Episode Transcript
[00:00:00] Good morning. Today I want to talk about Workday's acquisition of Sana. Now, Sana, or Sana Labs, is a company that we partner with directly for Galileo. We have worked with them for almost three years. They're a relatively small, but not tiny learning and development and knowledge management company. But at their core they are an AI expert. They pioneered in my mind the transformation of learning and development towards an AI dynamic content strategy. We use them for our agent, Galileo. Their agent, Asana Agent, is essentially like ChatGPT, but really a shell on top that allows you to use ChatGPT or any other LLM and integrate your own content in your own systems into it. They have essentially two platforms, the agent and the learning system. We call them Galileo and Galileo Learn in our implementation. And they're a great company, young people, smart, Swedish, very focused on elegance, beauty, functional design.
[00:01:08] The system is not cluttered up with lots of things that are hard to find. They work very hard to make sure that it's easy to use, it's very fast, it's very well integrated.
[00:01:18] They've worked extremely closely with us to optimize the Josh version, company Corpus and all of our data partners. So Galileo is a very high functioning application and they're now going to become part of Workday. Now, Workday was one of the venture investors in Sana, so they've known Sana for a while. But I don't think the Workday engineering team was fully aware of Sana's potential until maybe recently.
[00:01:45] Workday had been using Sana in their internal L&D department. But most of Workday's engineering has been on workday learning. And Workday learning is in somewhat, we used to call it a bag of doorknobs, which maybe is a little bit too harsh, but it's a very old sort of archaic system built originally through the acquisition of a company called mediasite that was a video platform. So the original vision of Workday learning was not to be an LMS at all, but to be a video distribution system. And of course what they found was that that's not really what people wanted, that was only a feature and they had to build an end to end learning platform with all the features of lmss, which are quite complex. On top of that, now the LMS market is very complex and very difficult. So let me talk about that. I started working in learning management systems in 1998 when I was part of a startup that was trying to build an online learning system. And at that point in time there were venture capitalists Investing in companies like Saba, Docent, Click to Learn, and many smaller companies that you've never heard of that have all been acquired or absorbed into bigger companies. Because at that point in time there was a massive interest in learning on the Internet. And I would say we didn't really know what that meant. And so what we did as a global educational and training and business community is we took the paradigm of instructor led training and we converted it to the Internet. So we built courses similar to the way a teacher teaches a class. And we built chapters and pages and interactivities and tests and assessments, and then later simulations and then much later videos to allow people to learn and educate themselves online using the same pedagogical paradigm that we used in a classroom. And most of you are a lot younger than me, but during the first 10 years of my career, I spent a lot of time on airplanes going to training and sitting in classrooms listening to subject matter. Experts, instructors and executives teach us, quote unquote, what we needed to know in our jobs. I did this at Exxon, I did this at IBM and it worked great because you'd go to a class, you'd meet a bunch of people, you'd learn a lot, you'd do some simulations, you'd go back to your office, take the workbook with you, and if it was something technical, you would have the workbook to go back and look at. And I was a systems engineer for five years at IBM working on mainframe stuff originally. And it was very technical and it worked. The Internet came along and the big benefit of elearning as it was called, was we didn't have to spend the time and money flying people around to classes. So there was an instant adoption of online learning simply to save money and improve speed. Now it still took months and months to build courses, and there weren't a lot of tools in the early days to do it. So you had to build a lot of handcrafted expertise in HTML and other tools to use it. And the courses were kind of complex and kind of long, and there were no platforms for those courses. So the learning management systems became not training systems, which used to be used for scheduling classrooms, but administration systems for this online learning experience. And SCORM S C U R M, which is the updated technology from an old aviation technology used by the airline industry to train people using CD ROMs, believe it or not, was the standard that was used to tell the learning management system what the learner was doing. And didn't tell them much, but it told them whether they started the course, how many chapters they'd gone through, what score they had, and perhaps if they completed things like that.
[00:05:33] So we. So this big industry of LMSs was born and it was brutal, it was very exciting, very competitive. Plateau was acquired by SuccessFactors, which was later acquired by SAP. Saba was essentially in a shooting match with Cornerstone. Cornerstone ended up outperforming them and ended up later buying Sava. But meanwhile, Sava bought a bunch of other companies. Oracle bought several of the companies. Taleo bought several of the companies. Taleo was then acquired by Oracle. So there was this period of, I'd say a decade where a lot of this learning technology was central to the HR tech market. And a lot of you remember that if you were around in the 2000s, you know, the first decade after 2000, okay, so what happened next was we got a bunch of new technology, we got video, we got phones, we got YouTube, we got Twitter. And these platforms got more complex. And so the vendors, the LMS vendors tried to adapt to all these new technologies to create social experiences, video tools, opportunities to infer skills, do skills based learning. And by the way, one of the sort of complexities of the LMS market is customer training, where you have to charge somebody for a course and they get credits and they can buy training certificates and so forth. So there was an E commerce part of this that had to be built.
[00:06:57] So vendors tended to differentiate themselves from corporate versus what called, what was called extended enterprise training. Sales training became a market for learning management systems, et cetera. And in the middle of all of this, of course, there was this big use case of compliance in industries like financial services and pharmaceuticals, where you have to legally track what somebody is certified on before they can actually do the job. And that is a mission critical, sort of unstoppable part of a company. So you can't rip that type of thing out once you have it. So while all these features were being built, the systems got harder and harder to use to the point where, I don't know, somewhere in the mid-2010s or so, people were so disgusted with their LMSs that they used to say things like, why can't we just stick all our courses in YouTube and just not use this thing anymore? And then there was this flurry of interest in what were called learning experience platforms, LXPs. The LXPs started out as a simple portal on top of this mass of content in the lms. By the way, you know, an LMS has thousands and thousands of things in it. Courses, documents, videos, Et cetera. So the LXP market, which was eventually dominated by Degreed and edcast, became the next big thing to try to make this system, which was basically designed for administrators, not users, easier to use. Because, remember, learning management system means it's for managing learning, not for learning. They were designed for the manager or the administrator, not for the user, because the learning experience was in the course itself.
[00:08:39] This is all going to make sense in a minute. So the LXP market takes off for a good five years. There was a lot of excitement about it. The LXPS also started the idea of a skills model, because what you could do in an LXP is you could take this gigantic pile of content that you had built in your LMS and you could tag it by skill, and then the user, the employee, could search by skill by, you know, level of proficiency within that skill and so forth. The idea being we're going to sort of take this legacy LMS and make it much more functional and easy to use and more adaptable. And that went fine for a while. Microsoft Viva Learning was essentially one of these. LinkedIn built one of these, by the way, in a parallel universe. While all this infrastructure was being developed, there were hundreds and hundreds of content companies building courseware. And in the early days, they were really hot. Later, they became commoditized to the point that, you know, you could get content in dozens and dozens of different places. And it was really only the reputation and validity or trust of the vendor that would differentiate them because there were all sorts of free courses all over the place, too. So the LXPs worked for a while, but of course, what happened was they got cluttered up, too.
[00:09:53] So big companies like Citibank and others used to show us their LXPs, and they'd say, well, you know, we've got this stuff all in here. And we built. We gave a version of it to the sales group so they could build their, you know, learning stuff and another group. And, you know, now we've got this big mess of stuff to manage in this lxp, and then we have to manage the LMS behind it. So, you know, every time we want to publish a course, it takes three weeks to just get it into the system because we got to put it into all these other systems. And all of a sudden we've got, you know, a hundred people in instructional design building courses, another hundred people in India publishing courses, and then another hundred people analyzing and measuring and running reports on all this stuff to see if anybody's even taking anything. And the user experience is terrible. I was at Deloitte where they use SABA under the covers of a bunch of portals. And I remember that once a year when I had to do my compliance, I had to basically allocate half a day to figuring out how to even get into the courses and slowly click through them to get certified on the various regulations that we had to do at Deloitte. So no fun, not really. Very strategic, very expensive. And HR departments basically said, you guys over there in L and D, just do your thing and leave us alone. And what happened as we talk about in our Revolution paper is the HR function moved in a new direction. And so the priority of learning, which was huge in the early 2000s, in fact, chief learning officers were peers of the Chro learning, got demoted and demoted and demoted. So that the corporate learning team was really, you know, an important team, but they mostly worked on onboarding or leadership development or sort of other corporate stuff. And everybody else did whatever they could do. Companies, the sales organization had their own learning stuff, the manufacturing groups had their own learning stuff, the IT group had their own learning stuff. And the company was spending a lot more money on learning than they realized. But they couldn't do anything about it because there was no sort of integrating technology. And the vision of one LMS doing all this was just very poor and difficult. Nevertheless, that's the way the world works today. And I think you would find my discussion here true in every single major company.
[00:12:01] Now for small mid sized companies, there are LMS vendors that are easy to use, integrated systems that have been built now thanks to SaaS that don't require this level of complexity. But once you get big enough that you have multiple divisions and multiple locations and lots of different languages to translate and so forth, even they become very complex to manage. So there's this big army of L and D professionals that do this. And by the way, L and D professionals are very smart, very well educated, very technology savvy people. They're, they're some of the most creative people in the HR profession and they're just trying to do the best they can with all the tools in the market. And they tend to be very good at trying new things out because there's a million experiments going on in corporate learning and different forms of education all the time. Okay, so that was all sort of what was going on up until around, I'd say, 2022, 2023.
[00:12:59] And then, you know, these big platforms like YouTube and the Talent intelligence systems like Eightfold came along and everybody said, and I was kind of pushing against this a lot. Let's do it all around skills. Maybe the whole company should be run as a skills based organization, which, as you know, I hate that phrase, but it stuck. And let's take all this content and organize it by skills. So let's take the skills tech, let's go by Eightfold, let's go by gloat, let's go by, you know, whatever tool you want. And now these LMS vendors have to build skills models and they have to go collect thousands and thousands of skills data into a big taxonomy, get it the data from Lightcast or somebody else. And so a learner or a user could go through their learning experience in their career based on the skills they have and so forth. And supposedly theoretically, this LMS is going to know what your skills are because it's going to know what you did and it's going to keep you upskilled all the time. That was more or less a flop. But you know, it sort of made sense at the time because we didn't have AI. I was skeptical of it from the beginning because people don't really learn things at work for skills. They learn things at work for productivity. You know, when I want to do something at work, I talked about this last week, I want to know how to do it. I don't want to learn PhD in the topic, frankly, maybe I do on my spare time, but right now I want to learn how to do this thing that I'm trying to do at my job. And I want to know how the best practices of others in my company do it. I don't need too much theory, I need a little bit. But so I don't really need a, you know, 17 level skills model on how to write a good blog. I would like some help on just doing it. And then later, if I have time and I decide I want to become a blogger for the rest of my life, I'll. I'll go through more education on blogging, for example. So that happened. I would say it was another flop, but you know, a lot of companies did, you know, pretty good stuff with it. And meanwhile the L and D department's mired in that. But most of the HR attention was really put on recruiting and employee experience because we went through, you know, an economic cycle in 2008. We went through the pandemic, we went through the shortage of technical skills and most HR departments could buy libraries of technical training. They didn't have to really build it, but they needed help in recruiting and they needed help in sourcing. And so the effort and intelligence and sort of energies of HR went into that. And the L and D departments still did their thing, but the clos didn't have that much to do. So a lot of the operational training was sort of flung out to the, you know, to the. To the line of lines of business where it belongs anyway, because, you know, I've talked about this many times. The best corporate training happens close to the work, not in the corporate training department.
[00:15:47] So, you know, we want a way to enable people to do their jobs. I really think the new word for corporate learning is enablement, frankly, not learning, because we don't. We aren't really here to learn. We're here to enable people to do their jobs better, enable them to grow, enable them to get promoted, enable them to become leaders. So. So I'm going to, you know, help you guys understand that. So along comes AI. And by the way, in a parallel universe, there's a couple of people out there like Joel at Sauna and others that are scratching their heads and thinking, you know, maybe AI could be used for knowledge and learning. By the way, during this whole period of time, there's another world called knowledge management. We got involved in knowledge management in my old company. It's a very old domain that mostly focused at the time on search engines so that you could go into a corporate IT department and search for things and find them, documents, videos, compliance, standard operating procedures and so forth. And that was always sort of parallel to learning. But the learning department couldn't do much with it. And when I wrote my first piece and did my first speeches on learning in the flow of work, which was something I, you know, sort of started talking about a long time ago, I knew from my own business experience that ultimately the only way learning was really going to propel itself to the next level of impact was to be more like knowledge management. But there were no technologies to do that. Now we're talking about a huge industry here. One and a half, 2% of payroll in many companies is spent on training. It's a lot of money. It's 300, 400 billion at least. And that's a little bit of a hard number to track because a lot of that money is in books and magazines and people going to conferences and stuff. That doesn't really track as the L and D budget. So we're not talking about a sort of sideline business here. You cannot run a company without training, and you cannot grow and adapt and maintain and retain your people without training. So, you know, this is a big deal. And when it got sort of demoted into hr, I think that was a problem because it needs to have a life of its own. It needs to have an owner of its own. And I think we're going to go in that direction because of AI. Now, what SANA did and what SANA really pioneered, and the reason we chose them in our world is they started from scratch and said, let's rethink the problem as a dynamic content problem. Now, we never had dynamic content generation technology until AI, but we did for a while recently, and then it came out as ChatGPT, of course, but it was, it existed before that. So Joel and his co founder at Sauna, you know, sort of built this dynamic content system before ChatGPT came out. And then when ChatGPT came out, they immediately pivoted it into a chat tool and a learning tool. And so what happens in AI, native learning is what we call it. And we're actually quite, quite good at this because we have a, we have a system. Our Galileo Learn has 750 courses in it that we've brought with us through our entire journey as a company. And any of you that want to learn about this, we'll show you exactly what we did and how you can do it. You can take any form of content, courseware, videos, audios, interviews, documents, et cetera, and load them into an AI platform and generate courses dynamically. You can give it templates, you can tell it what kind of instructional activities you want to have it do. Those are called cards in sauna. But. And then the user, the employee, can either go through a course the traditional way, or by the way, the system can translate it to multiple languages, or they can just ask the course a question. And in our case, there's an AI tutor that pops up in every course. And I was showing a demo to somebody yesterday. We have a course on a lot of stuff on well being. We have a course on Cadbury's wellbeing strategy. Cadbury is a company that's done well being for a long time. So I go into that course and I show the person I was demoing. I said, who's Cadbury? What's the history of Cadbury and wellbeing? How did Cadbury define well being? How does GE do well being? What's a well being strategy at Microsoft? What's the relationship between well being and learning engagement? What's the relationship between wellbeing and leadership? What's human centered leadership? Blah, blah, blah, blah, blah. It answers all those questions, and it's getting the answers from the corpus of content.
[00:20:05] So now the value of the learning platform is the content, because everything can interop with everything else. Now, this sounds a little bit too good to be true, but I'm not kidding. This is what it does, and we'll show it to you. You're a manufacturer, you're an airline, you're a retailer, you're a consumer products company.
[00:20:26] And you, you know, you're your l', Oreal, whatever. You gotta train a whole bunch of people on a whole bunch of stuff. And you've got all these different products and all these different offerings, and nobody knows where to look stuff up. The system will take all of that information and turn it into one massive corpus, and you can literally ask the learning platform a question. Now, the way we position this and the way I think about it is all the learning stuff that's ever been developed for the last 25 years or 30 years that I've been around in this has only resulted in maybe less than 2% of the level of interactivity and consumption of one year of ChatGPT. In one year, ChatGPT has grown to the first year about a billion users. And the reason for that is that we humans like to ask questions. And we don't always know what questions we want to ask. So we'll ask a question and then we'll ask another question, and then we'll dig in and then we'll go back later and finish. We don't like linear learning at work. We don't mind it in school because that's all we have to do. But at work, we have other things to do. So the way we learn at work is with fits and starts and big ideas and little ideas and tips and sometimes gigantic, you know, breakthroughs. And we want to learn in our own personal way. And so the way we think about Galileo Learn is this is your personal learning tool. Yours. It learns from you. It knows what you've done. It knows what skills you have. By the way, these AI systems like Sana build the skills taxonomy automatically. So you can give it your taxonomy and then. But if you don't have one, it'll figure out what it is, and it will tag content based on the language in the content, and you can decide what level of proficiency you want to use. We have three levels of proficiency in ours, so all of a sudden we've got this system that is likely to have two orders of magnitude higher utilization than the lms. What is it? Is it a learning system? It's really much more than that. It's an enablement system, it's an empowerment system, it's a growth system. We, you know, in Galileo we're coming up with a new tagline for Galileo. But it's something along the lines of get ahead, stay ahead. That's what people want at work. They don't just want to learn how to do their job today, of course they need to do that, but they want to do, they want to know what's next, what can I do to progress, what can I do to be better, what can I do to get promoted, what can I do to make more money, et cetera. So these systems are going to unleash this new idea and we may call it learning and development for a while, but I think that is a very dated concept and data dated name for this function and it's going to go beyond that. And I can't predict what name we're going to come up with, but the best I can think of is enablement. Because if you think about the super worker AI powered world that we live in, and this goes for frontline workers too. You know, somebody works in a manufacturing plant or a nurse or a coffee shop, people are very concerned about the impact of AI on their jobs. They're very concerned about automation and how they keep up. How do they learn how to use this stuff, what's the impact of it and how do I stay ahead and how do I, you know, sort of burnish my personal skills to stay relevant. I just had a big call yesterday with somebody about this. You can learn about your job and your career using an AI platform like Galileo, learn like Sana, and this is a new paradigm. Now architecturally, these are different systems from lmss. They don't use AICC and Scorm. You can load that kind of content in, but they're dynamically generating content.
[00:24:04] So when you, you know, go in and ask a question or try to take a course in another language, it's dynamically generating the content that you asked for without somebody having to go do instructional design. So honestly, the instructional design community is going to get smaller, it's going to change. You're not going to be able to sit down and take six months to build a course anymore. Nobody's going to have time for that. You're going to use these AI tools to build content. And in the L and D profession we're going to have a lot more people looking at utilization of content, quality of the corpus structure. What are people asking for? Where can we better Find subject matter experts. I mean, we've actually been advising clients to interview their experts in just an audio or video and stick it in the system to provide knowledge sharing. Now imagine that happening in near real time. Imagine that happening in an environment where the learning platform is available to functional business units, not just L and D. And somebody in sales says, hey, I just figured out how to beat this competitor and what is happening in this market. I got a 10 minute discussion I want to tell you guys about how to do this. And they stick that in the system and all the other salespeople can find it the next minute. You can't do that anywhere else. You don't have to. You know, even things like surveys, you know, authoring a survey, distributing a survey, getting people to take the survey, analyzing the survey, that's like a big pain. You can do that kind of stuff in these systems. And the way, in, the way I think about it is you're not going to really interface with this system through a portal. You're going to interface it through your own agent. So you're going to use Galileo or your, you know, favorite agent that you're sitting in, and you're just going to ask the agent a question and the agent's going to go into the AI learning system and it's going to answer it for you, or it's going to show you a course, or it's going to show you a video, or it's going to create a podcast for you. And you can literally go to Galileo, learn, and you can say, I'd like a five minute podcast on pay equity and I'm already a compensation expert, so give me sort of the expert view and it'll create an audio for you and you can just listen to it in the car. You can't do, I mean, you don't, you just can't do that in these old technologies. So Workday is in for a big surprise. I think they kind of know what they're getting into, but not really yet, because they have a big team of people working on the old Workday learning system. And I don't think they can do away with it, nor do I think they should, but they can layer this new technology on top of it. And we are, you know, very, very familiar with this stuff and have been doing it for a long time. And I think Workday is going to go through a learning curve here on what to do with Sana, how to position Sana and what I hope, and I know these folks very well, and I think Sana will do the best to prevent this from happening. But what I hope they don't do is try to turn Sana into a part of Workday learning, because it isn't. That isn't what it is. It's really the new generation of learning and knowledge and enablement technology for corporations. Now, Workday is a big company. They have a lot of different customers, and they'll come up with strategies, and we're going to work directly with them as much as they'd like us to, to figure that out. But this is a very sort of transformational deal in some sense to a major part of Workday's business. Now for Sana, it's great. I'm a little bit sad that they're not independent anymore, but we're going to continue to work with them as much as we can, as long as we can, because we're so committed to them. They're young people. I think part of this is an acqui hire to get this AI expertise into the workday team. You know, so if you think about the acquisition of Hired Score, which was an AI team, the acquisition of Paradox, which is an AI team, the acquisition of Sana, which is an AI team, they're building a team of experts here in different domains that know a lot about AI, despite the fact that the architecture of Workday is not oriented toward AI at all. It's really a workflow architecture for the web, for the cloud. It is not an AI architecture. This helps them innovate in different areas of AI throughout the workday platform with some really, really good technology skills in the company. I guess for those of you that are org design people, you know, the big question, of course, as a vendor and a consumer of Workday, is who does what and how quickly can they make this system do these new things? And I would guess I don't know this for a fact. I think what Workday will do is they will leave Sana alone for a while and let Sana thrive within the workday customer base and see over time what the integrations or replacements are between Sana and the technologies in workday learning, because it's a little hard to predict until people start using it, how the what the interoperability questions will be. And I'm pretty sure that's the what what they will do, because they don't want people from Sana to bail out and leave. Workday has a great experience with this. You know, Pecon, most of the people at peak on are still there.
[00:29:00] I know quite a few. I have a lot of friends in companies that are acquired by Workday and they've tended to stay for a while because they're a good company to work for. So. So this is a pretty big acquisition from my standpoint, by the way. So is paradox. Same sort of story in talent acquisition. So is hired score in the pre hire selection stuff that they do. So we are here to help you understand it. As an HR leader, I think the best way to learn about this stuff is to buy Galileo and Galileo Learn and get yourself started. Not everybody can afford Workday nor does everybody need workday. So you can get access to a lot of this technology through us. If you're a small or medium sized company, you're an individual and we can help you with your strategy. Of course, in the middle of all these transformations, this is the week of HR Tech in Vegas. So you can hear me talk about this on Thursday. I'll be talking about it unleash in Europe in October. And we're doing a big announcement with SAP at SAP Connect. SAP is not left out of this. They're going to do the same thing that Workday is doing, I guarantee you. I can't tell you exactly how, but they're pretty sharp cookies too. I don't know what Oracle's up to in this area. I got to believe they're going to try to do the same thing somehow. And there are smaller vendors like Sana out there that can do some of these things pretty doggone well so you don't have to, you know, buy Sana to do this. Docebo has been working on this for quite a while. They're pretty far along and others. But the traditional LMS vendors have a hard time getting there from here because they're building on a platform that's mostly a management system and not an AI engine. A little bit long today, but I really wanted to get this all in front of you guys so you understand this and we look forward to.
[00:30:37] By the way, we have a course in the Galileo Learn Academy about this. So you can go in there and just dig around and ask it a bunch of questions and it's basically got a Josh bot and you can ask me questions in Galileo Learn if you don't have time to work with us directly. Thanks everybody.
[00:30:55] Hope to see a lot of you at HR Tech in Vegas.
