How to Use AI to Deliver Personalized Learning at Scale
In this episode of the HR Leaders Podcast, we speak with Ryan Laverty, CEO of Arist, about how AI is completely transforming workplace learning.
Ryan shares how his team delivers 95%+ engagement by pushing personalized, AI-powered learning through tools like SMS and Microsoft Teams. The conversation explores how AI is making learning faster, more scalable, and more accessible for every employee, not just the top performers.
🎓 In this episode, Ryan discusses:
How to train the "non-5%" who usually miss out
Why agility is the most important L&D capability today
What AI-powered workplace learning actually looks like
How AI agents can build, assess, and deliver content end-to-end
Why SMS and Teams are outperforming traditional learning channels
Greenhouse is the only hiring software you’ll ever need.
From outreach to offer, Greenhouse helps companies get measurably better at hiring with smarter, more efficient solutions – powered by built-in AI.
With Greenhouse AI, you can generate stronger candidate pools faster and source high quality talent with more precision, streamline the interview process with automation tools, and make faster, more confident hiring decisions with AI-powered reporting.
Greenhouse has helped over 7,500 customers across diverse industry verticals, from early-stage to enterprise, become great at hiring, including companies like HubSpot, Lyft, Trivago, Crowdstreet and Gymshark.
Ryan Laverty 0:00
For HR leaders, the core challenge is, and has always been, and in the world of AI will still be a significant percentage of your workforce probably needs to learn something, and either doesn't want to or doesn't know that they need to learn that thing. And that's where push is still the foundation, pushing learning, pushing content, whether it's AI based or not, to people, is so important is because most people don't know what they need to know, and they are not in the top 5% of your most engaged learners who are just trying to practice all day. It's those folks where you're gonna see the biggest impact from from actually scaling learning and training.
Chris Rainey 0:42
You Ryan,
Chris Rainey 0:49
welcome to the show, my friend. How are you? I'm doing great. Chris, thanks so much for having me. It's been a while. I mean, when was we in New York? How ago was that? I don't even
Ryan Laverty 0:58
know, a few, a few months ago at least. Yeah, yeah. It
Chris Rainey 1:01
was super fun. What a view we had as well. What you've been up to, mostly
Ryan Laverty 1:07
traveling. Chris and I were just talking about how much time we spend in Vegas, which I think one one day in Vegas every year, is still a little
Chris Rainey 1:13
too much. We have to give context. You can't just throw that out there with no, we have no context. Hey, yeah, Chris and I are just talking about how much time we're
Ryan Laverty 1:21
in Vegas. Where does in Vegas all the time? Just, you know, for no reason. Why?
Chris Rainey 1:27
Yeah, why does everyone host HR conferences in Vegas? You live in the US. I mean, why? I don't understand what
Ryan Laverty 1:35
I spoke to someone who runs events. They said you can get a 30% attendee bump if you host something in Vegas versus anywhere else in the US, which doesn't make sense, shouldn't make sense, but I guess it does. It's
Chris Rainey 1:47
interesting, because most of the citros I spoke to from the from the last event, that's not I get the opposite kind of they're like, I think once you go, once you like, okay, yeah, yeah, it's cool. It's cool to see the sphere. It's nice to see the sphere and some of the attractions and stuff. But it's like, apart from that, I'm like, I want to get out of here as fast as I can as well. How are things on your end? Because when we last spoke, you had some super interesting things in the pipeline, coming up, some new product launches that I was like, asking you loads of questions about give us, give everyone an update. Yeah, things
Ryan Laverty 2:25
have been, have been going great. I mean, I guess just first quick context, the 10 second, for folks who don't know what AI essentially does is we use AI and SMS and Microsoft teams to build learning rapidly and then deliver it to where people spend all their time. Average person checks a dedicated learning system every three to six weeks, checks their texts every three to six minutes. And so when we combine ai plus these tools people use every day, you get customized, personalized content in a place you're already addicted to with no new tools. And the most exciting thing for us has been that these models are just becoming far, far, far more accurate, you know? So I'll give you an example, Chris, like we, we released a bunch of features around image generation, file embedding, and so you can, you can literally build a full end to end. Course, with AI. A lot of people say AI creating. What they mean is AI is going to help you with a question, or it's going to rewrite something, and it's going to kind of sound like chat GPT wrote it, and it doesn't feel real, right? The Cloud models that we've the sonnet models we've trained now, we can give it 1000 page document from like a Pfizer or Merc on a complex vaccine and studies, and say we want it at this reading level for this audience, with this outcome, it will use the pedagogical thinking, generate all the images, build all the structure, and then can deploy it that day, and it will build everything to spec that you don't have to touch a single thing before you actually push it live, even if you translated it to 20 languages, that's how accurate and how good the models are getting.
Chris Rainey 3:53
Yeah. What about speed now, in terms of how long it takes? Because in the past, you know, I sold various different load times, how long it takes to generate is that improved?
Ryan Laverty 4:02
Yeah. So generation time full, end to end, using images, highly complex content is now about three and a half minutes. Wow. So before it was nine minutes with no complex content, no images, now with everything illustrate, we're talking about infographics. Infographics illustrations is about three minutes. And again, what I always remind folks is people focus on speed and learning and spend a lot of time and attention on the speed to create. It's the speed to deploy that's the real killer. You can build a, let's, let's say, in a world, you could build a perfect workshop or PowerPoint deck in a minute and a half. It's then going to take you, on average, two to six weeks for a fortune 500 company to even get 20 to 30% of their employees or Salesforce to even go look at the thing or read the email. 50% of people will probably never see it. Once you've got over 20 to 30,000 employees, and so it's the speed to get something in front of someone, have them pay attention, have them respond, regurgitate. That's really where you want to focus your attention,
Chris Rainey 4:56
and that's where it's so important, like we're also focusing on that of our. HR agent to be able to distribute and deploy that straight into teams and text and slack, etc, so immediately it shows up in the tools. What's what? What's the typical engagement rate when you're the
Ryan Laverty 5:14
engagement? Engagement rates with text and teams are over 95% and doesn't even sound real. People think I'm buying. A discussion with our sales team, and we were like, Should we just start saying that it's like 76%
Chris Rainey 5:27
to make it more believable, right? Because 95 but I think the people
Ryan Laverty 5:31
tell me that that doesn't sound real because they're used to benchmarks for learning tools. And then I say, what percent of your Slack routines, messages do you think you answer? And they think about it, and they're like, Well, I read every single message and I respond every single message, and it's like, well, technically, we're not building a new behavior, which is where engagement rates have to be built. We are borrowing an existing, very addictive behavior, which is where so we don't actually build an engagement rate. We borrow one, right, which is a very different type of experience and gets a very different type of rate.
Chris Rainey 6:02
How do you manage what you can put and the limitations of text as a service delivery versus, like, a teams or a slack?
Ryan Laverty 6:10
Yeah. So it's actually, it's, it's almost identical in terms of the learner, in terms of the the admin experience. And so, like, if I'm building a course, the AI will set certain character limits, certain character limits, certain number of questions like it'll build it in a certain way. Right now, we actually keep the same exact structure for teams and text. You can go you can go slightly longer on teams, but it's not like like. And think about a single message viewed on a desktop, but because a lot of people view teams messages on mobile, it ends up looking from a stylistic perspective, very similar to text, and so we actually just make product decisions around assume it's mobile, keep it a bit more short form, keep it confined to that, because we can't control if someone's going to view that on like teams desktop versus teams mobile. Yeah,
Chris Rainey 6:53
is. I'm like, may sound like a silly question, but with most with most companies still having their employees have the teams app and then having the text within there. Why still have traditional text message?
Ryan Laverty 7:07
Yeah. So do you mean like, why do we still use text message? Why not just do all Microsoft Teams?
Chris Rainey 7:12
Yes, and slack and stuff like that. Why is text still relevant? Text is
Ryan Laverty 7:16
relevant because for a lot of employees, they don't have access to a slacker teams. And so I'll give you an example. Some of our large customers are folks like an Exxon Mobile. And if you think about the and we're seeing a lot of movement, actually, with we signed on recently on healthcare companies, you know, something like, I'm training nurses in large hospitals for a nurse in a large hospital. This one study we did, they work an average of 18 hour shifts, and for 17 and a half of those hours, they are not even on the floor that they have a nursing station with a laptop. They're not even on the same floor as that nursing station. A lot of times, the nurse is not even in the same building or in the same wing, and so you're putting all of their training, Let's even say it's teams right on a laptop, and then they have phones on them, but they are going throughout the hospital for an entire 1718, hour shift, and so it just becomes completely, you know, unrealistic for them. Some of our customers are large manufacturing plants, basically data centers, yeah, frontline and field workers as well. Yeah. Now that
Chris Rainey 8:18
makes sense. I mean, I was just thinking that they would have teams app on their phone, but then they still have to have the barrier of going to actually check that. So why not just get a text? And it
Ryan Laverty 8:28
really depends on the organization. Our goal is to meet your learners where they're already spending all of their time. And again, for some organizations, that's teams. For some that's text, et cetera. I love that.
Chris Rainey 8:39
One of the cool things that you mentioned to me when we last spoke was the idea of having an agent that can gather feedback from employees, managers the business, bring that back to the L and D team to then generate learning pathways. Have you made any headway on that? We're very, very close to releasing that. So that's in beta right now. So was the name, again, remind because it's on the website. I just forgot the name. Yeah. So teammate
Ryan Laverty 9:02
is the name of our of our product. And so there's a few stages to teammate. So the first you can think of teammate as essentially a end to end orchestration of several different AI agents. And so the first agent, which we already have released is our Creator AI agent that's by far the most used thing we've ever released. Essentially, that's give a few basic prompts, give it lots of stuff, build me full courses in seconds. We've had entire teams, 20x their output with Creator. The next segment of that agent that we've built out that's new since we last chatted is around analytics. And so I can respond to something, an agent can go look at my response, can grade that response, can route me to a specific scenario. So as a learner, I can get an open ended question assessed and then responded to. As an admin, I can go in and look at, okay, 30,000 sales reps told me how they described this product. Do I think I'm gonna have some type of a challenge with how people are articulating this? Is it? Do you think it's impacting reps? Revenue, right? And then the third component of that on the front end is the needs analysis agent. And so what the needs analysis agent is, is essentially exactly what you said, Go to senior stakeholders in the business. Let's say I'm building a course on company culture. Go to HR leaders, go to individual employees, go to the CEO and ask her about it. Go ask 20 or 30 different people. Come back and essentially synthesize for me all those interviews and then tell me where the gaps are and what I need to build. And so we're about, I think, a month and a half from the public release of that. And so that is when we combine it. The magic is when you combine all these things, when you combine AI needs analysis, plus it then takes the data it learned there with AI course creation, plus the delivery mechanism of AI risk, plus now AI, on top of these hundreds of data points we're collecting on every person, what you get end to end. You know, the ultimate goal of teammate is it essentially looks like a chat GPT style interface where I can just say, Hey, Chris, go interview these 50 people. Now go build a course on this. Now go deliver it to these 20 teams in these 30 languages, right? And that's really the magic behind it, is that all of this really complex orchestration can be very simple for the learner, very simple for the admin, but the outcome is really powerful. What's
Chris Rainey 11:10
I mean, I'm so excited. I think that that's this, that's a game changer. I thought, for me, that's probably one of the most exciting things when you first mentioned that, that that's all something that was done manually, that took months to achieve, which sometimes took millions of consulting companies coming in, charging a lot a lot of money to do that. And now the agent can go out there vitals in the flow of work, ask those questions, gather information, come back, present that to the team, and then generate the learning pathways. That's like, almost seems like, it almost seems unreal. What's the what's the biggest challenge of bringing it to life? Because obviously, you've been working on that for a while. What's the biggest
Ryan Laverty 11:51
challenge? Yeah, absolutely the biggest. The biggest challenge, really. So a few things. One, when a lot of people build AI tools, what they think of is, most companies, assumption of AI is, I'm just going to sit AI in an internal document repository, and it's just going to learn everything, and then I'm going to ask it stuff about my company, or ask it things, and it'll just know what to reference. The reality is that most company information is not just unstructured. It's very sporadic. A lot of it's wrong. There's no hierarchy to it. And so a very simple example is, let's say I fed an AI your calendar and your emails, and I said, prep me for an upcoming call. We are not a knowledge pulling AI system, but we tested a bunch of them. We went and did one of that with my calendar and my email. And what happened was, it started preparing me for a meeting that I had four and a half years ago. And so the context, the hierarchy, the way it can understand information. Let's say you told me something that was really important three months ago, but then told me something now that was less important, right? Doesn't have the context. Is it prioritizing the oldest info, the new right? Now that context for us, what's actually the reason we are slow rolling these elements, rather than just, you know, tomorrow saying, Look, we're gonna let all of teammate is actually company readiness. And so the Creator AI for like so for example, we do a lot of work in the pharmaceutical industry. Our Creator AI is fully source referenced, fully closed loop. It'll go reference medical, medical grade sources that these companies would want to see. And then it has, you know, human review. In the process, we found that, you know, that that is kind of the extreme example, because the pharmaceutical industry is so regulated, we've found that even outside of the pharmaceutical industry, 99% of companies still want close to that same level of rigor for even non sensitive, non proprietary, non anything, content. And so again, I'm seeing a lot of these startups that are coming to us and saying, we built this really shiny AI thing. Why is no company using it? And it's like because they're they're not there yet. A big, a big point of our product design is we'll go partner with a lot of these companies and do co design work. And so we we believe in slow rolling things out, and we don't want to roll something out until we believe that at least 15 to 20% of the market is ready to adopt it. We don't want to keep unrolling the shiny things and forget our current customers, just to have 1% of the market be ready for it.
Chris Rainey 14:10
Is there like a sweet spot you found of companies that you would work with to co create, because I'm sure otherwise, too many, too little could be too little and too What
Ryan Laverty 14:20
have you found? It really comes down to just having I think the right design partners are more than the having too many. I think five is probably a really healthy number. And when I say five, I mean, these are fortune 100 companies paying stakeholders, right? But really, what you need is a few good internal people who are innovative, but also have their eye on the ball of okay? They're not just like an innovation leader who's kind of separate from the business challenges, right? So for us, it's more important to have folks from different departments and functions and levels of kind of risk aversion within each of those design partners, and then have a few good ones.
Chris Rainey 14:53
What advice would you give to the HR leaders listening that want to bring a solution like this to the business? What's the best way? Or they should present it in order to get max, maximize the buy in
Ryan Laverty 15:04
absolutely present it, not in terms of workflows or even just direct outcomes. Paint a paint a vision of how the business will be different if this works right. And so I'll give you a really specific example. Ai says a technology that functionally by product. When you combine these things, we build stuff faster. We send it out to everyone. You can use that for reinforcing a workshop or telling everyone about benefits. You can also use that for cutting the time it takes to roll out a global vaccine training by 95% right? And so the difference there is, let's say I'm an HR leader, and I go to my leaders and I say, Hey, we're going to bring on this tool, and we've got this big stack, and it's going to go notify people of these things. That's great. You've just described a workflow, or maybe a learning outcome. Completion rates are going to go up, engagements going to go up. We call it the CFO sniff test internally. You know, is it going to pass the CFO sniff test versus if I'm an HR leader, and I go to and I go to the company and I say, right now, it takes us 10 weeks to get the most critical information to people. What's the result of that? Our competitor launches a new AI product. It takes us 10 weeks to respond. A new regulation comes out in farm. It takes us 10 weeks this, you know, we have a talent shortage. Takes us 10 weeks to get all of our recruiters ready, right? It takes us 10 weeks to do anything. Because when you have 50,000 people, that's how long it takes to get things to people. And so for us, that's why so much of our learning philosophy is around delivery and engagement, rather than just content, because companies have more content than the human race could consume, but what they're missing is is 10 weeks of delivery. And so for a chro, my recommendation is go describe AI in terms of organizational agility to your leaders, rather than in terms of capabilities. I love
Chris Rainey 16:43
that loved, and that's obviously what's going to resonate with those leaders the most. Right? Absolutely, you won't have to explain the value after that, because you're already talking the language when you think about the AI, I think you called it AI needs analysis agent. Do you give in the future when you release this? Are you giving the team the ability to choose what questions it goes out and asks? Like, how does that work? Because obviously they're going to want control over what's gone out. So what does that process look like? Yeah,
Ryan Laverty 17:13
it's similar. So right now, the way that even our course creator AI works is it gives recommendations, but lets you select your own that's a big part of our design process. That's the same thing with needs analysis. And so it will say, here's the recommended questions. Here's what I want to ask someone. But you can obviously go in and, you know, subtly in a text box, you would type like, Okay, make sure you go ahead on these five or six things. This is also, again, part of our product design philosophy. If I ask you, Hey, Chris, put in 10 questions. That's so much work, and you're gonna pause if I say, Chris, here's 10 questions. I'd recommend go change this before we hit
Chris Rainey 17:45
spot. Yeah. And I'm assuming you can have, like, some predefined ones depending on use cases that you already have, like, layered up, like, this is for managers. This is for, you know, exactly
Ryan Laverty 17:55
different. Yeah. AI is really good at essentially intuiting. What do you want to do? Like, with the with the Creator tool. If I feed it a bunch of pharmaceutical documentation and say, I want it for these reps, the learning outcomes it suggests are ones that actually make a lot of sense. They're about mechanism of action, which is how a drug works in a body. And people are like, Oh, I didn't, I didn't tell it that. It's like, well, it has the entire corpus of the internet on it. It knows that's a good question to ask or learning outcome to go for.
Chris Rainey 18:19
I mean, one of the things that stood out to me at unleash, I've never seen so many AI learning startups in my life. Every other Booth was an AI was an AI, you know, learning agent, you name it. There's this huge debate around what aI mean for the learning stack of the future, you know, is this the end of LMS is we've got LMS is lxps. You know, you're so, you're so, you know, deep in this space, I love to get your views on where you think where we are and where we're going.
Ryan Laverty 18:55
Yeah, absolutely. First comment there is, think back to what I said about delivery versus content. And so where I think a lot of companies will fail with AI is they will use AI to build a lot of the same stuff no one's engaging with. Let's build more animations, more PowerPoints, more content. There's this big focus right now on AI built learning content. If you go talk to learning leaders, the challenge isn't we don't have enough content. The challenge is we can't get people to actually pay attention to or read the over 40% of our time is spent on marketing, right? More, more broad level, if
Chris Rainey 19:27
we really 40% of that is on actually marketing, as opposed to actual to building interesting. Yep.
Ryan Laverty 19:33
So if you look at an L and D and led, about 80% is on two things, building content and convincing people to go take that content. A very small percentage gets to be spent on strategy, in depth needs analysis, right? And I think that's what we've we've 20x 50x the capabilities of some of these organizations. We haven't seen them say, hey, we need fewer learning leaders. We've seen them say hey, the nine month backlog now became the two. Week backlog, right? And that's really the output of that in terms of the learning stack of the future. Though, to your original question, there's three layers to it. There's push, pull and practice. And so the first layer of AI is going to be push based tools and learning. And so essentially, what that's going to look like are systems like ours that are going to be entirely focused on delivery. How do I get based on all these things that I know? How do I get the right content to Chris at the right time, in the right place, right? How do I say this is what you need. This is where you find it. The second layer is pull. That's where most learning today sits. But that's why the learning is not engaged with. It requires me to know what I need, to learn, to go find it, to go navigate, to pull it. And so when there's this whole debate around like LMS and Alex PS and the content and chat GPT, that's all just pull layer right, which is essentially, I have a question, I go find it. And so I think that the way that this space goes is either LMSs and lxps Integrate with something like a chatgpt or cloud intake their entire corpus, and then when you have a question, they can go find it. What's not going to happen is we have an internal company AI that goes and that learns everything when you have a question, that finds that for it, a lot of companies are attempting that right now. Again, I don't think it's realistic for these companies to think they're going to go compete with a Claude or slash anthropic or open AI like these companies are just moving so fast. Have such good talent, you need to build on top of those and build systems for intaking really good content. Have a hierarchy. And the third is practice. And so there's probably 50 AI scenario coaching, role play. You just mentioned you're at a conference like, How many times did you see? Hey, there's an AI avatar, and it will talk to you right that space is kind of the wild west right now, but essentially there's going to be a consolidation of that, and there's going to be, I think, a few key players that emerge. And the benefit there will be that again, the point of push is Get me the right info. That's going to be where most of my basic learning happens. The point of pull is when I have questions, I need to find things like a source that the point of practice is where AI role plays, plus the human element for a high level of expertise is going to come in. Ryan's prepared. He's learned the basics. He knows what's going on. Now he and Chris can have a really intense dialog to make sure that he can get to 100% Yeah, that's
Chris Rainey 22:14
kind of what we're focusing on right now. I was playing around the team sent me over as part of the learning pathways, the role play element with the with it, especially with the latest release for an open AI, it's unlocked a lot a lot more. And I was doing like a role play of like, how to give feedback to the to the agent, and it was just like chatting back and forth, and then it gave me feedback about how well I did. And said, Do I want to do it again? And I was like, wow, this is like, such a different experience to the the first two that you just mentioned. But it's still not perfect. It's still, you know, is still, you know, doesn't have something like require. It requires also like the context, because otherwise it could be just giving me feedback without the context of right of my current situation, the business, the industry, just everything.
Ryan Laverty 23:05
Yeah, I think, I think that's really where, you know, if you look at all these tools, like, if you go, if you go talk to, I've written white papers on this subject, and talk to people, who are, they look at that, and they basically say that that practice element is all we need, right? They're like, Oh, this is great. You know, someone can go find the info, and someone practices with them, and that's how they're gonna learn everything. The reason the push and the pull are so important to get someone to the practice is because a lot of times these, these academic studies on how people learn best are done in a vacuum, and the assumption in that vacuum is somebody wants to learn something, or knows they need to learn something, like, if you are, they'll go to like a Harvard classroom and look at the students there. The assumption they're forgetting is everyone a Harvard classroom wants to learn that and has chosen to give the time and invest right and they're paying for HR leaders. The core challenge is, and has always been, and in the world of AI will still be a significant percentage of your workforce probably needs to learn something, and either doesn't want to or doesn't know that they need to learn that thing. And that's where push is still the foundation, pushing learning, pushing content, whether it's AI based or not to people, is so important is because most people don't know what they need to know, and they are not in the top 5% of your most engaged learners who are just trying to practice all day. It's those folks where you're going to see the biggest impact from, from actually scaling learning and training.
Chris Rainey 24:24
Yeah, where do you I think I asked you this in New York, or where do you see you? You fitting into that landscape? Because you know you're not. I wouldn't call you an LMS. I wouldn't call you an LXP. I wouldn't call it a coach, because you got, you know, AI coaches every day, appearing as well, as well, like, where do you see yourself in the market?
Ryan Laverty 24:47
Yeah, absolutely, firmly, we're at the push layer of those three tiers, the push pull in practice, right? Which is the foundation, the the fundamental piece, again, that a lot of these things miss, is that you need the delivery infrastructure. You need people to pay. Attention to things before you go deeper. And so much of the focus today is on how deep Can we go? If you actually go, look at even the LMS of the world, they're seeing 1015, 20% engagement rates at all, engagement meaning someone went in and looked at the content,
Chris Rainey 25:15
right? Yeah. Not even wait for
Ryan Laverty 25:17
it, yeah. We see. We see 75% 80% of all even, even not counting, non mandatory questions asked actually get answered, right? And so think about the awareness component. Like if the way that someone learns something is they're aware of it. You space out over time, they'll they practice it. They retain that information if it's spaced. And then you can actually have them take action on that right, if it's in the right moment. And the challenge right now is that over 90% of content is never seen. The conversation never happens, and so everyone's going deeper with learning. You actually need to go a lot lighter and broader. And that's the space we exist in, in terms of where it fits in a stack. It's not a content library. It's on the LMS. You can even do certificates, right? It's a, it's a, it's a rapid, organizational, wide behavior change and transformation
Chris Rainey 26:01
tool. I love that. You've definitely said that a few times. But for a lot
Ryan Laverty 26:05
of leaders, they're like, I don't, you know, I don't know where that fits in my I know that's what I
Chris Rainey 26:09
mean. It makes it difficult for them to make a decision, because normally, if you come in and say, Hey, we're replacing your LMS that, oh, okay, that's what you are, and that's the line of budget you sit within, etc. And I see a lot of AI or companies in the space coming in, and because they don't fit within those traditional segments, they're like, the companies are like, where do we? Where do we, what are we taking away, budget wise, to invest in this? Or do we have both, like, there's a sort of confusion, like, you're almost creating your own brand new segment, which is both innovating, which is great, but also a challenge, because then you're like, where do we fit in terms of that, that budget line?
Ryan Laverty 26:48
Yeah, absolutely. Companies. I think in the startup world, companies love to say, Oh, we're creating a new category. I'm always careful of that, because then you have to just do a bunch of demand gen and build awareness around something and then sell it. Right? It actually be a lot easier if we were like, Oh, we're in LMS, but we do this, right? But fundamentally, again, think about the fact that for the past 20 years, most of these budgets have been structured around you get a content library and a system of record like an LMS or an LXP of some kind, and even AI role plays and coaches, technically, they don't fit in that scheme. Everyone looks at them and intuitively like, if you need to replace something, okay? Technically, it would replace a manager or someone that's coaching them, but it's not an AI coach is not a manager, right? A manager is doing such a wider range of things, and an AI coach is very specific, but it's enhancing capabilities. And so I think organizations, we're seeing them from the budgeting perspective, just take these large AI initiative budgets and go case by case based on, can you make the case? This will 10x or ROI. You just,
Chris Rainey 27:51
you just literally led to my next question. I was speaking to sutra this week on the show, and they were using one of the big LMS, is that we all know. And they were talking about how, you know, they've got Coursera, plugged in, Udemy, LinkedIn, learning, you name it. They had everything. And it was a lot of money, right? And so I was like, basically 90% of the value of your LMS is actually just these content, you know, resources you're bringing. What do you think this means for the future for those type of organizations, because most of these LMSs really are reliant on those content sources.
Ryan Laverty 28:29
Yeah, I think content is getting a lot more commoditized, and the expectations are shifting. And so what I mean by that is, you go into a library like a LinkedIn learning today, there's 100,000 courses, and none are fit for purpose and fit for you. The benefit of AI is that instantly, again, delivery matters, as we've talked about, but the benefit of AI is you can go in your expertise level is a two out of 10. You know this, but you don't know this yet. You need to practice this specific thing right now. If you want to go take a course on how to be a great sales rep, you've got to go hunt through videos and articles and playbooks and understand what you need to know and go find that and sift that's just far too much friction for someone to actually go do. That's why the engagement rates are so low. But the benefit of AI is in a world where it knows those pieces of information, it can personalize and serve that up to you. The challenge for these large the challenge for the large content companies is that's all getting replaced. There's going to be a world in the not too far future where companies are like, we don't need a content library, because when someone when someone needs something, they go find it. Is the whole reason people buy LinkedIn learning. But that also can arguably be described as a more advanced version of chat GBT, when I need something, I go to chat GPT to find it,
Chris Rainey 29:40
right? And it can create it in the context of your business, in the context of the role based around your specific product, so that sales training is specifically around that particular product, right? So, you know, and this, and it's level set, it to me personally, where I am mixing medium around in any language, right? So. It's new
Ryan Laverty 30:00
and it's relevant. I mean, so we, I'll give you kind of a teaser here. We, we were testing some of the early teammate capabilities with ourselves. We're in a demo with a customer, and we actually, we actually were testing it just on our cell phones because we were on the go. And so we set something up with our Product team where we can actually build a full course by texting a few prompts to a cell phone number. And again, this isn't something that most of our customers are sitting at a computer when they're building things, but the whole point of that, you know, I was sitting in a meeting and I really wanted to work on a specific thing that I was trying to teach myself. And so I thought about it, and I was like, well, technically, I could do this. And so I texted this number, and I was like, hey, I need a course on this thing. And so, you know, it asks me, okay, what's the what's the audience, what's your expertise level? I could just ask for some basic prompts. And then it said, Hey, your course is ready. Give me back your number, and I'll send you that course. Right? And I think there's a world in the not too distant future where, if Stage One is there's a bunch of content, learners go find it. Stage two is, admins build that content, and learners go find it. Stage three is admins can prompt that contents map you have to push right. You can't assume people are going to go do that, but learners can say, here's what I need, and they can get the perfect course back. That's really enticing as well. You know, I think that content libraries are going to go away. It's going to take a long time, just because people don't change as fast as technology. And then I think that the LMS LXP dynamic is going to get heavily consolidated some info, and LMS is valuable to an AI and then there's going to be aI interface that sit on top of them. But the current model of most LMSs today are like a quasi system of record and content library. I don't think organizations are going to be as reliant. And I think the few LMS is that do well will be the ones that can drastically change to build content a new way, not just sitting content sitting in a box somewhere, but that actually most of their value is going to be as a data source, right? EG, hey, we've got so much Employee Info, you can train your models on it and run your models in here. Yeah, I'm
Chris Rainey 32:01
seeing them. I mean, I'm speaking to most of the heads of product of those companies. They're all scrambling to do exactly what you just described. Or they're acquiring companies that are already they're already doing it as well. I love the idea as well, and of the fact that you have your internal learning pathways that you create, but then giving the employees the ability, hey, I want to learn around x, generate this and be able to create their own learning pathways. Do you see a future where then they would also have the ability to then share that with their colleagues, like, Hey, here's like, a learning pathway I built. Like, what do you think show the rest of the team?
Ryan Laverty 32:41
Yeah, yeah, absolutely. Again, the big focus for us, that's actually very easy to do, that actually would exist now. Like, if I built a course, it would exist, and anyone could go take it. The challenge we actually find is, it really is in the discoverability and the engagement with most learning. Like, the challenge is that there's too much content, you know, if, if you and I are all generating courses all the time. Yeah, and I want to go course on sales. 101, you've got yours, I've got mine, right? So, so a lot of our philosophy is push things to people. Keep it really simple, give them a few core options, if any options, and just start, just start the experience. Rather than Ryan, go find, go surf through these 25 courses to find this thing. Just Ryan. What do you need to learn? Oh, it's negotiations 101. Your manager said you needed to be a better negotiator on procurement calls. Great. We're going to serve this to you over a few days.
Chris Rainey 33:30
Do you think that you'll get to the point where in these platforms, there won't be any content library interface whatsoever, and it would just be the AI chat, because it's so good now that you don't need to browse through, like a YouTube, you know, or LinkedIn learning. Here's like 20 courses, and I'm scrolling through. Why am I going to waste time doing that when I could just ask the agent? Do you think that that whole interface will disappear and you'll just have the agent, and you'll just say, This is what I need?
Ryan Laverty 33:57
Absolutely I do. And I'll give you a really specific example. Think of, think of, think of the experience of searching on Google that we're also used to for 20 years. You google something, and then there's a bunch of links, and then you scroll through and you find the link, and
Chris Rainey 34:09
you click. You just gave me exactly. Now I just reach out TBT, and I don't even do any of that
Ryan Laverty 34:13
exactly. And so it's going to, it's going to take time, because, again, take longer to adapt than consumers. But think about how innate the behavior of googling was for you, and I think about how fast we moved to chat GPT and everything. I
Chris Rainey 34:27
don't do any. I don't use Google at all. I don't either. I ask exclusively. I mean, because to your point, I get exactly what I need instantly. And the more I use my agent, the more it gives it to me in the context of what it already knows. So it's even more tailored, not just what the sponsored ads
Ryan Laverty 34:48
libraries are, Google, right? I've got to go in there. I've got a search. I've got a search about, oh, this, Oh, this isn't exactly what I want. So interesting people, they're Google, but every link click is not reading an article, it's watching a 50 hour video. Do you think so? There's.
Chris Rainey 35:00
Do you think we're ready for that, though? Do you think there's still a need for people that want to browse? It's like, that's why I say
Ryan Laverty 35:06
consumers change fast, and people it will take time, right? Netflix is a great example. Netflix isn't going anywhere away anytime. That's true. Yeah. How many times have you not watched a movie because it took you forever to find what you want on Netflix, right? But, and they have a multi billion dollar content budget, but it's, but it's Netflix. Is the I'm sitting back Laissez Faire and going through it, okay? This thing, this thing, this thing, right? Like, there's kind of an experience to searching, and at the end of the day, content libraries in a company are not that. I don't, I don't enjoy the experience of sitting back and browsing, and maybe some subset of the population does, but my learning shouldn't rely on that. That should be a nice to have entertainment thing that sits in the corner,
Chris Rainey 35:47
yeah, beyond when you send it out. I love this conversation, by the way. I love talking to you like, this is such a like, it's such a cool topic. Beyond pushing it in the flow of work, how do you then keep the learner on, you know, like engaged. You send like, reminders, like, like, oh, you know, you left off here, and you just send a message trying to get them back in. Like, what does that look like?
Ryan Laverty 36:13
Yeah, there's, there's logic built into the system. And so essentially, what happens is, there's, there's a logic to, kind of when to nudge you back, you into remote use the also, the big part of this too, the content is built from the ground up to be as engaging as it is informative. And so it's not just the emoji use or the short form content, but it's if you actually look at how the courses are structured, you know, this is a thing that only I will notice. The easier questions are almost always first. And so there's a lot of logic to the way that the AI will build all the content. Like the thing people forget is that instructional designers have an academic background. And I can say it's, you know, I've written white papers with folks who have Masters instructional design, and the academic background is optimized for what questions do I ask? How do I get someone to the optimal outcome? They are not user experience people, they're not product people. You need to marry those two things, if you just have a product person, build a course, super sexy, super it's like masterclass, super sexy, super engaging. Doesn't actually teach me anything. Masterclass courses, sorry, passive content. They're not, they're not structured in a way that engages you, causes you to apply to scenarios, right? Yeah, instructional design courses are the polar opposite. You need both. And companies don't have product designers building their learning. They don't have instructional designers building new types of product, right? And so in AI, can
Chris Rainey 37:33
do both, yeah? Oh, man, I'm so excited. Yeah, no wonder you're so excited about what you do every day. Where do you like once teams comes out, teammate comes out, what I know that's already like, super exciting. What's kind of, what do you see the five years from now where we're going to be with this?
Ryan Laverty 37:52
Yeah. So the exciting thing with teammate is that it really is an open canvas for the orchestration of learning and enablement end to end in a company. And what I mean by that is, okay, great. We can build your courses, we can do needs analysis, and it can orchestrate that end to end. What happens when we can finally achieve sitting that on top of a document library or a company's corpus of information, right? Imagine that you're the CEO of 100,000 person company, and you can go sit down at your desk, and there's a chat GPT style interface, and I can say, hey, how well are people understanding this? And it can give you an accurate answer. Okay, well, what should What should we do accordingly? And it can give you an answer. Our competitor just launched a new tool yesterday, run the end to end enablement program that's going to get everyone up skilled on all of that instantly, right? And yes, you can talk about kind of a dystopian future where there's not a human doing that at a computer, the AI is just sentient and knows and can do all of that. I think that's really far out. But what's exciting is that, again, most of the work that's done in a company, and the reason why big companies move so slow is because of people, orchestration and communication, figuring out what's going on, what do people need, and then getting that to them, and AI can do all of those things, leaving humans to just make high level decisions. And so that's what's really exciting to us. Is okay, I'm a CEO of 100,000 person company competitor launches a new product. I need to be ready for that within if we have aI on top of engaged delivery structures, again, like a teams or text, right? So that I can get people to pay attention instantly. And I'm sitting on top of this massive corpus of information that a company has, and it's well indexed, and there's a good hierarchy to it. The end result is I'm essentially like a conductor playing the orchestra, where I can say, look, we've interviewed all these people. We know exactly we need. I can, I can, in two days, get the entire organization ready for something that used to take me a year and a half of planning and that level of organizational agility end to end is just
Chris Rainey 39:45
incredibly exciting. Yeah, and that would also require a whole cultural transformation. Yes, right? Why? I
Ryan Laverty 39:54
always hear me say people change slower than tech. Like a lot of the tech for this stuff, it's actually right. It's already here. Yeah, you're right. It's the you want to design with companies, so you build it in a way that it's a quick yes from the security and procurement team and from the stakeholders involved. And that's what takes more time than can we get the tech to do the thing, which is what I think a lot of these startups are missing
Chris Rainey 40:15
right now. Yeah, man, I could talk to you forever. Is there anything we haven't spoken about that we should have that we missed.
Ryan Laverty 40:23
That's a great question. Organizational agility. You hear me say that term a lot. Last thing I'll add, it becomes really obvious to people in moments of crisis, don't wait for a crisis. I'll give you an example. Our most used use case here in the United States for about a month was tariff communications, and the only people could use it, where our customers already brought in AI tooling, all this delivery infrastructure as soon as it let's say I've got a global supply chain. I'm a large auto manufacturer. As soon as something changes, I can get information everyone rapidly. I can upskill people really rapidly. People are going to feel the pain of their learning organizations not being agile enough when a competitor comes out with an AI product, when tariffs hit the US and change every week, and your your supply chain has to change, right? And so I'd say for leaders, the world we're living in is not one where this crisis just happened to come up. It's one where, when you consider AI, critical info needs to get to people, people need to upskill far faster, and that's only going to keep accelerating because of technology, and so you need to be ready to get people up skilled in days and weeks, not in months or years. And that's only going to the window for that is only going to get smaller. Man, I love
Chris Rainey 41:30
the agility piece, where can people connect with you? Man, if they want to reach out to you personally, and then obviously, learn more about the organization.
Ryan Laverty 41:37
Learn more about ai.co my email is just ryan@ai.co maybe I won't want to give that out giving you guys. To give that out my cell number yet. Don't do that or find me on LinkedIn.
Chris Rainey 41:49
Yeah, no worries. We'll link the link below to your LinkedIn and the website, man. But I always, always play to talk with you. I absolutely love what you're doing. It's super inspiring. And, more importantly, terrific you've been saying is truly transformational as well. So keep up the great work and look forward to chatting again
Ryan Laverty 42:06
soon. Awesome. You too, Chris, thanks so much. You.
More from the HR Leaders Podcast
Will Leahy, VP of People at Greenhouse.