Keynote: an enablement team in your pocket

0:00 Hi everyone, my name is Michael Yafi and I'm Eeris co-founder and CEO. Incredibly excited to share some really exciting updates today um and share our vision for an enablement team in your pocket. Um for now, please feel free to share where you're calling from in the chat. We'd love to learn where where everybody is joining from today. Um and we're incredibly excited to get started. And please also feel free to ask questions in the Q&A section throughout. Um so we have a lot a lot to cover. Um, we're going to start today by talking about the state of enablement. Uh, and we're going to zoom out a little bit and take a look at what is going on in general for learning and development and enablement teams. Uh, and what the state of play is. From there, we're going to talk about the state of agents and the state of AI. Uh, we're going to dive into what the reality is today. Um, and what is possible and what's not possible. Um, and also, you know, what we believe the future is when it comes to agents. From there, we're going to dive into a live demo of Aerys agent suite and show you all how to build complex enablement programs start to finish uh in minutes using AIS agents.

1:01 We'll talk about the impact this is having for some clients. Uh and then from there, we're going to dive into what you can do today uh in from a more tactical point of view. So, we have a lot to cover. We have a phenomenal number of people here from all over the world. So, incredibly excited to to dive in. Awesome. First, I want to say a huge thank you to all of the companies and partners that we've been able to work with over the past few years. Um, AIS started a few years ago initially as a way to deliver learning in war zones and we've evolved massively since then. Um, and today we're very very lucky and immensely grateful to be trusted by dozens of the world world's leading organizations. Today, uh, Aerys clients use Aerys for some of the world's most complex product launches, for some of the world's most important AI upskilling and technology upskilling initiatives, uh, and to enable their people to be absolutely exceptional. Um, and we are so so grateful to work with some of the world's best brands and some of the world's best learning and development and enablement teams. Um, we we truly could not be at this point without all of you. Um, and our journey to today has been uh, pretty interesting. uh you know as I mentioned we started initially as a way to deliver learning in war zones um and back in 2019 we launched the first course in the flow of work back in 2019 we realized uh that the future of learning was going to be conversational right we needed to meet people where they were um and so working with a few professors at Harvard and Babson College we created the first text message course uh and started delivering learning via text message well before learning in the flow of work was a widely understood or appreciated concept um we very quickly realized that delivering and learning the flow of work massively increases adoption and engagement. Um, and so in 2021, we had the first enterprises start adopting AIS courses. Uh, GE was our first client back in the day. Um, and we saw very rapid adoption of learning in the flow of work largely because adoption and engagement rates were so much higher than traditional e-learnings or live sessions. A few years later, um, as AI was starting to become more more meaningful and more significant and we were lucky to get early access. OpenAI back in 2020, we started realizing that AI could be incredibly effective at creating personalized experiences at scale um and could reduce the workload of creating, you know, highquality learning design. Um in 2023, we launched the first enterprise AI course creation tool on the market um which was very very rapidly adopted by dozens of leading enterprises. Um and then a few years later, last year we launched the first medical grade AI course creator.

3:30 Um and so by medical grade this enables our uh by being medical grade this enables many of our clients to deliver hyperpersonalized content that is incredibly accurate and meets life sciences standards. Right? Uh at this point a courses are not only referenced and cited the same way that medical writers do but oftentimes have lower error and hallucination rates than humans do. And so today we're incredibly excited to share the culmination of all of our work and the culmination of our vision. an end-to-end agent for enablement and an agent that can do all of the core roles of an enablement team from needs analysis to creation to delivery to analytics start to finish uh using AI and using agentic workflows.

4:12 What's really exciting is that the combination of AI and learning in the flow of work makes end-to-end agents possible, right? Um, we fundamentally believe that work is moving into the flow and we'll dive more in depth into why we feel this way. But it's really this combination of all the progress that's happened from a large language model point of view and the ability to deliver content where people are and engage with agents conversationally that makes this all possible today.

4:39 So let's zoom out a bit and and chat through a few core trends and a few core beliefs that we have about where work is moving and where enablement is headed as well. So the first observation that we've had is that work is broadly moving into the flow of conversations and every single large enterprise software company is moving away from dashboards. This is one of the reasons why you know a company like Salesforce spent so much on Slack, right? Uh they acquired Slack for 27 billion a few years ago. This is one of the reasons why Microsoft is heavily investing in Copilot uh and in Teams. uh and we're noticing is that every single large enterprise software company is realizing that nobody wants to go into dashboards, right? And the right way to engage with software is conversationally in the flow of messaging tools, the ones we use every single day. There's a few reasons for this. The first one is that adoption rates are dramatically higher in messaging tools, right? People check Teams and Slack every six minutes.

5:34 People check a tool like workday or a tool like their LMS once every four to six weeks on average, right? And so if we can deliver experiences in the flow of work, adoption and engagement rates increase massively. The other consideration is that conversations are frictionless, right? We can in natural language say exactly what we want to achieve and have an agent do it on our behalf. Um, and so what we're starting to see is with every single major enterprise software platform, the ability to, you know, instead of having to go through 20 clicks in workday or in SAP to figure out what you want to achieve or to get to the right outcome, being able to just naturally say here is what I want and have an agent orchestrate that platform on your behalf. So what we're broadly noticing is work is moving into the flow of conversations and conversations are becoming the place where work happens.

6:24 The second observation is that the agents that win are endtoend and proactive. They're not just automated workflows. And today, a lot of agents are trapped in dashboards. Um, for those of you that work in learning and development and enablement, you probably have seen tons of marketing from major HCM providers or for major CMS providers talking about how they have agents embedded in their products. Now, from our point of view, that's largely marketing fluff, right? The reality, and we'll talk about this more in depth here in a second, the reality is that the agents that win are the ones that can do an entire job description end to end, right? Are the ones where you can say, "Hey, this is exactly the type of outcome that I want to accomplish." And the agent does everything from understanding how to achieve that outcome to executing that outcome and working with other agents and working with humans to get to that outcome.

7:13 Right? The challenge today is that so many agents, especially the agents that are oftentimes in HCM platforms or in sales enablement platforms, fundamentally aren't able to do complex complex tasks. They're they're not able to do multi-step workflows really well. What they really are doing today is just embedding AI in certain parts of an existing workflow, right? Um this also means that proactivity is heavily limited, right? The same way that the best employees are heavily proactive, right? they understand business context and are able to proactively you know work uh and create ideas and work on projects to get to that business outcome. The same challenge exists today with agents right the agents that win are the ones that can understand where the gaps are and proactively solve them in real time without needing constant guidance from humans. And then the the third core trend here and in our opinion the biggest shift is that learning communications and surveys are collapsing into one enablement layer.

8:11 Right? Today in most organizations figuring out what the gaps are is oftentimes done by an employee engagement team or a business analytics team. Right? Uh that team usually spends time looking at data, doing surveys, doing interviews. Sometimes thirdparty consultants are brought in and gap analysis oftentimes is fully separated from creation and delivery and essentially creating the intervention that can fix that gap. The our point of view is that this is going to change pretty massively. Right? Today different teams and different platforms create learning communications and surveys.

8:46 That's all going to collapse into one enablement layer. And our point of view is that you should have one software platform that figures out what the gaps are and creates the right intervention instead of having multiple different platforms, you know, figuring out what the gaps are and creating the right interventions. The same is true for teams, right? We we are uh already seeing massive consolidation here where a lot of employee engagement teams and learning and development teams are combining into one and a lot of sales analytics teams and a lot of sales enablement teams are combining into one as well. And we believe that this is the future of every enablement function, right? They should be able to holistically solve gaps, identify gaps, and run that flywheel over and over again.

9:27 And as a result, this is what the modern enablement team's agreement is. At the end of the day, every single enablement team only has two responsibilities in a post AI world. They need to find talent gaps and fix them as soon as possible. Historically, a lot of enablement teams and learning and development teams have been goalled by their ability to create content quickly, but also by the volume of content that they create, right? If even even looking back a few years ago, a lot of learning and development teams were gold by how many resources or how many courses were available to their employees. Same thing is true for sales enablement teams, right? We we know multiple sales enablement teams that until very recently were gold by the sheer number of assets that every single rep had access to. The reality is that today every single rep and every single manager and every single employee has multiple different places where they can go to pull content. And what where they're usually pulling content is in an you know is in a a tool like a large language model right an internal LLM uh or an internal knowledge management system or going to a platform like YouTube if it's uh you know a non-complated skill. Um and so learning and development enablement teams have had to compete with YouTube and chat GPT for attention. This is a losing battle, right? It does not make sense to compete with chat GBT or compete with an LLM or compete with YouTube. And so the enablement teams that will win the next decade are the ones that can focus on predicting what the gaps are and pushing the right content at the right time instead of simply creating a lot of content and providing resources. This is a major shift, right? enablement teams are shifting from people in organizations that just create content uh and and you know let reps engage or let managers engage to really pushing content at the right place and at the right time.

11:14 And that takes us to sort of what AI can do today, right? Um today AI is very very good at creating content with incredible accuracy. As I mentioned at the beginning of our conversation, um AI today is very very good at creating content with human level or better than human level accuracy. And that's largely because most modern AI creation tools combine multiple different sub aents that can create content, verify content, you know, do inline citations, etc. This is what a creator does quite well. Um and also AI is better at creating content uh with really really good outcomes compared to humans, right? The reality is that an AI tool knows all the modern learning science, all the modern enablement research, um, and has tons and tons of data on what creates a good and effective course experience. What that means in practice is that anytime an AI model creates a course or an AI course creation tool creates a course, if it's properly trained, the outcome of that experience is usually going to be dramatically better than the outcome of a human created experience. We've seen this firsthand with many of our clients.

12:18 uh today on average uh reps trained using AIS created courses. So AI created courses in Aerys see about a 2x lift in year-over-year sales compared to reps trained using traditional methods. Right? And that's because AI is very very good at implementing research and understanding what people need and how to deliver it in the most concise and effective way possible. One interesting side effect of this is that courses end up being oftentimes more concise than we're used to. they end up looking a little bit different than we're used to because AI is very focused on just giving the absolute essentials, not wasting people's time with content that won't be relevant. AI is also very good at summarizing and analyzing data very very rapidly. And we'll dive into how this is relevant in a second, but we've been surprised at how much effort has been spent on creation uh and how little effort has been spent on data analytics and pulling out insights using AI.

13:11 On the right, you'll see an example of what this looks like in practice for many of our clients today. Right today, an ARIS client can upload 10,000 pages of content, instantly build 10 courses on key skills and key product info based on that content. ARIS will automatically reference and iterate and then we can instantly translate into 30 uh languages, 20 role-based personalizations based on job description and then deliver all of that directly in teams chat so that people actually see it. Right? So already with current AI tools and current AI functionality, a six-month process can be boiled down to about six hours.

13:45 What's more exciting from our point of view is what AI can do today, but what's not widely distributed yet, right? What no one knows it can do. Um and what's interesting is you know we've noticed that there's been a a significant delay in terms of how many uh you how many AI uh features and and sort of capabilities have been implemented into modern enablement and learning tools. Right? One good example is a large public uh learning management system just last week released an AI course creation tool. The reality is that our industry and most of the software companies in our industry are is largely behind, right? And many of the modern advances and many of the most recent models model advances have not been implemented well, right? A few examples of this at play today. AI voice conversations are very very humanlike and will only get dramatically better. Today, AI voice uh can largely replace surveys and interviews and drive dramatically more in-depth and dramatically higher quality conversations. And we'll show you an example of this here in a second. What AI can also do today is do real-time needs analysis with both structured and nonstructured data. Right? You can have an AI agent that's looking through all of your structured data in your CRM and your HR constantly to figure out what the gaps are. Um, and then AI can also practically create and deliver as needs come up. And again, because AI has gotten so much better in terms of quality uh of output, right, we can now proactively deliver the right content at the right time without needing to, you know, put a human in the loop for every single step. What this all contributes to is that today delivering the right content at the right place at the exact right time is finally possible because we can now fully orchestrate and automate everything from needs analysis to understanding, you know, what what the data is telling us to proactively creating content and delivering it with agents doing nearly every step of the work.

15:37 And so that that's a perfect segue into into what Aerys is announcing today. Our point of view is that true agents should match job descriptions and we've built our agent ecosystem such that every single agent perfectly matches an existing job description in an enablement team or in a learning and development team. So needs analysis does the work that a typical needs analysis associate or business analyst will do, right? It will pull internal data and relevant documents, interview stakeholders in natural language and tell you exactly where the gaps are.

16:12 creator creates content based on all the needs, analysis, recommendations, based on files and research. And then from there, we have a referencing agent that verifies all of this content to make sure that it's accurate. We then route it to the right people based on natural language requests. So we can say, hey, we want to route this course to all of our managers in North America that have been here for two to three years. And then we can do AI analytics to look at all of the unstructured data in a course and figure out whether or not we're actually improving the original business outcome that we were trying to solve, right? Um, none of this is possible without a really robust infrastructure layer, right? I think this is where a lot of organizations are going wrong.

16:48 You cannot build really robust AI agents without having an incredibly robust infrastructure layer that handles all of the nitty-gritty, right? All the nitty-gritty of how do you build courses, right? How do you schedule them? How do you route them? How do you deliver them to the right people? How do you manage all the compliance for both frontline teams and deskspace teams? And so essentially, you know, our core platform is an infrastructure layer on top of which agents can easily execute uh jobs and deliverables, right? Um and our point of view is that the agents that win are going to be the ones that can do an entire job description end to end. Let me show you what this looks like in practice, right? This is what the enablement process looks like today for most organizations. and and I would uh wager that that pretty much every single enablement team and uh learning and development team that's on this call has a process that looks really similar to this right so first to figure out what the needs are in an organization most teams today will either do surveys that have relatively low adoption and completion rates because it's no nobody likes doing a survey right they take a while and pull people out of the flow of work or assessments many organizations will do in-depth assessments either you know uh for their technology teams or for their managers. These usually are also quite lengthy. Or many organizations will do one-on-one interviews with subject matter experts or with stakeholders. And typically to build an enablement program, most of our clients have historically done 10 to 50 interviews, right? So 10 to 50 hours of interviews for one course. Um the really critical consideration here is that oftent times the interview component and the assessment component is outsourced to a consultant or a thirdparty agency.

18:27 So this adds a huge amount of costs and you know many surveys like employee engagement surveys only happen once a year right and so what happens is that by the time that we do the employee engagement survey and get the data oftent times those challenges are no longer relevant right today especially in an AI first world the pace of business is too fast for uh surveys to happen once a year or even once a quarter right and so what happens is that today the L & D team or the sales team or a consultant that they hire gets all of this data right from surveys, from assessments, from interviews.

19:00 Occasionally they'll have a business analyst look at their HR or CRM or talent data and then they'll take some time to identify what the gaps are and answer the question, what do we actually need to trade on from there what they will usually do is either curate a course from a thirdparty content library and keep in mind that these libraries are not updated super often, right? A typical course in a LinkedIn learning or in a skilloft is updated maybe once a year, right? And so they oftentimes will have to sort through tens of thousands of courses with unclear relevancy and unclear accuracy, right? And curate those and none of those courses are ever perfect for the business need that they've identified. Or they'll have to build a course in articulate or PowerPoint and then build images and videos either in house or through a third party. Um, and so this first part, the needs analysis component today will take organizations anywhere from two to six months. This second part, the part the process of creating content, creating content, you know, building media files, that part alone takes another two to six months. And then the translation and personalization takes another two to six months because we have to manually translate and manually verify and manually personalize every single experience and we're building really long form complex experiences, right? So the the the challenge for L & D teams and enablement teams today is that it takes us sometimes up to a year to go from first identifying a need to building a course and delivering it. And the challenge is that when we deliver the course, we deliver it via a learning management system, a CMS or a live session. And there's a few challenges with these mediums. One is that people don't go to them, right? The reality is that people go to the LMS once every six weeks. And most clients of ours have historically had learning operations teams whose sole purpose is to just get to people to attend. Right? Then the other challenge is that if somebody goes to a live session, it's not personalized to them and they forget 90% of what they learned within 30 days. Right? So we create oftentimes really incredible content and then we put it in a box far away from where people spend their time for nobody to see. Right? And this is the heartburn that so many teams feel, right? They're spending so much time and effort creating really really great content and trying to understand what the needs of the business are only to deliver courses in a way where nobody will engage with them. Right? And so the challenge is not only do we have a problem in terms of how we understand the gaps and how we create content, but we also have a problem in terms of how we're delivering content because we're not meeting people where they are. With Aerys agents, this process changes entirely. Right? Needs analysis is connected to your HRS and CRM and talent data and is reviewing that in real time to figure out what the gaps are. And then from there you can in natural language share business problems with needs analysis and needs analysis will go out and do you know personalized interviews with thousands of people.

21:53 From there once needs analysis has identified what the gaps are. It can pass off to creator to build and translate and personalize the right courses or communications. um we can curate videos and podcasts from existing AI tools like Cynthia or notebookm and then we can deliver all of the content directly in the flow of work ensure that people see it and then our analytics agent will actually tie back all of the content to the initial needs analysis gap identified right so a process that used to take 12 months or you know at best two to three months all of a sudden start to finish can happen in under 72 hours we've had clients go from having a need in the business to you know interviewing a bunch of people within 48 hours to building a course immediately after personalizing it, verifying it and then delivering it to people in one click right in the flow of work. Right?

22:40 This proc this is probably the most important slide in this presentation. This is the big process shift that's happening and every enablement team needs to adjust to this process. I think uh one of the most important considerations here is that the tool stack that people have today is not designed to support this process, right? Um as you'll notice, all of this happens in one agent. It can happen directly in Aerys. Um and we can work with other AI tools to to you know fill in the gaps as needed for more in-depth content or for for you know specific uh specialty areas like coaching for example because you can embed coaching experiences from tools like you lead directly in ARIS. Um but the really core consideration here is that legacy enablement tools will become largely obsolete right again as I mentioned today a lot of needs analysis happens in tools like viva glint or perceptics or worker um and those tools will largely become irrelevant as we can do continuous listening fully in the flow of work and uh we no longer have to do skill analysis because we can get a really granular understanding of where somebody's gaps are right um what we can also do uh what we can also do is instead of having to create content in articulate or pay a thirdparty agency or curate content. Aerys can create the perfect course for you every single time. There is never a need to buy packaged learning content again. We can just build the perfect course for you on the fly at the right moment of need.

24:05 Right from there, tools like seismic and skilloft and degree also become dramatically less relevant. Right? Because uh you know the whole purpose of an LXP was to have a a more engaging way for people to engage with learning. But we've found consistently that if you just deliver learning where people spend their time, there's no real need to send people to an LXP, right? Between an LLM and AIS, these tools will largely see usage rates fall. And we've already seen this happen with many of our clients, right? And then when it comes to analytics, right, today so many organizations rely on the LMS for analytics data. The challenge is that the LMS is really designed for compliance tracking. It's not designed for tracking performance. And as I mentioned earlier, one of the reasons why so many organizations feel a level of heartburn when they have to use their existing tools is because they have to use tools that were designed for compliance, but their remmit and what they're being asked to achieve is performance, right? And so uh we see you know Aerys will today will happily send data back to workday or success factors or cornerstone but we see these tools becoming obsolete as an analytics layer over time because the best data that we can capture to prove performance is actually directly an erys right when we can do a voice conversation with an individual after a course to verify their skill or if we can look at internal business data like HRS data or CRM data to prove that somebody has actually gotten better at what we wanted them to get better Right. The the the most important piece of this from there is that today for a 10,000 person organization, an average company will spend 2 to four million a year on this legacy tool stack, right? And it creates a ton of overhead because you have to manage tons and tons of tools and stitch them together with an agent first approach. Real time AI data analysis and AI voice and messaging interviews are placed and need to do needs anal needs analysis and gap assessment. Right?

25:56 Courses are created automatically. you deliver everything directly in messaging tools and then you have AI analytics on the back end. One uh sort of consideration from there is that today many organizations that we chat with are like yeah we use AI we're using AI to build courses and articulate faster or we're using AI to build powerpoints faster. What you'll find is that if you just do one part of the process here, right? If you just create courses faster like this step for example, you're not actually getting the full benefit of AI, right? You are doing uh you know you'll reduce your timelines from eight months to seven months, maybe six months, right? But the real benefit of AI and the real benefit of agents happens when you revamp your entire enablement process end to end. Otherwise, you will still be bottlenecked by other parts of the process.

26:44 And the best part is that if you take an agent first approach, it's not only way simpler because you're managing less tools, but it's dramatically dramatically cheaper. We see these budget categories going completely extinct over the the next uh five years and uh you know the public market and private market valuations of many of these companies and many of these categories have already plummeted 50 to 80%. Right? Um, from our point of view, uh, learning design agencies will have to massively shift, right? Um, and will, uh, will either become fully outsourced L & D teams or will, you know, will largely stop creating content. Um, learning design authoring tools will also largely become obsolete as agents become much much better at creating content. And already, you know, most of our clients have stopped or are in the process of stopping using a lot of their thirdparty tools. content libraries will largely become obsolete because nobody needs a static course when you can create the perfect experience at right time. Um, one of the things that's important to mention here is that you know tools like uh Claude and OpenAI um are largely trained on a lot of existing content library data, right? So there are very very few truly proprietary data sources where you need to have an expert platform versus just building a course using an LLM. What we found is that LLMs today are usually more accurate than expert-built courses uh in a third-party tool because um you know unlike a content library, an LLM does real time research to figure out what the most up-to-date data is um and hallucinates a lot less. Um we also see you know learning experience platforms uh you know over time diminishing in need um as people go directly into the flow of work or or go directly um to an LLM. Um the same is true for employee listening platforms right we believe that employee listening should happen in the flow of work um and fundamentally you know the the as you all will see in a second the engagement rates of the flow of work are so much higher than doing a manual survey right and the data that you get the quality of data that you get is dramatically better as well and the same is true for assessment platforms um and and then you know hiring consultants for performance gap analysis will still happen to a certain degree uh but the reality is that most of this work will be done by by agents So with that in mind, let's talk about what this looks like in practice. Um, and for the next 20 minutes or so, we're going to walk you through how the Aerys agent ecosystem works start to finish.

29:02 So let's start at the beginning with needs analysis, right? Um, needs analysis, uh, as I've mentioned lets you uh, rapidly do thousands of personalized interviews. Um so instead of manually doing you know a ton of interviews uh uh AIS needs analysis will manually reach out to every single individual schedule time on their calendar um or call them directly and we'll have a personalized voice conversation. Um so in this case in this example um we uh for an internal use case ask needs analysis to interview a lot of our reps and a lot of our executives to figure out what the gaps are. And this is a great example of what a needs analysis uh conversation looks like. So, let let me let we'll show you this in just a moment here. Give me just a second.

29:52 Awesome. We're just going to ensure that that you all can hear this audio and pinpoint how to >> So, please uh please let me know in the chat if you can't hear this audio. Hi Ryan, thanks for taking the time to speak with me today. I'm really looking forward to our conversation and learning more about your experiences and perspectives, especially in the context of your role as a CRO. I appreciate your willingness to share your insights with me.

30:24 >> Of course. >> Are you ready? >> Thanks so much. >> I'm glad we could get started. Can you tell me a little bit about what you're hoping to where all of our sales reps gaps are and paint a really clear picture of what we should do moving forward based on what sorts of challenges you uncover within the company. >> So, you're looking to identify gaps in your sales team's performance and use our conversation to inform your strategy moving forward. That makes sense. Can you tell me more about what you've observed so far in terms of challenges or areas where you think the sales team might be struggling? Yeah, I think we have really skilled reps, but the problem with having cool technology that you can use for a lot of different things is that a lot of our buyers will often try to use it for too many things at once. We want our reps to be really targeted and prescriptive about exactly what the best use cases for AIST are, where to point them, and make sure that they align with the most strategic buyers in order to get the most impact at the company as fast as possible.

31:30 >> Thanks for sharing that. It sounds like you're looking to help your sales. >> So, for the sake of time, we'll send over this video after our chat and everybody will get a recording of this session. Um, but what you'll notice here is that the the conversation is fully personalized to Ryan and all the context we have about Ryan's role as CRO at Aerys, right? This conversation goes on for seven minutes. Um, and what we found is that humans are so much more likely to engage with a voice agent and so much more likely to go in depth and really talk for a very long time even compared to a human for a few reasons, right? AI agents are better at listening than humans are, right? Uh they're really, really good at listening, at diving deep, at asking questions in a very, very thoughtful way. And they have far more context on Ryan and all of the internal data about Ryan than a typical human would. Right? So, we can pair structured data points about Ryan with this unstructured interview. Um, and this entire interview right here that you all just saw was set up in about a minute or so. We just told Ne's analysis, hey, here's what the gaps are.

32:29 Um and then needs analysis set up interviews for Ryan for the rest of our team and then went from there. Um so the the setup is pretty much immediate. Um and then the routing, you know, happened immediately. Uh needs analysis reached out to Ryan via email, scheduled some time for a call. Uh and in this case, Ryan asked for news analysis to call him directly. Just keep in mind for a second one thing that I'll mention here, every single company and every single executive every at every company has tons of use cases for understanding their employees more, right? Um, and one of the things that we encourage you all to do is go to, you know, uh, one of your peers or one of an executive that you work with and ask them if we could interview every single manager or if we could interview every single leader at the company or if we could interview every single frontline associate or sales rep, you know, what would we want to ask? Uh, and the reality is that there are uh, you know, massive needs for atscale interviews uh, for every organization. It just hasn't been possible until now. So we can take all this interview data and again from a one-s sentence goal in this case we literally just typed in hey can you tell us uh can you can you uh help us understand what the gaps are in ARIS sales team and ARIS sales process um AIS identified what the relevantmemes and stakeholders were reshot them for interviews and then we'll aggregate and analyze all the internal data points associated with those individuals and then you know it'll build a consulting output and there's human review throughout right the needs analysis showed us the questions showed us who uh it was going to reach out to for for approval. Um, and so we approved it, it reached out, and then we got a detailed consulting report. So, let me show you all what the consulting output looks like. And again, this was put together immediately after. We're going to show you all an example um from one of our clients uh that we've anonymized and redacted. Um, and this is a good example of measuring what high performing sales reps are doing and trying to understand how we can get the rest of our of the sales team to that baseline. Right? So in this case we interviewed you know over 60 managers and top performing sales reps and low performing sales reps and then needs analysis essentially designed a consulting report helping us understand uh what can we do to help all of our B and C players become A players.

34:39 So what you'll find is and again this was full this is fully AI generated and was fully generated by needs analysis um and based on all of the needs analysis conversations uh needs analysis gives us a really detailed executive summary right so we can quickly dive in and get a clear sense of here are the core skills that define top performers right um we talk about sort of the skill frequency based on managers and reps and we uh needs analysis gives you data to site every single claim right so in this case we actually pull pull out here are the customer oh sorry in this case, here are the sales rep quotes um that drove our understanding of this and here are the data points that drove our understanding of this in this conclusion. Right? So, we dive into sort of every single recommendation, talk through the data points that got us there and then um from there needs analysis will make a recommendation for the core work streams that will drive a successful enablement program here.

35:29 Right? Keep in mind that uh ARIS needs analysis doesn't just recommend an ARIS course, right? It will recommend holistic recommendations uh and holistic interventions including live discussions right uh you know adjustments to org design etc. Um and then for anything that it can create an era so we'll automatically kick off the creation process for right um and then from there it recommends a 90-day action plan to codify pilot and scale um and then we can just sort of execute from there and it also recommends uh metrics that we should track to prove that the intervention is working right so what we'll do later is once we launch this program uh we can then go back and ensure that we're tracking to all of these metrics and ARIS analytics will look at behavioral metrics and commercial metrics to prove impact back to needs analysis, right? Um, so again, what's happened so far, we've done tons of interviews that were all personalized. We have a really clear understanding of what the gaps are, and then we have a really detailed consulting output with recommendations for who to enable, how to enable them, what to enable them on, and what we should track to prove the impact of that enablement. And this has all happened for free, right?

36:40 Um just a few quick data points before we move forward. What we found so far in early uh in early piloting and we're incredibly grateful to Metronic and Coat Group for being uh you know two of our early participants here at Metronic. They saw 3x more participation um in needs analysis interviews compared to prior listening experiences whether that be surveys or or live interviews. So people were incredibly engaged and the length of those conversations was also dramatically more in-depth than previous surveys that they had done. At Codesc Group we found a 5x greater preference for listening in the flow of work in this case via messaging. Um and we found that people largely prefer being reached out to in the flow of Teams or text or Slack or voice versus having to click on an email and go to a thirdparty platform to fill out a survey or even do an interview with a human. Right? So there's uh hu like what we found is that employees dramatically prefer this approach.

37:33 Once we've created uh sort of that needs analysis recommendation, we can immediately pass off to creator to build courses. So let me show you what what what we've built and we're going to go back to our internal aerys use case. You know, one of the things that we identified in our internal aerys deployment was that we needed a lot more help for our reps explaining multi- aent systems to clients and handling some of the concerns around multi-agent systems.

37:56 Right? So we're going to go ahead and do is we're going to go ahead and build a course and I'll show you the behind the scenes of uh you know how the AI creation tool works um and what needs analysis can sort of or over time orchestrate and execute automatically. So so let's go ahead and dive uh dive into this. So, we're going to go into the Aerys platform here and this is this is sort of the Aerys creation platform and we're going to go ahead and build a course right um again I'm going to do this manually for the sake of this demo u but you know over time as organizations implement needs analysis this can be done automatically in this case we're going to go ahead and build us a course on explaining the key benefits and core concerns of a multi- aent infrastructure for sales enablement needs analysis and creation right we're going to go ahead and make this a skill building course or actually let's make this an awareness building course. Uh audience is going to be sales reps at Aerys.

38:48 Um let's say they're they're experts because they've been with us for for a while. And then REI is going to automatically recommend an action or behavior, right? So in this case, we want uh reps to be really good at handling objections. This was entirely AI recommended based on sort of all the context that we shared. Um in this case, uh we're going to pull together a few knowledge components. So the AI is recommending, you know, creators recommending a few knowledge components.

39:13 We want uh folks to really lean into concrete examples and case studies and then we're going to make this a seven lesson course and we're going to pull in a few resources here. So we're going to go ahead and pull in uh this really in-depth paper on multi- aent risks um that you know a few clients have shared as an example of sort of what they're what they're looking to address in their IT processes. Um, and we're going to pull in uh Anthropic's example of how they built their multi-Asian research system. So, we're going to pull in both of these very, very long docs, and we're going to pull in some of our needs analysis data points, right? So, let's go ahead and do that. We're going to go ahead and pull in that anthropic document. So, we're going to go ahead and uh let's do this.

39:55 We're going to go ahead and download this document, and we'll we'll upload it. And then we're going to go ahead and pull in that anthropic document. And then, we're going to uh paste this in. and we're going to go ahead and upload this 100page PDF. Um, at this point, we could ask the AI to just focus on the materials that we've uploaded. Um, this is really helpful for life sciences companies where we're uploading thousands of pages of documentation or we can ask the AI to do research online.

40:19 In this case, we're going to ask the AI to do research online as well. Um, we're going to go ahead and build a course outline. So, the I the AI will do some initial level of research um and build to get build a course outline that we can then review. So as you can see we have humans in you know uh as part of every step of the process here. Um and once this is done we can go ahead and generate course uh and generate the course immediately. Um let's go ahead and and uh you know I went ahead and built this course right beforehand just for the sake of time. So in this case we can go ahead and take a look at all the inputs right so here are all the inputs that we had. Here are the files that we uploaded. Here's the course outline that we built. It looks good. So I asked uh you know Aerys to build the course right away and Aerys builds everything. So we build all of the lessons, the whole structure, all of the interactive questions. We'll notice is that we have citations. So we li we line by line site every single claim that we make in the course back to the documents that we uploaded, right? Um so we're able to to have an incredible level of accuracy because we're actually citing every single claim. Uh we can see all the sources that we pulled and all the research that we did, right? So, so just to kind of give you all a frame of reference, start to finish in about five minutes, we asked an agent to build a course on a specific topic and creator went out and researched, you know, in this case, 40 different topics, did a ton of reading, read through thousands of pages of research, built a seven lesson course with images and infographics and interactive questions and lessons, uh, and then also cited every single claim to ensure accuracy.

41:53 And then what we can do from there is we can go ahead and upload our own images uh or photos or videos as needed. In this case, what I did is I briefly plugged in synthesia. Uh so I pulled in a synthesis video to kick off our experience here. Aerys then builds a really concise, really to the point lesson. Um and then we can include additional media files. So Aerys built us a really thoughtful infographic here that's fully to our brand guidelines. We have a series of multiple choice questions with instant feedback and evaluation. Um, from there, Aerys built um an open-ended AI evaluated scenario question. So, we can actually have people practice and then AI will evaluate whether or not they got the practice and scenario correct or not correct. Um, and then in this case, we're going to wrap up with a more in-depth seven minute explainer video from Notebook LM. So, we can pull in multimedia from other uh AI providers, right? and integrate that very very seamlessly and all of that will play directly in the Aeros course which we can deliver fully in the flow of work.

42:51 Right? So all of this was built start to finish in under a few minutes and what we can do from here is we can then translate this course into multiple different languages. So let's say we want to translate this to Arabic, Czech, uh and French Canadian, right? We can instantly translate this course or we can instantly create versions of this course for different audiences. So we can adjust the target audience for this course to you know let's say more junior reps, more senior reps, reps that are coming from different uh sort of companies or reps that are focusing on different industries. And then uh all of this is created automatically. Um and so in this case we have you know perfect translations for different languages.

43:29 What you'll notice is that not only do we translate all of the content but we also translate all the images and infographics as well, right? Um and then for the versions we have you know we've created different versions for CSMS. Um and for for more junior sales reps and we require human review for all of these to ensure accuracy as well, right? Um the I see some questions about Viva. Um the really really incredible thing here is that we can export everything uh as a word doc and we uh all of our exports are designed for medical and legal review. So everything can be automatic including all the citations can be automatically exported and submitted to Viva or any other compliance and verification tool. Um and courses uh courses are you know can be fully personalized as well based on role or based on geography.

44:16 So that's how quickly you can create content right it's incredibly incredibly fast. Um so let's go back to to sort of walking through uh the process here. Um essentially what we did is we just to kind of recap where we're at. We figured out what the needs were. We built courses. We translated them into multiple languages personalized them for different roles and then we can deliver everything from there. This is how Aerys thinks about uh sort of personalization and creation at scale, right? The reality is that we can't have full personalization in the way that Chad GPT does because every single enterprise has incredibly strict requirements for verifying accuracy of content. So our approach is that we start with the needs analysis recommendation. We then pull in internal data sources, research from creator, uh, and then interviews that needs analysis can do on your behalf and then we route all of that to creator to build one core course. Once that core course is reviewed by an AI and verified and once we have a layer of human review and verification and feedback, then we can instantly create personalizations based on different roles. And you can actually upload job descriptions into Aerys and Aerys will personalize content for different roles. From there, once we've created content for different roles, we can personalize for different levels within that role and then instantly translate uh each of those levels into, you know, dozens of different languages, right? And so we're able to get a full level of personalization based on geography, language level, role um while still maintaining all the compliance requirements that enterprises care so deeply about and that are so critical for ensuring accuracy.

45:49 Um what we found so far is that the the data from creator has been unbelievable. At Walter's clue they were able to launch a global AI upskilling program in 13 languages uh to you know uh hundreds of different roles in under three weeks. Two weeks of which was internal review time right um so they were able to launch this about eight times faster than their traditional process. uh and they were able to get you know 30,000 people trained and saw 120% lift in AI adoption just three weeks after that right so start to finish in six weeks they were able to build personalized training for 30,000 people drive massive AI adoption lift and save 1.62 62 million in terms of time saved and creation time as well. Right? So massive savings by using creator for creation and then for delivery.

46:36 Once we've created all the content, we can route it to the right people. So um we AIS has smart cohorts that can instantly route C, you know, content based on natural language needs and we can integrate with your CRM or your HRS to route the right content to the right people based on need, seniority, um you know, organizational positioning, etc. We can also trigger content based on critical moments of need. So if somebody was just promoted to be a manager, we can instantly enroll them in a course in that exact moment. Or if somebody is struggling with a discovery, we can enroll them in a discovery refresher in that exact second. Um, and then all of the operations are automated. You never have to follow up with anybody again.

47:12 All the follow-ups and check-ins and refreshers automatically are done by Aerys. The really cool thing is that uh, you know, directly in AIS courses are incredibly engaging, right? So if we go into an example here um you know this is my course on multi- aent infrastructure I can you know it happens you know directly in teams or directly in slack or by by text message and I can just watch you know this video >> sales enablement has a problem most content is created with >> I can read my lesson type back my answer to my question uh and then as you can see I'm getting automated uh sort of follow-ups and check-ins here and then I move on to the next lesson or the next component of my lesson from there right so AIS makes it incredible easy for me to engage directly in the flow of work.

47:56 And as you can see, I'm engaging here directly in chat. Uh so I don't have to go outside of Teams or outside of Slack or text message to get a really rich multimedia experience um and get, you know, all of my content and feedback, right? So nobody ever has to go to an e-learning platform ever again, right? And in this case, I can type in a natural language uh you know sort of responses uh and then I'll get instant feedback and guidance based on what I know and what I don't know as well.

48:24 Right? So in this case I get some constructive feedback with some additional content to help me along my journey and then I can watch this seven minute video about agentic AI from notebook alone. Right? But again everything happens fully in the flow of work. So our point of view here is that the era of going to an LMS or an LXP or a CMS is over, right? Uh internal LLMs and platforms like YouTube have solved the pull problem, right? And the problem of people pulling trading uh and predictive and proactive and embedded agents like ARIS are going to solve the push problem, right? And already are solving the problem of predicting gaps and pushing the right content at the right place at the right time. So uh from our point of view you know that the era of having to go somewhere to learn is over because really really rich in-depth learning experiences can come to you and they can be fully personalized. Um and what we found is that meeting people where they are enables some pretty in incredible results. Right? Just by switching the delivery method to uh uh in this case SMS for Novartis, they saw uh 70,000 enrollments across 140 programs very very rapidly. Um and uh again an unbelievable amount of scale uh in the first six months of a deployment.

49:32 Um and 80% of reps said that they saw immediate application of the training. Right? because when you deliver contextual learning where people are, people can actually use that content immediately and apply it very very rapidly. And then the last component that I'll cover here is our analytics station, right? Um, one of the really neat things is that as you all saw in Aerys because learning is conversational, uh, we track dramatically more data points than any other learning platform on the planet.

50:00 Uh what that means is that we can get a really clear sense of where people understand and don't understand and have a very clear sense of what the knowledge gaps are, what the skill gaps are on a granular basis, right? We can zoom in and say, "Hey, it seems like people don't understand this specific portion of the SPI feedback framework." Or, "Hey, these people don't understand this specific portion of our product." At one of our clients, a large public software company, they were able to train uh customer support reps on pricing and figure out exactly where the pricing gaps were ahead of a major launch uh and ensure that uh customer support reps were properly prepared to talk about pricing in the right way. And you can chat with our analytics agent, which we'll be releasing later this year, to create custom reports and answer questions immediately. So, you never have to manually go through data to figure out something for for uh you know, a leader or for a question that somebody has. you can just ask the AI agent and our analytics agent will build content for you or and we'll build sort of the right uh the right report immediately.

50:56 Last thing that I'll mention here is that we always tie engagement data back to the original goals, right? So, we're tying everything back to data in your CRM data in your HRS to ultimately prove the impact of the intervention. Um, and what we found is that, you know, when we're able to connect Aerys courses and Aerys interventions to business data, we see some really remarkable business impact, right? at organizations similar to EcoAB or Exon Mobile or RVs, we've seen not only 2x year-over-year revenue growth for com uh for folks trained on Aerys versus folks trained in traditional methods, but we see about a 3x increase in methodology adherence and 20% more meetings booked just by delivering personalized learning in the flow of work and tying it back to business outcomes.

51:38 So, from our point of view, the era of agents is here. Uh, and we're incredibly excited that AIS is the first end-to-end agent for enablement teams. Um, and we believe that this is a watershed moment for learning and development enablement teams. And it's easier than ever to start. Um, the reality is that enablement is a flywheel. It's not a oneanddone process. And all of these agents when they work in concert can create a really robust flywheel for understanding what the gaps are in creating the right interventions and helping your organization perform dramatically better. The neat thing as well is that these agents are actually much easier to implement than traditional software platforms, right?

52:15 Um, largely because they massively simplify how most organizations operate. And so our recommendation here is to map your before and after workflows, right? Map what uh your current enablement workflow looks like from needs analysis to creation to delivery. And we're super happy to help you map what the after looks like with agents at the center. And what you'll find is that not only will your workflows dramatically decrease in complexity, but you'll also save a lot of money in the process, save a lot of time in the process, and you'll free up your team to spend more time making people exceptional. And again, you don't need a strong data foundation to start because agents can do a really good job of pulling firstparty data through interviews or through the flow of work. You just need a critical or complex use case that you want to deliver quicker and that you want to deliver with better accuracy and efficacy.

53:03 Um and what we find is that the best organizations usually start with one of these three use cases, right? They start with major product launches, uh like the GLP1 rollout that's currently happening. Um you know, scaled gap analysis or interviewing to to very rapidly get a sense of where reps or managers are are are doing and how they're performing or organizationwide AI upscaling, right? These are the scaled use cases that organizations tend to find the greatest level of success in.

53:27 So, thank you all so much. Uh we are unbelievably excited for the next generation of talent and the next generation of enablement. Um and so so excited to share with all of you an enablement team in your pocket. Um we would love to spend more time with HTV1. So please go to aristo.co or a.comdemo for a demo. Um we have incredible reps that are all trained using Aerys and they would love to help you. Um and we're just so excited for the journey ahead as well. Thank you all so much.