Summary:
In this episode of the Sidecar Sync, Mallory Mejias sits down with Jason Marchese, Learning Content Specialist at Sidecar, to explore the exciting frontier of AI-generated interactive web apps. From solving everyday meal planning problems to analyzing U.S. migration trends, they demonstrate how anyone—even non-technical folks—can build useful tools using Claude and ChatGPT. The conversation then shifts to a groundbreaking internal project: the Learning Content Agent (LCA), a dynamic AI-powered system that streamlines and customizes educational content at scale. Whether you're curious about AI's impact on personal productivity or ready to revolutionize your association’s learning programs, this episode is packed with inspiration and practical insights.
Timestamps:
00:00 - Introduction
🎉 More from Today’s Sponsor:
Member Junction https://memberjunction.com/
📅 Find out more digitalNow 2025 and register now:
https://digitalnow.sidecar.ai/
🤖 Join the AI Mastermind:
https://sidecar.ai/association-ai-mas...
🔎 Check out Sidecar's AI Learning Hub and get your Association AI Professional (AAiP) certification:
📕 Download ‘Ascend 2nd Edition: Unlocking the Power of AI for Associations’ for FREE
🛠 AI Tools and Resources Mentioned in This Episode:
Claude Dinner App ➡ https://shorturl.at/BQzBq
ChatGPT Dinner App ➡ https://shorturl.at/2ccPX
Census App ➡ https://shorturl.at/ehNKF
ChatGPT ➡ https://chat.openai.com
Gamma ➡ https://gamma.app
ElevenLabs ➡ https://elevenlabs.io
Claude Code Overview ➡ https://shorturl.at/EKlHa
https://www.linkedin.com/company/sidecar-global
https://twitter.com/sidecarglobal
https://www.youtube.com/@SidecarSync
⚙️ Other Resources from Sidecar:
More about Your Hosts:
Amith Nagarajan is the Chairman of Blue Cypress 🔗 https://BlueCypress.io, a family of purpose-driven companies and proud practitioners of Conscious Capitalism. The Blue Cypress companies focus on helping associations, non-profits, and other purpose-driven organizations achieve long-term success. Amith is also an active early-stage investor in B2B SaaS companies. He’s had the good fortune of nearly three decades of success as an entrepreneur and enjoys helping others in their journey.
📣 Follow Amith on LinkedIn:
https://linkedin.com/amithnagarajan
Mallory Mejias is passionate about creating opportunities for association professionals to learn, grow, and better serve their members using artificial intelligence. She enjoys blending creativity and innovation to produce fresh, meaningful content for the association space.
📣 Follow Mallory on Linkedin:
https://linkedin.com/mallorymejias
🤖 Please note this transcript was generated using (you guessed it) AI, so please excuse any errors 🤖
[00:00:00] Mallory: We're talking about AI here, and we're talking about, you know, coding and interactive web apps, but I think we're really talking about something much greater, which is helping humans learn and retain content in ways that they never have before.
[00:00:14] Intro: Welcome to Sidecar Sync Your Weekly Dose of Innovation. If you're looking for the latest news, insights, and developments in the association world, especially those driven by artificial intelligence, you're in the right place.
[00:00:27] We cut through the noise to bring you the most relevant updates with a keen focus on how AI and other emerging technologies are shaping the future. No fluff, just facts and informed discussions. I'm Amith Nagarajan, Chairman of Blue Cypress. And I'm your host.
[00:00:43] Mallory: Hello everyone and welcome to the Sidecar Sync Podcast, your home for all things associations, innovation and artificial intelligence.
[00:00:52] Today, we've got a special edition episode lined up for you. We're doing an evergreen topic or as evergreen as you can be in the world of AI. We'll be talking about building interactive web apps using primarily Claude, but some other tools as well. My name is Mallory Mejias. I'm one of your co-hosts along with Amith Nagarajan.
[00:01:11] And I am joined today by Jason Marchese. How are you doing today, Jason?
[00:01:19] Jason: I'm doing great, Mallory. Excited to be here. Honestly.
[00:01:21] Mallory: I'm excited to have you here. So you just joined the Sidecar team recently ish, right? A few months ago.
[00:01:28] Jason: Very recently. About a month and a half.
[00:01:31] Mallory: Oh, wow.
[00:01:31] Jason: Rearing towards two months.
[00:01:32] Yeah.
[00:01:33] Mallory: And what is your role at Sidecar?
[00:01:36] Jason: Uh, my role is learning content specialists. So essentially I'm going through, reviewing all of our content, optimizing our content, making sure it's all up to date, and then also customizing the content for, um, new domains and new specialties.
[00:01:51] Mallory: That is awesome. You've told me a little bit before we started recording about what you've been working on and I cannot wait to cover that in this episode.
[00:01:59] Can you tell us all a little bit about your background, though? I know that you have some association experience.
[00:02:07] Jason: Sure. Um, so I'm just coming from Morocco. I was there working with an association for about nine years, almost 10 years. Wow. Um, but I've been in education now for 15 years.
[00:02:19] Mallory: Okay. So that there's so much to unpack there.
[00:02:21] So you were in Morocco for how many years?
[00:02:25] Jason: All together? About 15, but oh my, in this last position, nine.
[00:02:29] Mallory: Yeah. Okay. In Morocco for a total of 15 years at this association for nine years. And then you said your background is primarily in education. Is that right?
[00:02:40] Jason: Um, so I got my start, um, in Peace Corps actually.
[00:02:44] Um, and then I just kind of stayed in education and training. Um, I enjoyed it. It's more of a passion for me.
[00:02:51] Mallory: Awesome. Well, I think that'll be really relevant experience for all of our listeners on the Sidecar Sync. If you all have listened to any episode prior, you've probably heard me say I'm the non-technical person on the podcast.
[00:03:04] AME is the technical person on the podcast. Even though I will say through co-hosting this podcast, maybe I've become a bit more technical. So, Jason, just to set the stage for this episode. How would you define your technical expertise? I know when I'm talking about interactive web apps, I'm coming from the non-technical angle, so I want you to share a little bit about your expertise.
[00:03:27] Jason: I would say I'm pretty non-technical. Okay. Um, I have a joke with one of our engineers saying that. I took one C plus coding class back in college and it was the last course I would ever take. It was not for me, for sure.
[00:03:44] Mallory: Well, that's one more coding class than I have taken. Actually, I will say I did try to do some self-taught, like HTML coding, you know, back in the, the MySpace background days, if you remember those.
[00:03:55] So I guess I've dabbled, but you've definitely, I think you're more technical than I'm.
[00:04:02] Jason: I don't know. I don't remember much from that course. I, you know, what is that? That was more than 12, 13 years ago at this point. Um, but. In this day and age, you really don't have to be technical, which is a huge advantage.
[00:04:15] Mallory: That was my point, exactly. When we start talking about what we've built, what Jason has been working on, as someone who considers himself non-technical, I think you all will be quite impressed. So as I mentioned, we're talking about interactive web apps today. First I wanna do a little bit of foundation setting.
[00:04:33] Um, show some examples that are more fun, easygoing, lighthearted, uh, just to establish what we mean when we say. Interactive web apps. So we'll look at two different examples. One is a fun one that for me kind of helped me solve a household problem that I often have. And then I'll have Jason share one that still was quite easy, I think was built in a, a single prompt with a few edits, but is more on the like data number visualization side.
[00:05:00] So for all of you that are just listening in audio only, I'm going to do my best to talk you through. What I'm sharing on my screen. So right now I am inside Claude. I mentioned a household issue and for me, I'm the person in my household that cooks. I really like to cook. I don't like having to decide what to cook, especially when I've got all these kind of random assortment of groceries, uh, in the fridge and in the pantry.
[00:05:31] So for me, often with ai since. I talk about it in such a business context all the time. I actually have to break my brain a little bit to apply it to my personal life because most of the use cases I'm thinking about are association specific or sidecar specific. So I thought I would challenge myself and build a web app.
[00:05:50] That applies to my personal life. So I gave it a really simple prompt. I said, build me an interactive web app that allows me to choose common groceries I probably already have in my fridge, and then generates healthy recipes based on what I have on hand. So easy enough, um, that was the only prompt I gave it.
[00:06:10] I'm going to show you the little web app game, if you will, that it created, it gave it a title. It's called What's in My Fridge, select Ingredients You Have and Discover Delicious, healthy Recipes. It broke down all of the ingredients into different categories, which I didn't tell it to do. So I've got proteins, vegetables, fruits, dairy, and eggs.
[00:06:30] Pantry staples, and basically I have a whole bunch of check boxes in front of me, so I'm gonna check a few of these things that I often have on hand. Salmon, eggs, black beans, typically got yogurt. I'm gonna click some vegetables on here as well. Some fruits and. Cheese, butter milk, things we most often have.
[00:06:55] And then I'm gonna say I have rice and pasta, so I selected 15 ingredients. I'm gonna click generate recipes. What I really like about this, again, I didn't provide it any instruction to do this, is that it actually gives me a percent match on recipes so I don't have to hit the recipe a hundred percent, but, uh, one of the recommended ones it gave me was black bean and avocado salad.
[00:07:19] Mm. I will say I. Didn't click avocado. So I do think this recipe's not an ideal match. It's only 67%. Uh, scrolling down, protein packed spinach scramble. I've got the eggs, cheese, and garlic. I don't have the spinach. Again, I built this little lap in one prompt. Um, but it did. Really good job. Now, I sought out to create the same exact app on chat, GPT, just to see.
[00:07:48] I provided it the same prompt, and what I got was a bit more lackluster compared to what I created in Claude. I've got a really plain, simple screen here. It says, what's in your fridge? It only gave me maybe 15 ingredients to choose from, and then what happens when I click a few? Uh, it actually won't find me any recipes, so I have to make sure that I have clicked the exact right ingredients.
[00:08:18] Or it won't generate recipes. So now I've clicked almost all of them and it gave me a veggie omelet with some really simple instructions on the Claude app. Actually got detailed kind of step by step recipes on how to cook. And I also have a tomato pasta that literally says, boil pasta pepper, prepare sauce with veggies mixed together.
[00:08:38] That's the whole recipe. So these are my two kind of fun examples to show how. Quite literally anyone right now listening to this podcast could go to Claude or to chat t and build yourself an interactive web app. You can take this thing and integrate it with all your association data sources, right? And make it fully functional without any technical expertise.
[00:09:02] But if you just wanna play around with this tool, you can. And now Jason, I wanna kick it to you so you can share screen and demo perhaps a slightly more serious version of an interactive web app.
[00:09:16] Jason: Absolutely. But how useful is your version?
[00:09:19] Mallory: Thank you. I mean, it's pretty useful and if I had given it more instruction, more ingredients that I actually like, recipes that I like, I think I could have a fully functional little web app that I use every day.
[00:09:31] Jason: Absolutely. Already a big time saver. Exactly. I mean, how much time do we spend researching that online, looking up different recipes and it doesn't match what's in our fridge? So true.
[00:09:40] Mallory: So true.
[00:09:41] Jason: So I think some of you can guess what's up here now. So looking at a bit more data focus. So this one in particular looks at US population migration.
[00:09:53] So Mike Mallory was saying some of this can be built by just one simple sentence, prompt. This one in particular was built by Build me an interactive web app that uses publicly available census data showing population migrations in the US over long periods of time. Now, we did have to do one minor fix to this, but it gave us this incredible interactive chart.
[00:10:20] Where we can look at all of this sense of data, specifically state by state if we wanted to, and we can see the increases decreases to get a sense of that flow of migration. State by state.
[00:10:34] Mallory: What state are you in right now, Jason? I.
[00:10:37] Jason: I am currently in Louisiana. Oh, great. So, so it looks like people are leaving us.
[00:10:42] Mallory: Oh yeah. Looks like Louisiana's got like a consistent outflow. You're
[00:10:46] Jason: not doing so well. Not doing too hot.
[00:10:47] Mallory: Can you click on Georgia? That's where I am. Absolutely. Georgia's growing. So we've seen kind of some consistent inflow, the top inflow coming from Florida. So I would say that's not super surprising.
[00:11:01] Um, but yeah, I think this is a great example of a more serious thing that you could do with Claude's interactive web apps. Jason, I'm curious, do you have a favorite tool? I mean, people on this podcast, if you've listened, you know that I favor Claude. Claude was also building the Better Interactive app.
[00:11:20] Claude's also better at, uh, writing, in my opinion, in my humble opinion. So Jason, I'm curious if you have a favorite tool.
[00:11:27] Jason: I have to say a few months ago it would've been chat, GPT. Mm-hmm. Hands down. Um, I was obsessed with it, loved how it, um, basically responded to me and I liked it. Sociability basically.
[00:11:40] But to be honest, in the past few months, I am leaning more towards Claude. Claude is winning me over. Um, they've made some really, really great advancements. I. In terms of content generation and analysis, it does a really great job, both Claude and Gemini. I have to say Gemini 2.5 Pro is also one of my favorites, so I'm kind of bouncing back between those two.
[00:12:04] But chat GPT, I've kind of left behind. I feel bad about that.
[00:12:09] Mallory: Left behind. That is a controversial take, but I feel you, I, I use chat GBT for, uh, image generation. Do you, uh, what do you use for image generation if that's something that you do? I
[00:12:20] Jason: have been using it for image generation, but kind of indirectly.
[00:12:24] Mallory: Okay.
[00:12:25] Jason: So building content. I've been using Gamma a lot and Gamma does a lot behind the scenes and it does work with Dali, but it also incorporates a lot of other agents.
[00:12:33] Mallory: Very nice. So Jason, you worked at an association for many years. I'm curious, was the thought of even building any sort of interactive webs, like web apps like this, these are obviously fun and not necessarily association related, but was that even in the realm of possibility?
[00:12:52] Would that have been something that's just so difficult to do that you would think, well, we'll we'll put that on the back burner.
[00:13:00] Jason: I think it would've sounded impossible to us. Um, no one in the association really has a technical background. To give an example, we've been working, or had been working on building software for about five years.
[00:13:12] So, um, it was taking a very long time. And to see these kinds of capabilities now and seeing that you can have something developed in a matter of minutes that would've taken weeks or a month before, it's, it's insane.
[00:13:27] Mallory: Yeah. And when did you start your own learning journey, uh, with AI tools like this?
[00:13:34] Jason: That's a great question. Um, I would say about a year ago.
[00:13:37] Mallory: Oh wow.
[00:13:37] Jason: Um, I was casually experimenting with it. Um, but I got more heavy, I would say about seven months ago. So I was doing a bit more project management and an accreditation process. Mm-hmm. And I really needed to design some complex dashboards.
[00:13:54] That would track basically 12 different locations and their progress and put it into one simple place. And I was favoring chat GPT at the time, and it did help me do that. Now I would go with Claude. Um, I think it would be faster with chat, GPTI had to go through a lot of different iterations to make sure it all worked correctly.
[00:14:14] Um, but it was phenomenal. It gave me a lot of information and just put it at my fingertips. To track 12 different locations and where they were in this process and in that development would've been impossible with one-on-one check-ins or looking at individual spreadsheets. But to be able to pull it into one simple location, have it designed and donut charts so I could really get a visual aspect of that.
[00:14:38] Progress was phenomenal.
[00:14:40] Mallory: Yeah. Well, we've seen a dinner plan and a census data tool, which are pretty impressive. I think we can all agree for just simple, easy prompts, but I wanna brainstorm what associations might be able to build with the same approach. So if we're talking about standalone tools that solves specific pain points, I'm thinking maybe.
[00:15:01] A professional development quiz that recommends certain training resources that your association has based on the answers to the quiz questions. Maybe it's an interactive map of your upcoming annual meeting where people can click around the expo hall, maybe see what vendors will be there, see what sessions are in certain breakout rooms, see where the food is, you know the important things.
[00:15:23] And these are all useful and fun. But would be, as Jason mentioned, quite resource intensive or virtually impossible for many organizations to build, even just a year ago, a few years ago, but. Here's the thing. While these standalone apps are neat and helpful to demonstrate what's possible, they're still just individual solutions to problems.
[00:15:44] They don't fundamentally change how you operate, but what if we stopped thinking about standalone apps and started thinking about. More integrated workflows. So instead of building a single tool here and there, we used AI powered web applications to completely transform entire processes that currently take months and turn them into something that happens in days.
[00:16:05] Maybe weeks, maybe even less, maybe hours. So Jason has been working on exactly that with educational content creation. He's been working on a system that doesn't just make one part of course development easier. It kind of revolutionizes the entire pipeline from content assessment to final delivery.
[00:16:24] We're talking about going from teams of instructional designers working for. Over a year to create comprehensive courses to a single person producing the same output in a month or less with infinite customization possibilities. It almost sounds like what we're talking about doesn't exist, but it does, and I think this is where real transformation happens.
[00:16:45] The individual apps are fun, but when you start reimagining entire workflows, I think that's where you see the true power of artificial intelligence. So Jason, I want you. As best as you can. 'cause I know it's kind of a lengthy process to give our listeners and viewers an overview of the content process that you have been working on.
[00:17:07] Jason: Absolutely. Um, and I think what you mentioned about teams and bringing it down to one person is exactly what we're looking at. So from my background, I can say. You know, to create a new course, it would've taken a team of teachers at least a year to create. Absolutely. But now I can say I just spent a week creating a course, at least the content for the course now.
[00:17:32] And our current process, what that involves is leveraging a lot of Claude, um, leveraging Gemini, and we reference our old material to help construct our new material. Just to make sure it's built with our same kind of philosophy and the same kind of intuitions that was built into that other material. Um, so essentially we construct a prompt that puts all of that information together for us, um, and we can ground that prompt with other additional resources and specific examples from that curriculum.
[00:18:06] And then we can do. I can ask it to use that to construct, um, curriculum that is more focused towards accounting. I can use it to focus more towards construction. Really any and every topic that we can think of, we can do that. We can fine tune our content towards that. But it's only one step really.
[00:18:27] That's just the creation process, right? So we all know in education, you actually have to implement and use that material, teach that material, and this is where the real excitement comes in. So, you know, for those of you who do asynchronous education or even in-person education, you know the delivery is a whole other side of things.
[00:18:47] You know, how well was it delivered, how well will it be delivered? And not everything goes to plan, but really what we can do now with AI is we know exactly how it's going to be delivered because tell.
[00:19:03] In this conversation I was, I'm explaining to the things to you. I might backtrack, I might, um, lose my train of thought. Both ai, that doesn't happen. You stay, um, on topic throughout the entire time and you're following the guidelines that you want followed. So a lot of you out there in education would be familiar with Addie, Sam, experiential education, um, Cope's Learning Cycle.
[00:19:28] Um, and what you can do is you can make sure that your delivery is fully adherent to those types of, um, learning structures. So you're not guessing, you're not just building something, you're not assuming that, um, someone's gonna deliver it the way that they should, but you know that it's gonna be delivered the way that they should.
[00:19:47] And the timing of your check-ins, the timing of your questions, the timing of your learning assessments is automated for you, and it's done exactly when it should be done in learning process.
[00:20:00] Mallory: Okay, so I, man, there's a lot to unpack there and I feel like we haven't even gone through step by step, but I want to clarify for those of you who might be tuning in for the first time or who aren't familiar with what Jason's talking about, he's referring to the content inside Sidecars AI Learning Hub.
[00:20:16] So that is a library of asynchronous courses all around educating association professionals and leaders on how to leverage artificial intelligence within their organization. So it's all very association. Specific AI education. So what you're saying, Jason, is that long ago, uh, Amme and I and a few other team members worked to record all of the content in the AI Learning Hub for the second time.
[00:20:43] So I think Amme had focused primarily on like the first edition, the second edition. We had more team members involved, including myself. I believe that was, we launched it in September-ish of 2024. Right now we're recording this in June of 2025. And the team have taken that existing content, static content, if you will, and then used artificial intelligence to kind of revamp it, update it, and convert it into content that could be customized.
[00:21:12] Do I have that right?
[00:21:14] Jason: Exactly. And one thing that you can see now on the Learning Hub is there may have been some videos that were a bit long in the past.
[00:21:22] Mallory: Yeah.
[00:21:22] Jason: Um, but now they've been shortened and they're more concise and they're very much focused on the information and knowledge that was supposed to be delivered.
[00:21:30] Mallory: Okay.
[00:21:31] Jason: So, in a sense, our first round was optimizing the content, right. Making sure that it is ready for adult learners, making sure that it's delivered and the delivered with the right structure. Essentially making sure that there's a nice flow to that learning and to make sure that the learning is really happening on a genuine level.
[00:21:51] So what we're doing now, or the next step in that was looking at that and thinking about things in a mathematical sense. So how can we use variables? And these variables can allow us to customize the content later on. So say, initially, yes, we created content for associations, but isn't that information still applicable to people in business?
[00:22:17] So what do we need to change exactly to make it more contextualized for people in business? We don't wanna talk about members, right? We wanna talk about customers. We wanna talk about clients. There are key words, key phrases, and. Some slide changes, um, that we can make that will make that content more specific for that new target group.
[00:22:39] Mallory: Awesome. Okay, so reviewing all the old content as kind of the first step and then adding in variables. I think it's really interesting what you said about having AI assist you with the content itself, but also the delivery. And I think many people who work in education, rightfully so, might hear this and think, well, yes, sure, I can create educational content with artificial intelligence, but will it be good?
[00:23:04] Will it check off the boxes in terms of what I want my learners to walk away with? I think it's incredibly interesting that you mention being able to incorporate official criteria, uh, or benchmarks or whatever that may be, educational standards into your prompts. So the new content that you're creating with artificial intelligence is in line with those.
[00:23:25] Can you talk a little bit more about that?
[00:23:28] Jason: Um, so exactly what you were saying about, um, criteria Mallory. Um, so what we can do is, you know, say you have very, very specific things within your learning structure that you need done. Mm-hmm. Um, so if you're doing experiential learning, for example, or you're adhering to adult learning principles.
[00:23:49] There are very specific things that you wanna look at, right? So let's say experiential learning. Um, we wanna make sure that we're having that reflective observation phase. And within that one piece of criteria, um, you have to decide how you're gonna judge that criteria or how you want the AI to judge it, you know, so we have the definition of what it means, right?
[00:24:08] Mm-hmm. But. AI doesn't know how well you, like say what your optimal rating is.
[00:24:13] Mallory: Mm-hmm. And
[00:24:14] Jason: what your last optimal reading is. So what is a fail in terms of that criteria and what is a 100% pass? So you do have to think about this a little bit, but you can use AI to build all of this for you so you can have these conversations.
[00:24:28] So let's say with reflective observation, adults deep in learning, when prompted to analyze experiences, connect to prior knowledge and articulate insights. But what does that mean for the ai? It's it's words, right? That it can contextualize, but for you and your organization, that might need to be defined a little bit better.
[00:24:48] Mallory: Mm-hmm.
[00:24:48] Jason: So for me, what I did for that one in particular, um, a five for me is strategic reflective prompts throughout, connect to prior experience, progression from recall to analysis. And a fail for me might be no reflective prompts or processing questions and maybe one way transfer and one way transfer is pretty much lecturing the entire time.
[00:25:11] So we're not engaging the learner, we're not asking them to reflect, we're not asking them any questions. We're just giving them information and hoping they remember it. Um, and right now I have a list of about 25 criteria. Um, I'm expanding it to about 31. Um, and for some locations or organizations that might be a hundred, you might only have 10.
[00:25:35] Um. But the idea here is, is you can really customize it towards your context and you can use it to review all of your old material and then start building your new material.
[00:25:45] Mallory: Mm-hmm. Yeah, that makes a ton of sense. For an association leader who's listening to this and maybe has some fantastic. Evergreen educational content that members love, but is slightly outdated.
[00:25:58] And we know in the world of ai, basically as soon as we put this podcast out, like all the material that we talked about's outdated because everything's changing so fast. But what I wanna get into the nitty gritty of is for that association leader, in my example, uh, if they were working with their education team, the process that you just mentioned is that.
[00:26:17] Fairly manual for now, or is there an automated way that you can, like bulk review all your content, bulk, run it through, uh, that criteria, and then create new content
[00:26:30] Jason: for now? Mm-hmm. It is a little manual. Okay. Um, but what we have automated is essentially the next step in this process. Okay. Um, which is that actual delivery and assembly of all that material.
[00:26:43] Okay. You know, let's say we built something, we have our new slides, for example. We know exactly what our audio script, we know exactly what it wants to be. We know the tone that we want delivered in, and what we can do now is we can plug it into something that's been developed for sidecar and LCA. This is a learning content agent, and what it does is it pulls that image.
[00:27:07] It pulls the audio scripts and it sends it to various different AI agents, um, and it interprets it for us and puts it all into one single mp four. So I now have the content I wanted. I have an AI avatar delivering it in exactly the same way I wanted it to be delivered in. And, uh, the words are exactly what they need to be.
[00:27:31] Um, it's presenting and delivering the exact content on each slide that we need delivered.
[00:27:37] Mallory: Wow. Okay, so that's where we're getting into that interactive web app piece. So the LCA, you mentioned learning Content agent. I'm not sure if that will be the name forever, but that's what we'll call it for now, is basically a place that you can go in and then it renders the content.
[00:27:54] Is that right?
[00:27:56] Jason: Yes. Okay. And it's rendering it based on your criteria, um, and it's rendering it exactly the way that you want. So say in your educational content, um, you have a lot of updates to do, but there's still some slides or some bits of information that are still relevant and still current.
[00:28:14] Mallory: Mm-hmm.
[00:28:15] Jason: So you can go in and produce, um, or create your new content. And in this LCA what you can do is you can rearrange everything and you can do it in any way that you want. So say you wanna take two slides of new content that you just created, but you wanna take five additional slides from, I don't know, different courses that you've offered in the past, and put it into that same presentation or into that same video presentation, you can do that.
[00:28:42] And it will again, adjust the voice. It will adjust the delivery exactly according to your criteria. So you're taking the old content and refreshing it and reassembling it with your new content and exactly the way that you want and, and according to how you need it delivered.
[00:29:01] Mallory: And so the final product at the end of this process is an MP four file, a video that shows the slides and has an AI avatar delivering that content.
[00:29:12] Jason: Exactly, so for now it's just MP four, but we are expanding this to be different kinds of documents, different types of sort of micro projects and different activities, and we'll have more downloadables like PDF. And the thing is it'll just be completely automatic. So it might be old and we're refreshing it or it might be recontextualizing it entirely or in rebuilding different pieces of it.
[00:29:36] But essentially work that took, again, like we said, years, several months is now taking weeks, and this assembly process takes a couple days. When you're looking at a high, a high bit of content. If you're just looking at a couple videos, matter of minutes,
[00:29:55] Mallory: I mean. Just to really dive in on how crazy this is.
[00:30:00] I mean, for example, if we had created content in the AI Learning Hub about general large language models, and we did that fairly recently, so maybe we kept talking about Claude 3.7, one of my favorite tools. Well, at the time of the recording of this podcast, Claude four has been released in any other version of this world.
[00:30:19] We would not rerecord that entire course because Claude three seven had changed to Claude four. We just wouldn't, and sure, you might think it's just semantics or a title change, but the models are different. They do different things. Claude four, in my opinion, is much better than Claude three seven. And so having the ability to change even a word or one feature that rolled out with Claude four that Claude three seven didn't have in a matter of minutes is what you're saying.
[00:30:49] That is insane like that, that is transforming what is possible with education right in front of our eyes
[00:30:58] Jason: and even think about the recording process. Yeah. You know, people make mistakes. When I've been speaking here, I've made some mistakes, right? And you can't always go back and record, right? Sometimes you just accept that mistake.
[00:31:10] But with this kind of delivery system and this kind of assembly process, there are no mistakes. Is a, it's a quick fix and it nonexistent. Phenomenal, insane.
[00:31:23] Mallory: So Jason, I would love for you now to share screen, if you will, and show us this learning content agent or LCA, just to give us a little idea of what you mean by, um, using this app to actually render the content.
[00:31:38] And kind of, if you can talk about, so you mentioned refreshing the old content, uh, explaining how that connects to this would be helpful as well.
[00:31:49] Jason: Absolutely. This is the, the fun part of everything. Yeah, I think so. Essentially what you see on screen now is an inside look at our LCA, um, and you see this long list of things.
[00:32:01] These are basically all of the videos that we've recently generated.
[00:32:05] Mallory: Mm-hmm.
[00:32:06] Jason: Um, but I want to really walk you guys through what a process looks like. So, for example, if I build a. First off, we can, let's see, uh, just demo name. Mm-hmm. We're just demoing something. And as I mentioned before, we'll have various different types of activities available.
[00:32:30] So discussions, lectures, readings, interactive content. For now, we have just our videos, um, our MP four content, and let's go ahead and say we're gonna build something for our general learning hub. Um, so what I. Once I've put in those names, I wanna assign a course. Let's say we're just doing a foundations course and I wanna browse my assets.
[00:32:56] Now what happens is when I say browse assets, I'm getting access to every, um, presentation. Every video we've ever done, I. And it's hosted on for us box.com, and we've got it put into these different folders where I have generic accounting content association content, um, different use case libraries. What I can do is say, I can go into accounting, we're building something for, um, general associations, but let's pull something from accounting.
[00:33:29] Accounting is applicable to everyone. Um, and let's go into foundations. So here I have one slide that's accounting specific and I can see an image of that slide. Lemme just go ahead and select one. I'm navigate back. Um, let's say after that, let's go back into accounting and I wanna pick from a different set of slides.
[00:33:55] Um, let's see. This one we want to talk about time drain reality, which is definitely applicable to everyone. Let's also go into my general association assets. Unfortunately, I don't have thumbnails for these generated right now, but that's okay. Um, so what I'm doing here essentially is pulling from content from various different courses and from what we call different schools.
[00:34:24] So a different school would basically be a different content focus. So we have account and associations. Um, and then I can go into my general asset bank, which is meant to be applicable for all types of organizations. And maybe I just pull something in from my AI agents course. And what I can do once I have those, um, what we call assets.
[00:34:46] Assets are basically images of particular slides within a slide deck. Um, I'm gonna come down here, you selected. This is where phenomenal magic happens that, um, as a non-technical person, um, I could never explain to you, but that, um, cloud Code has done a really great job in supporting us in creating. Um, so I can review my slides, I can put them into different order.
[00:35:15] Maybe I didn't click them in the correct order when I first selected them, so I can rearrange all of that. Um, maybe I'm looking at the images and they don't make sense in that particular order, so I can easily rearrange that and I can save it or I can submit for approval. And what happens here is that magic that we talked about before when I submit for approval.
[00:35:37] In the background, it is consolidating all of this content, um, pulling out the scripts for the audios and pulling out or connecting that to, um, our AI avatars and then assembling it in a nice, neat package, um, as an MP four. So way I can download that video and then post it onto our learning hub, which means within this processing would take.
[00:36:05] Uh, maybe five to 10 minutes as we've only selected a few slides. But it's doing all of this generation. It's generating the audio, it's paying attention to tone that I've selected and I've chosen, and we can take a look at that too after this. Um, and again, it's following that learning criteria that, um, we had set up previously.
[00:36:26] So it's, it, it's absolutely insane. I mean, think about how long it would've taken you to create that slide again. Um, or, you know, copy a slide from this presentation, copy a slide from that presentation, put it into a new presentation.
[00:36:41] Mallory: Mm-hmm.
[00:36:41] Jason: Um. Then give it to someone who's going to, um, write out the script that they're going to read or perhaps impromptu, um, do that recording rerecording if they made mistakes, all of that time.
[00:36:55] Um, all of that extra investment that we used to have to do is. Old school, um, we don't have to worry about doing it anymore,
[00:37:04] Mallory: old school. That that might be a fun, uh, phrase we can put into the topic. Like, don't build your education the old school way. I want to give our listeners a little more context, so.
[00:37:15] At Sidecar, we've had many associations come to us and say, we love your AI education and we want to do something like this for our members. Can you help us? And we have done, uh, private bootcamps based on like the specific associations, industry or profession. We've done webinars like that, so on and so forth, but we can't really.
[00:37:38] Scale that we only have a certain amount of people at Sidecar, we can only travel so much. We can only do so many webinars. So it was very difficult for us to bring AI education to association members. But what Jason's talking about here with having like the association, AI Learning Hub, and then the accounting one allows us to do that.
[00:37:58] So if we have an association with lawyers as members come to Sidecar and say, we would love to bring a version of the AI Learning Hub to our members. You can do that, we can help you do that. And we can do it really, really quickly as Jason just showed. So I wanted to give you all that context, but even if you're listening as an association and you're wondering why you would need to customize your content, um, I can think of a few examples there.
[00:38:24] So maybe if you have different segments of your membership in terms of profession. So maybe you've got students as a part of your membership, maybe, maybe you've got early grads, mid-career, maybe you have retired people. Imagine taking a really successful piece of educational content and then tailoring it for those segments.
[00:38:41] So maybe if you're talking to students, you wanna use terminology that's relevant for them. If you're talking about mid-career folks, they probably don't care about student terminology. Another example of that is if you're in healthcare as an association, uh, in a specific specialty like cardiology, and maybe within your membership you've got some nurses, you've got some PAs, you've got some nurse practitioners, you have some doctors.
[00:39:05] For example, example, you could take the same piece of content and customize it for each of those segments and make sure that the terminology you're using is most relevant for them. And you can do that before, as was saying. You just wouldn't have done it because it would've taken so long. But now you have this available to you, basically at your fingertips.
[00:39:26] Um, the last thing I wanna cover here is that when we started the episode, we talked about interactive web apps. We were just using regular old Claude, regular old chat, GBT. The thing that Jason demo, the learning content agent LCA was uh, not just built by going to claude.ai and like typing in a single prompt.
[00:39:47] I want you all to know that that involves some more heavy development work. We have some fantastic developers on our team who've been working on that. Heavily using Claude Code. So if you wanna just build a simple interactive app like we did with the census or dinner, you right now can go to claude.ai, give it a prompt and you would have a nice prototype that you could hand off to your development team.
[00:40:08] Or perhaps something fun that you could show your kids or, uh, you know, help you create dinner at night If you wanted to create a web app. That's more transformative, like the one that we showed you in the educational process, that would be a bit more involved. So you'd probably wanna have some skilled developers going to Claude Code or another AI coding tool for that.
[00:40:29] Jason, you did mention tone in that demo. Uh, what do you mean by controlling delivery and tone? Like what options do we have for now?
[00:40:40] Jason: You limitless. Okay. Li Limits. I love it. Um, yes. So basically, um, we leverage 11 labs audio, and with any 11 labs you have a, a plethora amount of options available to you. Um.
[00:40:57] You can adjust tone, you can adjust delivery speed. Um, you can choose from a variety of different accents. Um, you can choose for people who are more educators or people who are more into marketing. And if you think about it, you know, when you watch a commercial, it's not the same tone, it's not the same kind of personality that you would see in an educational environment.
[00:41:20] Right? They might actually deliver the same information, but not in the same way and not addressing the same kind of audience. Um, so you can fix all of this within the structure so I can show you and give you a little bit more insight into that. So with that same platform that we were walking through, we have something that we call dictionary.
[00:41:44] And what I can do is I can do this on a course by course, slide by slide, a lesson by lesson kind of basis. And I can customize exactly who my avatar is going to be. Ah, I can customize how that avatar introduces themselves. So right now we have them introducing themselves as AI instructors. Um. You can customize exactly what voice you want to use.
[00:42:13] You can customize some of those different variables that we had talked about before. So say you're pitching this to maybe as an association, you're pitching it to your members, right? I. Let's say you wanna use that same exact content and you wanna pitch it towards, um, a company, right? You would no longer be using members, but you would be using clients or you would be using customers.
[00:42:36] This allows us to automate all of that, so we would automatically change those specific, um, vocabulary words to the context that you're targeting.
[00:42:47] Mallory: I'm sitting here smiling. This is just so cool. I mean, I, it's insane. I knew we were working on this, but I think seeing it, this is actually my first time seeing it demoed to me, and it's fairly new, right?
[00:42:59] Like when was the LCA system available?
[00:43:03] Jason: I. It is extremely new. Every day we're implementing new features. So about a week ago, I couldn't make these changes per slide or per lesson, but now I actually can. I can go in and set up any variable that I want across individual slides, and this is a really unique opportunity because same.
[00:43:23] You want two people to have a conversation, you know, um, when you go through and you edit these podcasts, you're able to display one person talking and then another person talking. You can actually do this now within this content. So yes, you're not using real people, but you can still have your avatars achieve the same goal.
[00:43:43] So one avatar could say something in their particular voice. Right afterwards, another person could say that, you know, sorry, I say person, the other AI avatar can say the thing in their particular voice and you're mimicking that kind of conversation or you're mimicking that same back and forth. It's, it's mind blowing.
[00:44:03] Really.
[00:44:04] Mallory: Oh, it's so mind blowing and so incredibly exciting. So. As someone who, I mean you, you were there for like the inception of, of this learning content agent. Yes. Learning content agent LCA for association listeners who are hearing this, and hopefully their minds are blown as well and who are thinking perhaps we should do something like this within our organization to better our educational offerings.
[00:44:32] As someone who. Was fairly new to ai, new to sidecar, new to the LCA. What advice would you have for them? Are there any kind of things that they should have in place before this, or are there ways that they should start thinking about building their current educational content? Maybe they're already in the works of a new webinar, a new asynchronous course.
[00:44:55] Things they should keep in mind while they're building those, if this is the direction they wanna go.
[00:45:01] Jason: Absolutely. Um, the first thing I would suggest is joining our Learning Hub. Our learning hub walks you through this entire process actually and gives you all of that background information that you would need to move forward.
[00:45:14] Um, I think there are a lot of steps involved in this, but really the first one is experiment. You know, throw in a couple random prompts, um, test really the limits of what's out there. Um, don't say this isn't possible. I think that kind of thinking before of, oh, we can't do this because it would take too much time, or we don't have the resources.
[00:45:35] Try a prompt. See where chat GBT gets you see where Claude gets you, um, and start having that conversation. And I think you're gonna end up building something on accident. You'll end up with something that, um, really changes, um, changes the entire game plan.
[00:45:55] Mallory: Yeah, I mean, imagine some industry report coming out and you as the association being able to create.
[00:46:02] Some educational content on that in, I don't know, instantly. Instantly. Or, I mean, I guess it would take you a little bit of time. That's where going, we're close. You're right, that's where we're heading. Maybe a few hours, maybe days. Imagine that providing that level of service and value to your members and what that would mean for them and how that will impact them as, as humans and as professionals.
[00:46:23] So I think we're talking about AI here and we're talking about, you know, coding and interactive web apps, but. I think we're really talking about something much greater, which is helping humans learn and retain content, um, in ways that they never have before. So I think that's incredibly exciting. I am gonna have to take that class on the, uh, that course, on the AI learning hub myself.
[00:46:45] 'cause I'm, I'm quite interested in like, the inner workings of this, you know, as I slightly become more technical day by day co-hosting this pod.
[00:46:54] Jason: It's, um, it's a phenomenal change. And I say instantly because again, I'm coming from this, you know, background of where this took a tremendous amount of time.
[00:47:04] And like if we're talking about hours, days, or even a week, that's like instant in my, in my book. Absolutely. Um, but we are, we are getting to a point where this would take minutes, which is. Uh, mind blowing.
[00:47:19] Mallory: I mean, this is truly, truly crazy and so relevant for associations. Is there a course, Jason, that you are most excited to launch in the AI Learning Hub?
[00:47:28] Is there a feature on the LCA that you're most excited about? Anything that you are looking forward to on the roadmap?
[00:47:35] Jason: Um, what I am most looking forward to is taking all of that content review and content development process. Mm-hmm. And automating it.
[00:47:42] Mallory: Yeah.
[00:47:42] Jason: So essentially we can now build something Right.
[00:47:46] But. You should always review what's been produced, right? Nothing is ever perfect. So what I want to look into next is really leveraging that criteria that we used to build and using it to judge. So once we've built some of this new content and we've assembled it, and we have our final MP four, what I want to do is I wanna send it through that process again to be reviewed, just to make sure we've covered everything that we want to cover.
[00:48:12] Mallory: I love that. So having AI assist you with the final product review as well, which Amit talks about all the time, kind of having one AI review, another AI's work. Uh, wow, that's fascinating. And you
[00:48:24] Jason: said you were very close.
[00:48:25] Mallory: Very close. Um, oh, the last question I wanted to ask is, and this might be a better question for am.
[00:48:32] Do you know how many developers we have working on this? I think it's only a couple. And it's not even full-time, right? I think they're just like working on it periodically. Is that correct?
[00:48:44] Jason: Yes, it is one of their priorities for sure. But they're focused on other different projects, um, at the same time. So it started with, um, three.
[00:48:54] Mallory: Yeah.
[00:48:55] Jason: And now it's. I mean, really. So I was in a team where I would see these updates week to week, essentially, and what you would see week to week, and it's insane. I mean, this is, you're seeing things that you wouldn't have seen for months and you're seeing them be implemented and ready to use in a week.
[00:49:16] It's, it's insane. Crazy.
[00:49:18] Mallory: I was just asking that because I imagine, you know, we have some listeners who are like, that's great, but you know, we have access and we do to some f fantastic resources and very AI savvy developers. But you heard that only two developers right now, and they're not even dedicating all their time to this.
[00:49:35] And we were the Guinea pigs. So I will say, if this is something that you wanna implement in your association, go and take the course in the AI Learning Hub. Is it available? Jason right now, or it will be, there's
[00:49:46] Jason: not a specific course on the LCA, but everything that's in there now really tells you the inner workings of it.
[00:49:53] Okay. It's not direct. Yes. But all that information that's in there would really set you up to launch, essentially.
[00:49:59] Mallory: Exactly. So we were the Guinea pigs. You can go and. Learn all that information that j Jason just mentioned, and I would bet you could get something up and running like this with, I don't know, maybe even less than two developers, uh, since we've done some of the legwork.
[00:50:14] So everyone go to the AI Learning Hub. If this is of interest to you, please head to claw.ai or chat GBT, your preferred tool and make some interactive apps. Figure out what you're having for dinner. Uh, create a fun little quiz for your kids. Play around with this. Even if you're non-technical, especially if you're non-technical, because you have so much information and resources available to you at your fingertips if you are technical, even better, everybody, thank you so much for tuning in to today's episode.
[00:50:48] I'm hoping you all got a lot out of it. And Jason, thank you for sharing your incredible experience and expertise with listeners. I think this has been a good one.
[00:50:58] Jason: Absolutely. Thanks for having me on. It was great.
[00:51:05] Intro: Thanks for tuning into Sidecar Sync this week. Looking to dive deeper. Download your free copy of our new book, ascend Unlocking the Power of AI for Associations at Ascend Book. It's packed with insights to power your association's journey with ai. And remember, sidecar is here with more resources from webinars to boot camps to help you stay ahead in the association world.
[00:51:27] We'll catch you in the next episode. Until then, keep learning, keep growing, and keep disrupting.