw your answers into the chat area, JD and I can read it, we can reflect on it, at the moment I am reflecting on the weather reports from different parts of the world. That is exciting, [inaudible 00:06:48] that is exciting

Donald Taylor 

Because we really want to know is how engaged are you in the AI conversation? I guess, I know a lot of people who are, if you like, peripherally engaged, they want to know more but they are not actually engaged in the doing. So I would interested to know any thoughts people have got around that.

Liam 

We are starting to get some comments back for you. So, JD what are you seeing out there.

JD Dillon 

I think that the biggest thing for me in terms of engagement comes around the fact that there is big conversations outside of what we, right, so I think engagement in AI as related to workplace learning starts with how engaged is your organization overall and your industry in terms of leveraging this type of technology and engaging this conversation. If you see, if you look at some of the quotes on screen and the references of screen, about how many executive teams are looking at this technology as a critical part of what they are doing overall. The fact that the shift is being made, like I mentioned earlier in terms of focusing on how we are augmenting the workforce and what inherently how work is done and what it means to be human in the work place, I think are critical consideration before we jump into where our prospective is. The last one from The Verge I think is interesting given that there is so many entities out there engaging in a conversation, throwing content around around the concept of AI that don’t actually use the technology or are kind of misinterpreting what AI is and how it relates.

JD Dillon

So I think when it comes to getting engaged and I agree with Don where he said, it seems like people are very quickly starting to explore this concept in our industry, kind of understanding a good starting place and what is and is not real, I think is critical to get started.

Donald Taylor 

And, JD couple of comments came through, [inaudible 00:08:47] said that a limited engagement, interested but not well versed, I think that is pretty typical. People in AI are keen to know more but are not actually doing it yet. Pascal from Berlin comes in, I think, with a very good comment saying look “I am quite perspective, people in learning and development may be thinking about AI as the land of milk and honey.”

Donald Taylor 

I think that is so true, new technology comes along, bang it is going to solve all our problems, and it is not. But it is I think going to change things, so it may not solve our issue but i think we better get ready for the new world of L and P that it is going to generate, but I do not want to spoil your thunder JD; you crack on.

JD Dillon 

Ha no I completely agree, no thunder stolen. I mean if my drawers were full all the things that were going to change the world learning and performance I would need to buy more drawers. Remember when social learning was everything and what not, but that is where I kind of come to back to the point and I think this is an important conversation and the reason we are starting to explore this and kind of open dialogue is that, for me, there is a difference here and because it is not just about something we are deciding to do to change what we do. It is something that is fundamentally changing work and what it means to be a person in the workplace and I think that is where the conversation really begins understanding what that looks like.

Liam 

Don, you have been some sort of research in this category as well haven’t you?

Donald Taylor 

Well absolutely every year I do a survey, I call it the Learning and Development Global Sentiment Survey and here is the results for 2019. What I do is I ask people around the world and this year it was 2000 people, roughly, from around the world answered. What will be hot next year? Very much like regim and pascal were saying there is a lot of interest, in not just AI but [inaudible 00:10:42] I just want to look at these are a second and then we will go on to the next slide. We will have a look at the AI side of interest but here you can see some of what JD was talking about being reflected in the general experience. So I say what is hot; I ask people that question because I want to know what are people talking about now because typically, with the group of people I am discussing this with it becomes either standard practice or abandoned in within two to three years.

Donald Taylor 

You can see here that some things have fallen down the table, look here at the bottom of the table. Curation, is 14th it was number 9; it was number 9 for the previous three years. I think the reason for that falling down the table is, in this case people have found it a bit complicated to make curation really work. I am not saying that is not important I think it is very, very important, but it is not something people find easy to use. Video has been falling steadily down the table over the years; not because it is not important but simply because it has gone from being hot to being business as usual. But if you look, by the way the URL is at the bottom of there, is so good we printed it twice, which is the URL to download the report for free. You can download that one and all the ones from the previous years as well.

Donald Taylor

Let go to the next slide Liam because if you look at the top three. The top three options this year for the first time in six years are all about one thing. They are all about data; personalization/adaptive delivery, artificial intelligence, and learning analytics. These are things that about a third of people that I surveyed across the world, the results come from 92 countries, but a third of these people voted for one of these or more than one. They are all driven by data, you could easily argue by the way that personalization and adaptive delivery is driven by even artificial intelligence but it is seen as being the end which you get to by the means of AI.

Donald Taylor 

So there is very, very strong interest in this and AI has moved up from third position last year to second position this year. It is very unusual to have something move up the table from third position because there is a lot of votes there, it is very difficult to get that much support. So in other worlds what we are saying is around the world bags of interest in artificial intelligence just as our audience has indicated.

Donald Taylor     

The other bit of research I did, we will come back to this later on, is around the capability, the skills of the learning and development profession. So, the institute that I chair, the Learning and Performance Institute has this map, this road map of 25 skills. We spread those across four levels and those skills are both traditional, if you look at the one at the bottom there; Facilitate Face to Face Learning; that is classroom delivery, that is something where we all come from and there is a lot of other skills there which we need for the future including things like; on the right hand side under performance and impact, performance consulting and data analytics. These are things which are crucial for the 21st century learning and development department, having this set of skills.

Donald Taylor

We will come back later on in the presentation to see how well we are actually doing against this set of skills. This has been in place for six years; we revamped it last year, the results are fascinating. Again the URL there at the bottom is where you can go and do a free, you can test yourself free against this. Not test yourself, sorry, you can go and self assess against the framework for free using that URL if you choose to. We will come back to this later on to see, well okay what are the top skills the industry has got, what are the skills where we do not reckon we are so good, and against the context of what we are talking about today it is going to be important.

Donald Taylor 

Let us go on to the next question.

Liam 

Fantastic, thanks Don. So, again we would be really interesting in gathering your thoughts on the next questions. We got some fantastic questions comments coming through last time so please do feel free to chat away in the box. So question number 2 to the audience and to yourself and Don, you are not going to escape that easy.

Liam 

What are we getting right so far with regards to AI in workplace learning? And what are we getting wrong? So, you have already talked Don about some of the Frankenstein types idea coming out there and some of the things that we have seen in the [inaudible 00:15:07] business where AI is clearing making a difference but there is some skepticism about what is working and what is not working.

Donald Taylor 

I am going to hold back on any specifics for just the moment. I would like the audience to come up with their thoughts, but listen and I do not want to lead anybody but I will say this we are getting a lot of things right I am seeing some interesting stuff done particular by start ups in this area because they are small enough to be quite nibble about it. What we are getting wrong I would say is being confused by what we mean by AI, and also imaging the consumer experience that have with AI is sort of the right thing to be doing in the learning field.

Donald Taylor 

So when people talk about the Netflix of learning, well I am very skeptical about that. We know that there are some strong algorithms behind Netflix that compare millions bits of experience by consumers of Netflix and then use that to reflect what you might want to do. Fine, but I do not think that is necessarily the right approach for learning. Learning is not about entertainment and it is not about what you might fancy doing, it is about what will help you [inaudible 00:16:17] consciously or unconsciously. I think there may be things that you do not that you do not know and in that case a Netflix of learning ain’t going to help you.

Donald Taylor 

Pascal as said look what we are getting right is that AI doubles the future of life and that causes major impact on learning; I am with Pascal on that one. That we definitely in a world where we have to recognize the impact that AI is going to have. JD any thoughts that you have got?

JD Dillon 

So wait a minute, you are telling me that technology built based on the value proposition of keeping me watching something, whatever the system wants me to watch, as long as possible is not an applicable discussion?

Donald Taylor

Haha, yeah, I hope we are on the same page with this one. Yes the goal of Netflix is not the goal of a learning and development, absolutely.

JD Dillon 

Agreed.

Donald Taylor 

What are we getting right JD?

JD Dillon 

So I thing for my perspective I think that, I start on kind of the flip side of that. I think in order to get things right we have got an inherit challenge that I think is constant within our industry, which is we are very buzz driven, marketing driven profession, and again someone tell me if I am wrong. But as you see when we work our way very quickly through these various hype cycles where you know, elearning was going to save the world, then mobile learning was going to save the world, then social learning was going to save the world, and gamification was going to save the world and because of how quickly we move between conversations then the people who turn a lot of content and a lot of discussion, and I will say it out right; vendors and what not; put words on their projects.

JD Dillon 

Recently I was a very large trade show walking around and astounded by the fact of how many vendors have suddenly integrated Artificial Intelligence into their product; when like six weeks ago they did not. So, I do not understand how that moved that quickly but I think there tends to be a race to get into the conversation from all sides, whether you are a practitioner, whether you are; let us throw big air quotes around the term thought leader, a vendor because everyone one is trying to get engaged and I think a lot of people are trying to be helpful. But when you jump in without the right level of understand and the right level of established credibility, I think that is where things start to go sideways and we get very focused on the hype and the buzz around the term rather than distill it down to what is real and what the fundamental principles can be.

JD Dillon 

I mean just to make the comparison, I think mobile is a great example. Where how long have people been using mobile devices in their everyday lives and in the workplace and have we really figured out where that type of technology fits with what we do. In a lot of cases we still have not, so and that is a very example as compared to what I think this technology can do.

Donald Taylor 

Couple of points from the chat box.

Donald Taylor 

Coming from Poland is saying, we are promising too much, let us not get too excited. Yeah things will change but it will be an evolution not a revolution; Something in that I think.

Donald Taylor 

Pascal is saying look we should not just be chasing after AI for the sake of it, and of course he is right with that. I think, I always think back to second life and everyone getting so excited about it and then it is fizzling out to virtually nothing.

JD Dillon 

Aha, I am sorry I cannot keep a straight face when you say the word second life. You did not want to have a third party avatar fly around to watch a PowerPoint presentation in a virtual world. Sorry, but i think that kind of relates to my next point.

JD Dillon 

Which would be we are often fast to jump to application, right? This new thing arrives what can I build with it or how can I integrate it into what I am doing today. I think that is, giving the pace of change and how fast this type of technology is moving, and I listen to a podcast with a gentleman who wrote the book Human and Machine, recently and he spoke to the fact that AI is the fastest, the technology is moving its way into the workplace faster than any other technology has in the past. When you look at how quickly it is moving and how fast the pace of change is in organization already, if we are too fast to jump and say how can we use AI to support how we are delivering courses today. I think we have missed the target.

JD Dillon 

So I think it is important for us to pause, as learning and development professionals, to understand how this foundationally can impact the people we are supporting and the work that we do, but not pause so long as to let things pass us by. You know, our value comes into question because we are not up to speed with what we could be doing to support an organization.

Donald Taylor 

That is a really good point. I think this idea that, I have seen so often in the past that we have jumped on certain things and been very superficial about it and I think mobile learning is a good example. There was a lot of really bad mobile learning around 2006, 2007 and then it went away and then it came back and of course now it is pretty much business as usual. Not necessarily being used as well as it could be but I think it is something that we just except that we have to do it.

Donald Taylor 

But there are also other things where we have failed to engage with things until too late, and I think actually analytics and understanding how to interrupt data and put it to use is something that we are way behind on. Lets not get ahead of ourselves. I think you are right we have to get right at the right pace, JD.’

JD Dillon   

Yep and I think that the biggest thing that, I think we are getting right; at least in a lot of the conversations that I have with people who are in the mix, facing the challenge, starting to have these conversations internally with their peers and stake holders; is that they are starting with how is this technology influencing my business, right? How are things like automation and machine learning being applied within the conversation already. So again, is it not, I highly doubt that the learning and development team is going to be the one to introduce AI to your business. I highly suspect that something is already happening, something is changing with regards to how your organization is using data, how they are making decisions, how they are automating tasks, and all these various things that this type of technology can do. It is more about, and again in my opinion, learning how it is influencing the workplace and then figuring out how in that AI enabled workplace can we take advantage of similar types of tools and technology to help people do their jobs better.

JD Dillon 

I get encouraged when people come to me in a conversation talk about what is happening elsewhere in their business when it comes to machine learning and data, and then start to relate that to how they are going to apply it in their practices.

Donald Taylor 

I am seeing people [crosstalk 00:22:53]. Sorry, go ahead guys.

Liam 

No yeah I was going to say, Don do you have any sort of comments on that? [crosstalk 00:23:01]

Donald Taylor 

Yeah I was talking to someone last week working for a very large UK retailer and you know in retail you can absolutely expect that have got, I would not necessarily call it AI, that have got apparently intense sophisticated algorithms for managing things like supply chains and what have you. And of course they are looking at using that data effectively in learning and really stitching things together properly so that you can look at the impact of any, both interventions and support work done so that people are learning effectively and you can see the impact of their learning and you can see what, how the impact of the learning can be better or worse depending of certain variables that you change. Because they are able to find, combine, and use the data effectively we are not quite talking about I do not think AI there, we are talking more like probably some algorithms and some smart machine learning.

Donald Taylor   

They are talking about really getting data to work, but it is complicated in that something that we are going to come back to later on, in the difficulty is getting this data out of the silos that it exist in within the organization and making that, getting permission to do it and then effectively accessing the data turns out to be quite a lot of work, especially with a large organization employing several hundred thousand people.

Donald Taylor 

Sorry JD.

JD Dillon 

No I disagree but back to Liam.

Liam 

Well we can open that up for further conversation because as you said we have seen some interest in use cases across the business and JD you talked about how you are seeing great potential use outside of learning. So I guess really what we would like to see is [inaudible 00:24:56] what are some of those potential use cases for AI in workplace learning? So perhaps if you audience, if you are seeing some applications or perhaps some challenges that you felt that AI might be able to support you, we would love to hear that and sort of debate where that might go as well.

JD Dillon 

Whenever I ask this to folks the first word out of peoples mouths is chat bot. So I am especially interested to see if folks are exploring talking about or seeing application that have nothing to do with a chat bot. Not saying that chatbots are not a potentially a great idea, but I think that the reason to talk about the potential use cases is to blow open the doors of all the ways that are already in play. So I think that our conversation today is not about theoretical world of what may happen; it is what people are already doing with this type of technology to hopefully expedite the conversation within an organization because it is really right now it is not a maybe in the future type deal.

Donald Taylor 

I think fair enough. Well why do you think we are doing this Webinar rather than doing a face to face, just to avoid your travel [inaudible 00:27:03].

JD Dillon 

I get on a plane on Friday.

Donald Taylor 

I get on a plane tomorrow. I think JD has a fair point there, that we could talk about these very grand cases, I mean chatbots are a valid case of using AI; I think there are other ones as well. The automation of daily tasks is an important one. I think actually, JD, I would say some people will resist it. They will resist it just because it is a challenge because then what else do you do with your time. I think all change will have some resistance to it.

Donald Taylor 

Pascal is saying from my point of view it depends on the people you like to support. I have not seen that many use cases but would it be possible to tackle with some other tools.

Donald Taylor 

Maybe he is talking about performance support might have suitable use cases around it and I think there is lots of cases where AI could be used with performance support.

JD Dillon

Yeah I agree and for me I agree with what you said with regards to change. For me it comes back to where people place there value, right? For some reason my professional value is wrapped around, like if I literally negotiate travel logistics for a living I am going to be concerned. So I think that comes back to how is this influencing the work conversation and where people can be of greatest value at work and then where the technology can support us and then how it relates to learning.

JD Dillon 

So I think if we jump up, I think the big conversation for me is a consideration; looking about the current use cases that are out there and making sure that we keep in mind, where are they today what potential value in their current iteration could they potential delivery at grand scale but specifically for your organization and for then how wide scale are they in terms of adoptions or how proven are they. I think when you start to, this is not meant to say this is every application of AI in workplace learning that is alive today, but when you start to parse things down based on value, adoption, and then who is the main audience.

JD Dillon 

So first looking at the types of application that specifically focus on the employee experience so that is where you start to see chat bot as a performance support mechanisms start to introduce, the content recommendation engine where a lot of them are not AI machine learning driven, but the idea of connecting people to content, to use natural language processing in order to better inform the search experience for people trying to find answers.

JD Dillon 

So I think, there is again a wide range of user facing or employee facing application already in play. Personalization obviously being near and dear to my heart as something that we do at Axonify which is figuring out how do we provide the right support at the right time across the board of whether it is training delivery, performance support these types of ideas. But for me that is our huge value proposition already in play in terms of leveraging technology and data in order to be where people need us to be given how complicated scales can be for an organization and the fact that I do not think anyone here has unlimited and budget and resources.

JD Dillon 

So how do we feel those gaps that typically exist in our world, given that we support a very complex audience and very large numbers across the world but with a finite amount of resources.

Donald Taylor 

JD sorry, you were going kind of fast there for just a second, where did you say [inaudible 00:30:13]-[inaudible 00:30:14]play on this grid.

JD Dillon 

So I mean there is multiple places where Axonify specifically applies this technology already on screen, but I mean personalization for me is one of the biggest value proposition in the world of AI machine learning today and I think it connects back to what you said early around adaptive learning and personalized learning being set and enabled by this type of technology.

JD Dillon 

So it is less a conversation about AI, it is more a conversation about personal need of an employee and the fact that people change grow over time, needs changes as businesses evolve, and how can we leverage this type of technology to be where people need us to be to help them whatever they need help at and to give them time.

Donald Taylor 

I would add to that I am seeing a lot of activity in the top left quadrant here, the smart coaching and smart assistants piece. Well i totally see that as being, well I do not know what your road map is, but I do not think it is beyond imagination to say look add in personalization and makes it look like an offer for the moment. With enough AI during this smart coaching area already effectively.

JD Dillon 

Agreed. Along those lines I think it is important to not look at AI and one lane, right? That is where I kind of come back to say AI does not equal chatbot, that is the way that we do this in our organization and I think it is important to explore and consider all the potential range of application again already alive in the market place to figure out how do the pieces fit together for you not which one do you chose.

JD Dillon 

I think if you jump ahead there is the administrative side of this conversation as well. I mean just to highlight content authoring as an example where today machines can write better multiple choice questions than we ever could. There is a variety of articles you are already reading on the internet that are written by machines that you do not necessarily know are written by machines. But how does this technology start to change our relationship to content authoring? Where people today their entire job may be writing questions or writing articles, what could we be doing with our time and our skill and our inherit human capabilities where a machine comes up behind us and picks up some of those pieces that moves us faster and makes us more agile in terms of content delivery.

JD Dillon   

I think the other one I would hammer on is impact analysis. Which you mentioned early, in terms of that ability to actually understand how the work that we do directly influences results for an organization for business and then again that is something we have been doing with my teams for awhile now. It is interesting to see the level of, lets just as, out right confusion when I say to someone no I can actually tell you directly how much training is influencing the results of your business because that is not a conversation they have ever been having before, are familiar with. Meanwhile if you said I can tell exactly how much in terms of sales this particular advertisement influenced, people okay I understand that. But it is the same thing to a certain degree using similar technology we just have not opened the door to that conversation yet.

JD Dillon 

So, I think that is another one that is very interesting to think about. Using data to piece together how we are actually influencing change in our organizations.

Donald Taylor

JD, you know, you are dropping that in here but that is a whole Webinar itself and totally you have to do that in the future because it is so hot.

JD Dillon 

Agreed and I think there is one more that you have to look at. There is a kind of splint between the way these tools can be used to support most on kind of the back end that we do and the front end facing an employee. Translation is a great one to highlight just because how long does it take, how challenging is it, how expensive is it often to get material translated into a variety of languages that you support and even if you maybe just support four languages in your business formally, but how many languages are out there in terms of your employees preference with regards to how they would actually benefit from a greater representation of content that was more relatable to them. There is a lot of conversation around AI technology especially coming out of Google with regards to their ability to rapidly and accurately translate spoken word yet alone text content and these types of things.

JD Dillon 

So I think there is a wide range of capabilities out there that can impact organizations in different ways and I think it important for anyone who is interested in this topic to start digging into what is already happening and the value that is being recognized in organizations today to start to realize how can you start to light up these capabilities and not stay in one particular lane because that is the one you heard about first.

JD Dillon 

Which is where I tend to dig on chat bot because it tends to be the one that have heard about already because it has been a little bit faster to market than some of these other ideas.

JD Dillon

For me to just kind of wrap up the idea for use case. It comes back to, for me, not picking one use case but looking at how does this technology become a foundational component of how we do what we do. Again to liken it to mobile learning, I do not look at mobile as you make a decision as a L and D person to say are we doing mobile learning this time or not. People use these devices everyday as their work so they become a natural constant consideration in how we do what we do.

JD Dillon 

So for me and the work that we do at Axonify everything is underpinned by a couple key considerations. One, what is the priority of the business, what problems are we trying to solve, how are leveraging data to validate those challenges, identify where the gaps are, leveraging AI machine learning to make better use of that data so that we can become more targeted in the ways that we support people and really hit that and match up that employee need with that business priority.

JD Dillon 

So I see AI as less a specific application or use case, more as a fundamental component of how we do what we do as we evolve our practices in L and D.

Donald Taylor 

Well that is really interesting, it is not an add on it is fundamental.

JD Dillon 

Right.

JD Dillon 

I think that is where we miss with a lot of the trendy things. We look at them as if this or that. Recently I did a presentation where someone asked what is the difference in adaptive learning and self directed learning. I really wanted to say, “and part of the challenge is that question” because when we put things in boxes and then determine when are we going to pull out each trick I think we do a disservice to what we could be doing with these ideas if we distill them down to their fundamentals and really plug them into how we do what we do everyday rather than put them in a box and pull them out when necessary.

Donald Taylor 

Yeah.

Liam

Okay, so there is a danger that I am going to do that with this very question about simplifying this. Because one of the things that I certainly get asked about well this data question was fantastic but what is that data that we really need for L and D need to apply AI?

Liam

So I recently read something about the Tesla car and how many data inputs it has to have in order to able to drive a car across the street without the driver, you know everything from the distance to the weather to how fast the car is going there is so many data inputs needed just to make that car drive alone on the street. Now obviously L and D not quite as complex as driving a car down the street, you may argue differently, but I am really interested to understand what people think of the data sources we need. Then get your thoughts Don and JD in terms of what you are seeing that you would suggest that people do need those sort of types of data really.

JD Dillon 

For me the biggest thing is the realization that none of this conversation matters if we do not get better with data. Right, because a machine simply needs data in order to learn, in order to find patterns, in order to make recommendations and decisions and to take action. So if we still think of data the way that we certainly have in terms of training, none of this really matters. We have to get better at this first before we can step forward with some of the use cases we just talked about.

Donald Taylor 

JD, can you expand on that a little bit.

JD Dillon   

Sure so I mean, what can you do with course completion and test scores. It is really what it comes down to. So if we continue to stay in our lane, in our box with regards to the only data that exist is the data that we have historically generated course delivery the greatest AI in the world still cannot do anything with that. Obviously when I talk about things like translation there is still ability to leverage the technology in certain ways but if we really want to get to the use cases like personalization, understanding where there are gaps in a business, determining what the impact of what we do is on results we have to expand the term learning data to be much more of a grand conversation around the different types of data that influence people and what people generate and how work is done.

JD Dillon

Yeah I think the term learning data has often been separated from the term business data and in my view they are all part of the same grand puzzle. So we have got to get out of our data in regards to data to make this conversation about AI meaningful.

Donald Taylor 

You really put it very well and it comes down to what data have we got, to which meet our goals and talk about it that way. It is not just in the learning department of course the course completion they are useful for somethings tactically and [inaudible 00:39:37] within the department but they are not useful for the goals you want to meet.

Donald Taylor 

Pascal in the chat raises a point. Pascal says “In my mind you need to first know what you are trying to achieve and gather the data fitting to it. If you gather data then try to hope for the best and hope it will figures things out for you then you are going to go wrong.” I think that is a fair point.

Donald Taylor   

So I guess that the implicit in this question is what are we trying to do. Well what we are trying to do is influence the business if you are trying to influence the business positively you need business data. I think that is the result there.

Donald Taylor

That means a couple of things, that means we have to have access to the data, we have to know where it is, we have to have access technically and permission to have the access. We have to have the skill, the skill to be able to analyze it as well and do we have that. Remember earlier we talked about the LPI, the Learning and Performance Institute Capability map, and the 25 skills on that. In the first 6 months, you can see there at the bottom, 1145 people have been through the self assessment process and perhaps unsurprisingly the top, the skills in which people rated themselves most highly are the traditional areas that we have come from; face to face training, designing solutions, and creating content. Well you know when I used to be a face to face trainer back in the 1990s that was what you did all day long so it is unsurprising that we have a lot of core people with that skill. But where are we weak?

Donald Taylor 

We are weak in of the 25 skills the 23 one is marketing and communications, which absolutely actually is a skill which everyone in a department I think these days at least needs to be aware of and at least have a good strength across your department area. Not so warned about 24 because that is a specialist skill and you could have one person in the department or indeed you could have someone outside the department that you work very closely with. What is number 25? What is the skill that we across the population, this is worldwide sample of people, rate ourselves the lowest at? It is data analytics, which is actually understanding what we have been talking about for the past 45 minutes. I do not find that depressing, I just thinking that is it an interesting view of reality. My view of this is that you do not have to go and become a data analyst, you do not have to go and become a data scientist, you do not have to become an AI expert, but you do have to get those skills into your department one way or another. Almost certainly from borrowing them from elsewhere.

Donald Taylor 

JD you have a great phrase about this don’t you. I cannot remember what it is exactly but you will be coming to that later on, about where you can find the data analytic skills. JD what is your view on all of this.

JD Dillon 

I think it like a lot of think, where the first step is recognizing that you have a problem. [inaudible 00:42:30] It is like workplace data AA to a certain degree. So I think we have been locked into a variety of models. On screen you can see the Kirkpatrick Model as the example of, I generally think there are two things that I think if I go up to anyone in this profession worldwide, there is two things they can tell me. What ADDY means and what the Kirkpatrick Model is. I think the more that we have become accustomed to this being the definition of measurement and data in our profession, whether it is this model or a variety of others, we have limited ourselves to what we have been capable of in the past. What technology could do, what data looked like? Now we are struggling to still get passed level 2, right? Very few organization admit that they can get to level 4 and as long as we are having a level conversation I do not think we are doing what we need to do.

JD Dillon 

So when it comes to, i completely agree, with expanding the definition of learning data to be more about business data and the data that influences performance and the results that we see from human performance. I think that, again taking a lesson from marketing, and realizing that we need to expand our definition of data in a variety of ways. The Vs of Big Data is a marketing concept but recognizing we need to get faster at how we access, generate, and analyze data. We need a greater variety of data, in terms of the type of information that can point us towards gaps in performance and how it relates to what people are doing on the job. We have to make sure the data is trustworthy because if you bring in bad data into the smartest system in the world you are still going to get junk out the other side. It takes a lot greater volume of data to make this conversation work. It comes back to what I think folks said in the chat earlier around the value of data. It has to be about what the business is trying to achieve, not data for the sake of data in order to get better and say we have got more. That is not sufficient.

JD Dillon 

I would love to be able to say these are the seven data points you need to collect in your organization to really drive forward with AI and machine learning and personalize learning and things like that. For me it was a wide range of potential answers to that question. Here are some things to consider. I think today we are great at demographic data, knowing who they are what department a person works in, where they have come from, how long they have been with the company, who their manager is, and these types of things. But really exploding out the definition for a data profile for an employee to include everything from what is influencing this persons performance, you know if you work in a retailer there is a lot of seasonality that takes place in that type of environment or product launches will directly influence performance and results and what not. So taking everything like that into account. What people know not just what they are consuming but how their knowledge is changing. How their behavior in real world is changing and how that relates to their change in results.

JD Dillon 

Because a big part of this conversation on data is being able to leverage data to identify when we can support from a learning prospective and when learning actually is not the problem. When it is something that needs to be, we need to bring in addition types of support. When you speak to the idea of smart coaching earlier, right? We can identify if we have the right types of data, that a person knows how to perform they have got the knowledge, they have got the skill, but they are not seeing the results. How can we inform a conversation with their manager to make sure that it is not oh they need more training or maybe there is a motivational gap, maybe there is a resource gap, a process gap and all these other things. So I think there is a wide range of ways that we can use the right types of data to identify where the gaps are and close those gaps. I think it is something that you have to look hard at your organization to understand what data is out there and how can it be related to the work that we are trying to do.

JD Dillon 

Any thoughts from you Don?

Donald Taylor 

This is a great point the smartest data person does not work in L and D. That is not to say we are not smart in L and D it is just that the skills have been developed to go out and beg, borrow, or steal those skills if you possibly can.

Donald Taylor 

Serene who works here in the UK, if I am correct. Serene correct me if I am wrong. “Hang on guys, data analytics is greats but we have to be mindful of what the business wants, wants us to measure, and analyze. We often get feedback from stakeholders are the numbers just to show me how business the L and D team are?”

Donald Taylor 

I think, Serene, just to reiterate the point here the data we are talking about is not the data from the learning management system. Correct me if I am wrong JD, it is not the data that shows how busy we are. We are talking about getting data in from various different parts of the business and then the analyze which allows the correlates made with us to say, these activities in L and D are correlated with this success and also perhaps with these failures. So we are using this to improve our processes and deliver better performance to the organization. That is the purpose of the data analytics and if we are just showing that we are busy we are doing nobody any favors.

Donald Taylor 

JD fair enough?

JD Dillon   

Exactly, for me I think a lot of this conversation about expanding our use of data is not something is not that is necessarily involve the stakeholders or the people that we are working with. I think transparency is a critical part of this conversation, anyone should have or be able to have an understanding about how data is being used to support and enable their performance. It should not be a mystery as to why, like today it is a mystery why Netflix keeps recommending Boss Baby to me because I am not going to watch it, but it keeps happening.

JD Dillon 

An employee should not have that same why is this happening to me that we have when we browse the internet and keep getting ads for certain things. So we need to handle data in a different way with a level of transparency and build that trust, but ultimately I think it is about raising the right information, the right recommendation, the right metrics to the people that we work with, to stakeholders, and to executives to really understand how their business is changing and where the needs are for them to get the results of trying to achieve not just numbers that say “oh look I have got better numbers now” because most people frankly do not care.

Liam 

Perfect well that makes perfect sense JD. So I think it does pose this next question quite nicely in term of how can L and D professions start preparing for the introduction of AI in the workplace learning strategies? Now where we are sort of pressing up against the question time. We have allowed the back end [crosstalk 00:48:54] so JD so perhaps if I you have so points in what we can start doing around this area.

JD Dillon 

Sure because I think a lot of times the questions will come up and again I have had these experiences in my different L and D roles in term of, I did not necessarily have time, money, and resources to go out and play, to try new things, to go buy new technology all the time, but how can I start getting better and prepare myself for a world where my technology will start to leverage AI and I need to be better at data by the time that I get there.

JD Dillon

For me it really starts with drilling down the connection between what business is trying to achieve, the results we are looking for and the actions that we take. I think that if you are not able to take step forward today with the types of use cases that we have mentioned, really starting to understand how results are connected to expected behaviors, what people are expected to do, how that relates to what people are expected to know and what you are expected to kind of convey in terms of communication and then how that informs the decisions you make. Whether is is building a video, building a course, building a piece of performance support. Really having those connection built because then you are opening the door to better measurement strategy because as you start to expand the types of data you can collect be able to collect data about what people know, how they are behaving on the job, what types of business results we are seeing really inform this conversation.

JD Dillon 

So taking a results oriented approach is a great starting point. Then here is a wide range of different steps I think people can take to start their machine learning, their AI journey in their organizations . when you look at the list on screen I think exploring how your organization is already involved, what you are already doing with these types of technology. As Don kind of mentioned buy lunch for your data scientist, we have had this strange adversarial relationship with IT for most of our live and now I think the data people are becoming potentially as important if not maybe more important and we cannot show up at their door when we want something. We need to already have engaged them in a conversation in what we are trying to do.

JD Dillon 

Then experimentation asking around, if you are working with vendors, challenging them to say how are you starting to take advantage of this type of technology to really help me do what I am doing better and to better support our people, I think is a critical point. The last one for me is what is the point, what is the purpose, what are trying to do, how are your trying to help people, what are you trying to achieve as a business. Then where does that enable a purpose behind your use of this technology. I think that is what was missing in a lot in the previous biggest conversations ever in L and D where they did not really define why what is the point of going down this path and how it going to over achingly make us better at making people better at what they do, or helping people get better.

Donald Taylor 

Pascal says that is a great list, he also says earlier learn more about the actual needs of the business and the people, and I think that is a fair point in fact very much goes with your last point of the list there JD. I would add other thing on the list here, which kind of a few of them allude to, which is talk to people about how AI or indeed machine learning or indeed basic algorithms and data are being used else in the business for business ends. What are people using the data for? That may illustrate a couple of things. Firstly, data sources that you were not aware of and secondly, ways of thinking about the business that you have not contemplated. Having those conversations, what is already happening may spark some thoughts that may be able to go off and do some great AI of your own.

JD Dillon 

Yeah and a lot of instances a lot this is already there, right? So I always make [crosstalk 00:52:36] aware. If you work in a contact center you are working in a data mine. You do not need all of that data but potentially, but what data could be informing your practices. If you work in the logistics industry or like trucking as an example, an amazing amount of data is coming off a truck as people drive a truck and soon as trucks drive themselves. How could that data be used to inform what you are already doing? So I think it is a conversation that could get started right now even if you do not have the resources to make leap when it comes to technology.

Liam 

Brilliant, well thank you both for your inputs. Now we are going to open up to the audience for questions, so if you do have any questions please fell free to type into the chat and we will post those to Don and JD. We are also interested in your comments. Obviously we opened up about is AI is it real or is it hype. So is AI in the workplace learning real or hype? So if you have got a gut reaction to this if you could type in real or hype into the chat we would love to see what your thoughts are. And start seeing what the general sort of consensus is, is this something that is really happening or is it a bit of hype? So we got the first real coming in.

Donald Taylor 

We cannot lead the audience. We should let people vote and then we will reveal the results real or hype, let us find out?

Liam 

You think I am going to lead them down a path do you?

Donald Taylor 

I suspect we have skewed non representative sample here. Oh yes.

Liam 

As those are [crosstalk 00:54:10]

Donald Taylor 

Meanwhile as they come in, as they come in perhaps [crosstalk 00:54:11]

Liam 

We can answer some questions.

Donald Taylor 

Yeah look there is a great question here from Pascal saying “JD can you share how is Axonify using AI in products already”

JD Dillon 

Sure so to short hand because of limited time, Axonify is inherently a personalized adaptive learning platform, so every time a user access Axonify, which tends to be for just a couple of minutes a day kind of along the realm of micro learning principles, we take a look at their individual data profile and start to change the experience in any given set Axonify session based on where their identified area of greatest need is. So we could be pulling in data based on proven knowledge, maybe data based on behavior observation, maybe business data based on sales results, or efficiency scores or really whatever data is our there about how an individual is performing on the job.

JD Dillon 

We then leverage that data to one personalize every interaction with Axonify so we are focused on again where are you experiencing gaps based on the priorities the business has land down and two we are using that data on the back end to both inform coaching conversations with managers. So managers can look at an individuals profile and can understand where their strengths are, where their potential areas of opportunity are, and then on the back end as an administer you can take a look at where you are focusing your training efforts at for Axonify and how those topics and subjects are actually impacting the results for the business. So, being able to break down business result to say this is the degree to which training has influenced a result and these are the other factors maybe influencing the outcomes for your organization.

JD Dillon 

For me, Axonify is kind of a wide range of application use case when it comes to AI and machine learning.

Liam 

So JD what about if the business is not asking that question? What if they are just asking tell me how many people have completed this learning? They are still in that mindset of putting people into classroom is the right thing to be doing. So how would you help people through that?

JD Dillon 

Sure as a person who has experienced some type of layoff in literally every company I have worked in, my recommendation is the question is coming. Even if it is not being today, it is something that you should be working towards because someone is going to raise the question what value are you delivering to our organization when it is time to make a decision. So I do not think we necessarily, even if people are still asking how many people went to the class, how is our compliance looking. I think it is a yes and story, still giving people what they are asking for in terms of where your value lies in their eyes, but being able to start growing that conversation to something that we know is more meaningful because it will eventually get there whether we are ready for it or not.

Liam 

Great, great point. We are coming to the end of the hour, so please do feel free to keep firing through questions. We will also share JD and Don’s contact detail so you can also contact them afterwards.

Liam 

So, JD just sort of final thoughts to wrap of the session?

JD Dillon 

I think folks in the chat really nailed in term of the connection point between what we are trying to do and how this technology can help us drive results for our business. So when we, I fundamental believe, and I think we are on the same page here Don, that this technology and what it really means to us, has the capability of fundamentally shift what it means to do work and then how we enable that work to create a successful business. But we really have to identify what that purpose is and what we are trying to achieve before we start leaping down the path of data and technology.

Donald Taylor 

Yeah totally. The guys in the room really kept us honest on this one, Pascal, Barteks, Serene this are the names that come to my mind. But yeah identify a goal but I think it is something that is going to fundamentally change some things. It may be revolutionary, it may be evolutionary but it is going to fundamentally change things. We can get ahead of it. There is tremendous opportunity here for us, there is also risk if we ignore it. I think the people in this school, the smart ones, are going to get ahead of it and make it successful and make themselves successful as a result.

Liam 

Brilliant well thank you for that, that was great session. So we have got Don and JD’s contact details here as well. So feel free to carry on the conversation post the Webinar. As I said we will share the slides and the recording with you as well. I am sure Don and JD will be more than happy to sort of carry on the conversation offline.

JD Dillon   

Of course.

Donald Taylor   

Of course always happy to.

Liam   

And if you want to carry on the conversation with Axonify we also have a blog that you can sign up for. So axonify.com/take5 where you will a weekly digest of information and there is also a link to L and D Global Sentiment Survey that Don had mentioned earlier, and also JD has an area where he curates everything. We talked curation earlier but everything for around AI and learning content as well, so you can find that at learngeek.co/ai.

Liam 

So I would just really like to thank yourself JD for your inputs today and also Don for a really good session. And the audience for some fantastic questions and responses throughout. So that brings us to a close today, so thank you all and I wish you a good rest of the day.

JD Dillon   

Thanks everybody.

Donald Taylor 

Thanks guys.

Liam 

Thank you.