Edtech Insiders

The Rise of AI-Powered Educational Tools with Yair Shapira of NovoDia

• Alex Sarlin • Season 10

Send us a text

Dr. Yair Shapira is the founder and CEO of NovoDia, an EdTech company innovating educational publishing through AI technology. Previously, he founded and led Amplio Learning to significant growth before its acquisition. With expertise in technology and educational tools, Yair is dedicated to enhancing educational equity and efficiency. He holds a Ph.D. in Biomedical Engineering and is an active contributor to the EdTech community.

💡 5 Things You’ll Learn in This Episode:

  1. How NovoDia is transforming educational publishing through AI-driven solutions.
  2. The role of generative AI in creating dynamic, engaging, and adaptive learning materials.
  3. Why current curriculum models fail to meet diverse classroom needs and how AI can fill the gaps.
  4. How NovoDia enables districts and teachers to customize curricula to meet local and individual requirements.
  5. What the future holds for AI-powered tools in education, from multimedia lessons to personalized learning experiences.

✨ Episode Highlights:

[00:02:26] Yair shares his journey from Amplio Learning to founding NovoDia.
[00:05:37] The unique challenges of using AI in education and meeting pedagogical rigor.
[00:11:01] How NovoDia addresses the mismatch between district-purchased curriculum and teacher needs.
[00:18:22] Customizing curricula with AI to support diverse student populations and regional standards.
[00:30:43] Enhancing engagement with multimedia tools and interactive content in classrooms.
[00:37:55] The evolving role of AI in education and NovoDia’s vision for the future.
[00:44:04] Ensuring pedagogical accuracy with AI systems and human SMEs.
[00:46:59] Resources and advice from Yair on staying ahead in the fast-moving EdTech space.

😎 Stay updated with Edtech Insiders! 

🎉 Presenting Sponsor:

This season of Edtech Insiders is once again brought to you by Tuck Advisors, the M&A firm for EdTech companies. Run by serial entrepreneurs with over 25 years of experience founding, investing in, and selling companies, Tuck believes you deserve M&A advisors who work as hard as you do.

[00:00:00] Yair Shapira: It continuously improves. Think about it this way. It's an AI system that sees a curriculum as a molecule.

This is the molecule built of atoms that eventually boil down to specific activities, even specific images and phrases. very, very long taxonomy. And while building it, because it's a system, because it's a computerized system, it ensures the consistency and the accuracy. There is no way that you'd find what you find often in textbooks, that page 74 is contradictory to page 26.

Because it was built by a different, it was written by a different person, maybe in a different year.

[00:00:45] Alex Sarlin: Welcome to EdTech Insiders, the top podcast covering the education technology industry. From funding rounds, to impact, to AI developments across early childhood, K 12, higher ed, and work. You'll find it all here at EdTech Insiders. Remember to subscribe to the pod, check out our newsletter, and also our event calendar.

And to go deeper, check out Check out EdTech Insiders Plus, where you can get premium content, access to our WhatsApp channel, early access to events, and back channel insights from Alex and Ben. Hope you enjoyed today's pod.

Dr. Yair Shapira is the founder and CEO of Novodea, an EdTech company innovating educational publishing through AI technology. Previously, he founded and led Amplio Learning to significant growth before its acquisition. With expertise in technology and educational tools, Yair is dedicated to enhancing educational equity and efficiency.

He holds a PhD in biomedical engineering and is an active contributor to the edtech community. Yaya Shapira, welcome to EdTech Insiders. 

[00:01:56] Yair Shapira: Thank you for inviting me, Alex. 

[00:01:58] Alex Sarlin: Ah, I am really excited for this conversation. You and I have been chatting for a number of months now about the amazing work you're doing at NovoDia.

Let's start, you know, when we first talked, you gave a little bit of an overview of NovoDia. How you came up with the idea for Novo Dia and how it came from sort of some core capabilities of what generative AI can do. So let's start there. Give us the story of Amplio and how you moved into the content and publishing space with Novo Dia.

[00:02:26] Yair Shapira: Sure. So, uh, I started Amplio about a decade ago with, this is my former, uh, education company. And Amplio was focused on special education and succeeded quite well. And eventually was acquired. And I stayed for another year or so and left, uh, late 2022. And now, if you remember, late 2022 was the early days that, uh, we layman had, uh, some capability to touch generative AI and were owed by the capability of Dali, which was now looking back two years ago, it seems like, uh, prehistory, right?

But it struck all of us that this is going to change. Any and every content area in the world, whether it's marketing or legal or animation and education eventually is a content industry. So it was clear to me that definitely generative AI is going to change the textual, the visual aspects. of education.

It was clear to probably hundreds of edtech entrepreneurs. So we all went to think, what is it that we can do and need to do with the generative AI to change education? And my thought path was, I think a bit different than others, because generative AI has more or less the same promise in all industries, right?

It says, we're You can create great content faster, cheaper, that's the idea, right? But each industry has its own, uh, characteristics dictate how AI is going to enter the mainstream. And if you don't follow the characteristics of the industry, you find yourself not in the mainstream. And in education, going mainstream means, I believe one thing, at least now in 2024, 2025, it means that you need to satisfy the needs of the schools, the districts, the states, eventually they are the regulator, and what they require, I think justifiably, is that there is pedagogical rigor, and that it's standards aligned.

And if you do that, then you can maybe find your way Into the mainstream of education. But AI in 2022, also in 2024, as is just does not meet these requirements. I mean, we've all seen, uh, glitches and hallucinations and even if you package it with nice prompts and user interfaces and many feature tools do, it still doesn't get closer to the rigor that schools and administrators and states, uh, would require.

It can still make a great tool for teachers. But it won't enter the mainstream classroom, and behind it, I think, is a fundamental difference between what large language models that the LLMs provide, and what education is all about. So, LLMs are designed to statistically please, right? They find the statistically most pleasing response.

And every teacher knows that education is exactly the opposite, right? Even parents know it. A good educator challenges other than pleases and the intricacies of education make it quite difficult for AI to go mainstream. And this is what we were set to do. 

[00:05:37] Alex Sarlin: Yeah. I mean, I think you said hundreds of entrepreneurs probably had the same realization, but you know, I've talked to hundreds of entrepreneurs that I'm not sure that's entirely true.

I think, you know, everybody's sort of realizing the capabilities of generative AI at different. Paces and in different ways based on their prior, you know, knowledge or prior understanding prior sort of, you know What they've been working on and I think what's so interesting about your take on content creation and delivery is you're saying hey Well, we know that there's a compliance and standards and you know There's all these things that have to be true of curriculum for it to make it into a classroom in a meaningful way yet ai also can make incredibly compelling, engaging, interactive material that students really find exciting and compelling.

So instead of thinking about a textbook as always being, you know, that stock picture of Frederick Douglass and that, you know, that stock picture of the Pullman riots, you know, I mean, it's just like, there's this vision that we have of educational content and you say, well, it doesn't have to look like that anymore.

And it can still fulfill all the pedagogical rigor, standards, compliance. I think that's something I actually haven't heard very many people say. So tell us, you know, your take on educational publishing and why Novodea could be such a disruptor to it. 

[00:06:52] Yair Shapira: Yeah, so educational publishing, you know, let's start from the end.

At the end, 70 percent of what teachers share in the classroom is teacher sourced. It doesn't come from publishers, it's not funded by, uh, by the district. So the reality is the districts go and spend tens of millions of dollars, I guess, each district on curriculum. It doesn't reach the classroom. They practically reject it.

I think one of the first things we did was to go and ask teachers, why do you do it? Because it was quite amazing that teachers are spending an hour a day sourcing their own content when they are mandated to use the curriculum. And teachers said, and again, you'd not be surprised. They said, Well, the content that we get from the school is not engaging for our students.

It doesn't meet their needs. It's not always relevant for who the students believe and feel they are. It's too rigid, so we don't have the autonomy that we use. Keep in mind that teachers are the rock stars of the classroom, right? They truly know what's best for their students. They know what the students want.

So if they are doing it, and they are doing it on their spare time, weekends, nights, probably the curriculum just doesn't meet what the classrooms need. Add to this another point, because there is a main premise in publishing. The main premise is that publishing is very difficult and the costly, expensive, slow.

And because of that, because of those economical reasons, the large, the main publishers, focus on the center of the bell curve and they focus on the center of the bell curve maybe in two dimensions one is topical so they focus on the main subjects math and DLA not to say that they don't provide other things but that's 95 percent of what they do is math and DLA and they also focus on the center of the bell curve in terms of the population they Target the average gen ed students.

Now show me an average gen ed student. Nobody's average. All schools are different. I mean, they teach so many different subjects and they teach it to a diversity of students. And there is a need for solid curriculum for every subject, for social studies and for CL and for arts. If it's taught in school, why doesn't it deserve a great curriculum?

And what about the student populations that are not exactly at the center? What about newcomers and gifted and struggling learners? They too deserve, uh, a curriculum. So current publishing is full of holes today. And again, the proof is in the fact that teachers reject and don't use it. And that's our job.

[00:09:21] Alex Sarlin: Yeah. And I think, you know, that is a realization that I think, you know, everybody who's been in a classroom in sort of any capacity over the last few years has seen in real time, this idea that Teachers go to Teachers Pay Teachers, teachers go to YouTube, teachers go to, you know, the internet, they go to just sort of all different places to try to supplement or even replace the curriculum that the district has bought for them, that sort of the mandated curriculum.

with different types of activities, different types of material to engage students, to make it more personal for them, to make it more relevant to them, or just to sort of make the classroom experience more fun. And, you know, we've all seen that. We all have seen that for everybody in ed tech. I've seen that for years, but I think what's so interesting about your take on it is you say, well, that actually creates a fundamental.

mismatch and sort of an incentive misalignment between the districts that are trying to buy high quality, terrific curriculum that actually gets into the hands of teachers and then students and teachers who are saying, I have to spend all this additional time, sometimes even money, of course, adapting what I know I need to teach and getting the materials that, you know, making my own materials, buying materials, you know, online.

That's such an interesting dichotomy that I think, you know, we all see, but I don't know if very many people have sort of recognized it as the scale of problem. It is. So tell us a little bit. You've talked to a lot of teachers. You've also talked to a lot of districts, you know, tell us about how that works on the ground on a day to day basis.

How does it affect the school culture and the sort of the whole? Um, Curriculum situation when people are spending, you know, millions of dollars on curriculum and then teachers are having to go supplement or replace with additional material. 

[00:11:01] Yair Shapira: Yes. So, you know, the teachers, as I said, that the spending time, as you said, they go to YouTube and Kahoot and teachers pay teachers and they bring a new stuff and they are right because they know what the students need for the administrators.

It's a nightmare. Right. We hear it time and time again. I mean, the leadership just loses oversight on what's going on in the classroom. There is research from Rand showing that leadership is terrified and they don't know 50 percent of what's going on in the classroom. So they're terrified by what they know.

They don't yet know what they don't know. And then they start telling us that they cannot defend teachers who bring their own stuff. How can you? And it's only getting worse, and they see it day to day. I mean, newer teachers, younger teachers tend to reject publishers content even more. With the proliferation of AI tools, it's even easier.

Why go and search in YouTube? If you can just create whatever you want in a minute with a tool or directly with a GPT, it's easier for teachers to throw away the formal curriculum. And I think most importantly, it's for the students. I mean, they go home, they watch Netflix, they watch Disney, they go on TikTok or whatever.

And then we expect them to learn from the content practically comes from another millennium. I don't know if you know it, but if you Google for 1900s educational materials, you'll find that 125 years ago, education was at the forefront of art. It led what art looks like. I remember myself as a child in the 1970s, you know, flipping pages in the encyclopedia just to see images of faraway nations and wildlife and so on.

It was just fascinating. All this is somehow gone. And I think there is a way to go back to fascinating educational materials. 

[00:12:53] Alex Sarlin: Yeah, well, and I think we all sort of know, even if we hadn't recognized it, you know, why that has happened, because, as you say, educational publishers, there has been a really high barrier to entry until recently.

And so, if you're a publisher creating K 12 or college textbooks, you're on edition six of your textbook, and you had the same textbook. book with, you know, relatively small updates for decades, and it's still going to be the textbook of record. It's still the same one that people are buying again and again.

So they haven't had to change. Let's put it that way. Even in a world where there's so much competitive content online, where there's so many different people trying to create educational content, they haven't had to adapt because they have a monopoly on the standards aligned content, especially for the first generation.

Big states like Texas and California and New York and Florida. And I think it brings us to something that you mentioned, but I want to double click on, because I think it's so interesting, which is that the middle of the bell curve point you made, right? That publishers, because it does cost a lot to make content and because they're trying to sell often similar content across the country, they've always had to sort of hue to the average, the quote unquote average, right?

The classes that are taught the most and you sort of align to students with, you know, that are right in the middle, quote unquote, they don't want to push people too hard. They don't want to make it too remedial, but like everything has to sort of go right to the middle. And every teacher knows that is not how any class looks.

So tell us how that is changing in this AI era. 

[00:14:23] Yair Shapira: Yeah, so NovoDia, my new company, already a year and a half. Is it an AI publisher? So we go into the exactly the same model that traditional publishers are following. We offer full curriculum to districts with only one difference. We do it with AI, not only AI today, AI maybe take 70 percent of the editorial work and maybe in the future it will take 95%.

And the rest, which is the most meaningful part, is done by subject matter experts, the SMEs. And with that, once you take away the element that things are so complex and slow and expensive, and you plug in something that is more agile, that is faster, that is not that expensive, but still maintains the rigor and the standardization, and even significantly improves the engagement.

You can do so many things For example, you can address subjects and topics that are just unserved all the long tail Of social studies scl civics arts foreign languages, you name it You can just go and address it immediately because it makes economical justification You can even adhere to the new local and evolving standards, which is a trend that It's all over the nations, right?

States make their own standards. And in some areas, the local standard is not because they have a different view. It's because they want to teach local history, right? And local history in Maine is different than local history in Washington. And you can do that. And no publisher in his right mind would even think of developing Something for history social studies in it's just too small.

It's not justified And you can create a specialized curriculum for groups of students those who do not respond to the existing curriculum And you can customize and develop curriculum on demand not just for LAUSD but for any district that comes and says We need this, we have a special view, we have a special community that we serve, we have some special needs and so on.

Why not create a version of a curriculum or just start from scratch and we need a specific curriculum for what we see and what we need. And most importantly, bring curriculum to what 2025 students and teachers need. And expect and deserve, which is so far from traditional curriculum and is the reason that is rejected.

So we get requirements from all sorts, things like resiliency studies and Native American history and culture and LGBTQ studies and science and ELA for students with behavioral challenge and all kinds of curricula for English learners, and then so many more. And we see a trend. We see a trend where first districts.

Before us, it doesn't have to do with us. Some districts are so fed up with publishers content that they go and develop everything in house. It's strategic for them to become their own internal publishers. Maybe it will take them years, but they will save so much money, and they will tailor make it for what they need it like.

And I think we can be very instrumental in this. And the second thing that we see fairly new is that once they see how fast and how accurate the curriculum that we provide is, they start to set free to the imagination and also hunger for curriculum. They start, you know, thinking about curricular gaps that they have in topics that they never.

even dared thinking that they can have a curriculum in. Things that were totally overlooked. They think about specific students groups that are unserved and now they can tailor things for them. And they even think about new curriculum modalities like boosters for learning loss and after school resources which are always neglected because they are not in the center of the care.

So I think it is already moving to the point that it creates its own demand. 

[00:18:22] Alex Sarlin: I mean, I think it's worth even taking a moment and unpacking some of the different use cases you just named, because they're so powerful individually and it's all possible with this new era of AI publishing, but let's just unpack them a bit because this is some of the most exciting stuff I think happening in the space, you know, We talk a lot in edtech about, you know, quote unquote, personalized learning.

And I think, you know, it's a hot term. It's been around for a long time. It's had lots of different meanings. I've sort of migrated my thinking to be a little closer to what I call precision learning, rather than it being personalized as if it's always literally for every individual, which I don't think it actually necessarily will be.

It's precision, meaning it gets to exactly what you need. people need, whether that's a single person or a group or a state or, you know, a subject. And what's so interesting about what I'm hearing you say here is like, you're saying, well, look, you can fill in, you can basically adapt to the changing standards of an individual state.

You can adapt to the regional needs of an individual state. I think one of your pilots was with New Mexico and it was about Pueblos, right? There's like the idea that they have standards in New Mexico about learning about Pueblos. Pueblos because it is a huge part of their local history and no publisher is going to go very deep if they mention at all Pueblos because that, because New Mexico is not their core audience, but it is very easy with a tool like Novotia to make a unit about Pueblos that can go very deep or have any kind of, you know, included standards and be engaging and exciting and interactive.

So the subject matter is there, the changing standards is there, and then perhaps the most exciting one. For everyone, I think is you can adopt it to different student needs, and that could be an individual, but it also could be multi language learners or English language learners. It could be for students with behavioral issues.

It could be students who have a certain interest, even right. It could be a people who are want to do an after school program about something, and they just there's no curriculum for them because You know, maybe it's an after school Japanese learning club. Hey, school isn't used to being able to support that, but they can when curriculum is this quick to create.

There's so much opportunity that I think you're really spotting when curriculum creation, the cost and speed goes down. So far, one aspect of Novo Dia that you haven't mentioned so far, there's two that I want to. That I really want to dig into because I think when people are trying to envision what this looks like, it might be hard to envision it when it just feels like, oh, you can do anything is there's these sort of two levels, right?

There's curriculum that can be adapted and bought by a district like Pueblo Studies for New Mexico district might buy Pueblo Studies or the state might buy Pueblo Studies. And then the teachers actually still have the opportunity to customize. As they deliver that curriculum to their individual classroom.

And what's interesting about that is that customization serves some of the, you know, ideally they don't have to customize it that much. It's different than what's happening from original, from traditional publishing, but they still have the opportunity to customize it and they can make it even more precision, even more precise to the exact needs of their classroom.

So tell us about, you know, the flow of a district. Buying content and then an individual educator adapting it further to make it even more precise for their needs. 

[00:21:38] Yair Shapira: Yes. Spot on Alex. I think that the mega transformation that you're hinting to here is that, and again, it's because of economies because we are in a business and so do publishers, but when everything is based on the cost of manufacturing and the cost of development, Then your entire attention is internal, your entire attention, you put your back to the, to the customers and you look only internally.

How do you save costs? How do you reduce editorial time and so on? Once this is solved and with AI, whether you reduce 70 percent or 95%, it's only a matter of time. You can turn around 180 degrees and look at your customers, at your clients, at your users, at your beneficiaries. And these are the districts and the teachers and the students.

What do you need? What do you need to have better life, right? And you'll get different answers from different people, but you need to serve them just like in, in software. You serve your clients and you always want to satisfy them and to, and to please them, and this is exactly what you're hinting to, and we want to satisfy all of our users from the district, through the teachers, and eventually most importantly to the students.

And with precision, because the world is not fair, the world is precise. Everyone should get what they need. And we do it through the process that we work with districts. So we meet district leaders and we ask them, what are your gaps? Sometimes it's not such an easy question, because sometimes they have to open up to thinking what the real gaps are.

So you speak to C& I leaders, and they always tell you ELA pacing guide. Okay, let's think broader. And eventually they come up to, why not, why can't we somehow weave in social studies into our ELA classes? Or how are we able to deal with the socio emotional learning, or with the resiliency? Or how do we also find a way to provide math to kids with a learning loss?

And it takes them a bit of time to broaden up the scope and not be so terrified by talking about things that are not ELA and math for middle of the bell curve. This you have, for now. So they open up and start raising more and more, uh, needs. Sometimes they have to speak with their C& I teams, with the curriculum directors, with the coordinators, with the teachers.

And it's a rolling thunder. You just hear more and more voices all around the district. And eventually we often get Either a need that we already developed for, or new stuff, which we qualify because we want to also to make sure that it's a real need and not just an imagination. The next thing we do is if we already have a curriculum ready made because we've created it for another district, then we ask them, do you need any customization, any specific population, any specific points of view, sensitivities?

And we can easily do that because in our system, the entire curriculum. All the way from the curriculum, through the units, through the lessons, through each and every activity, is structured in such a way that it's mirrored in a knowledge base, in a taxonomy, that can be easily adapted to whatever the district would like.

So sometimes they need customization. Sometimes they convince us to create a whole new curriculum, and it's not very difficult to convince us. We are happy to do that. 

[00:25:18] Alex Sarlin: Right, 

[00:25:20] Yair Shapira: it's part of our business and then we develop an on demand bespoke curriculum for them. We deliver it very fast within a week or two.

They already have the first lesson plans. Of course, everything is done in tandem with the district. They need to vet it. They need to validate that what we bring is what they wanted. So it's a very quick process. We do all the heavy lifting. They just have to imagine and we start pushing it into the Classrooms also with the training and professional development if required they start using it We start receiving feedback and the curriculum grows and grows and grows and goes into the classrooms Now comes an interesting part.

I told you earlier that one of the reasons that teachers reject Publishers curriculum is because it takes away their autonomy Now it's a balancing act, right? How much autonomy do we want the teachers to have If they use ai, if they go to GPT and invent lesson plans, then they have full autonomy. Uh, but who knows what they're doing And what about consistency and coherence and quality and vetting and so on.

And we allow them to use AI for a personalization and differentiation, uh, in the classroom. So whatever the curriculum includes. That is vetted by our SMEs and by the district. When it arrives in the teacher's hands They have the capability to modify it within a lane keeper AI model. Let's call it a tamed AI that is very pedagogical, maintains the rigor and the standardization of the content and Allows the teacher to change the character.

So if the character is, the character can be a female male Can be an African American or a Caucasian, whatever they choose, whatever age, whatever gender, they can change the level of readability on the spot. You know, they can change it to 10 different versions, the depth of knowledge, and they can change the language.

What about kids who do not speak English? Okay. We can create a Spanish or a Russian. version of the same content and all that can be done in the classroom or a few minutes before the Classroom and assigned to the students who need it. So there is autonomy. There is flexibility The data is collected. So we know that the fidelity is high and we also know how much autonomy different teachers take And that's great for the administrators.

[00:27:42] Alex Sarlin: They can change the interest, right? I mean, they can change the sort of theme of the material based on what their students are doing. They can change the names of the people, so that they can match the names of the people in the curriculum to the students in the class or to, you know, to People that, you know, do anything they'd like.

I think this model is so fascinating and, again, just want to very rapidly unpack some of the things you said because you say them as if they're so obvious. These are new things that people have never been able to do, which is like, you know, take vetted curriculum that is standards aligned, district approved, and then adapt it in all of these ways that keep the pedagogical rigor.

Intact, but can, you know, tweak in all of these ways. It can change the depth of knowledge of some of the questions. It can change some of the reading level. You know, there's lots of debate about whether changing the reading level changes the rigor, but, you know, let's leave that alone for now. But you can change the, as you say, the representation.

in the curriculum of what race or gender they are. You could change the language and adapt it. This has really not been true. You know, in a world, flashback to five years ago, and you have a teacher about to teach a lesson on, let's say, you know, the quadratic formula. I always like to go to that one, you know, quadratic formula.

And they're saying, okay, Well, I have my curriculum from an educational publisher and they've done their best to try to make it, you know, kid friendly and fun. But wow, this is too dry for me. Kids will not see themselves in it. They won't see the relevance of it. They're not going to relate. I'm going to go on YouTube and I'm going to find a really amazing video explaining the quadratic formula and why it's important.

And then I'm going to go to teachers pay teachers and see if they can find something that relates it to, you know, Music because my kids love music and then I'm going to go to Kahoot and make all sorts of quizzes to make it more engaging and interactive. And they do all of this work just to take this publisher curriculum and align it and even with all of that they still wouldn't be able to, you know, inject character to it or story into it.

They wouldn't be able to change the depth of knowledge of the questions, you know, automatically they might have to do that on their own. So like, all of these things are sort of like, It's so much work to do that level of differentiation. And all of that is literally just built in. They're just, it's just a user interface in Novodea.

You say, Oh, let's change the race of all the students and let's change the depth of knowledge. And for this group of students, let's make it in this, you know, let's make it in Hmong. I mean, it's a totally new world of what curriculum means. And it actually brings me to my next question, which, again, is something I think, I've seen the Novodea platform, and one of the things that jumps out immediately about it is just the look and feel of the materials that are being created.

They don't look anything like traditional educational materials. made in all of these different formats. There's all this interactive games. There's all of these, you know, stories and sort of almost like filmic, you know, elements. So let's talk about that as well, because that's built into Novo Dia, this sort of sense of, well, let's make it feel and look a lot more like the level of, you know, engagement that students are used to in their regular lives when they're home or when they're on their phones.

Tell us about that and what role that plays in the Novo Dia story. 

[00:30:43] Yair Shapira: Yeah, that's one of the most important pillars within what we do. My partner, Tomer, he comes from the art, from the creative world, and he brings this view and he's the guy who told me what happened to education. How come it was at the forefront of art and it's not, how come.

His kids, he has two girls, how come they go to school and they learn from something that really looks like from the 1900s and they go home and they watch Netflix? How do we expect them to even learn from this? So, we are trying to, I don't think entertainment is a bad word, we are trying to entertain the students and the teachers while learning.

In the same way that, uh, you know, Sesame Street did it 50 years ago, right? So, it's part of the AI elements that we are entering into our system. There is a textual part, the solid, pedagogic, standards aligned part, and then there is the visual part that brings learning to life. And I think that it will dramatically change, I don't even know how, I'm just watching what happened in the last year in the multimedia capabilities that AI provides, and I can just imagine what would happen in the coming year or two with multimedia capabilities.

The capability to create, you know, high quality interactive videos is so powerful in education. We have to think how to use it. And if you create it ad hoc on the spot, you know, publishers have spent millions of dollars in creating short animations That look like a shadow today, those weird clips from three years ago.

And who would imagine in a half a minute, you can create something so much better than this and the masses and tailored to what the school wants, what the teacher wants and what the student needs. So I think this has been one of our pillars and is going to be one of the main. Yeah, 

[00:32:44] Alex Sarlin: I mean, already it's just takes a moment to look at nobody a content to say, Oh, this looks very, very different than traditional education content.

And then, as you say, I mean, there's a number of different game formats. There's lots of it can create these sort of. animated videos with voiceover where all the images are created by AI, but they're all represent different. It's almost like a animated slideshow and movies. It can just do so many different things.

And then as you say, it's enabled to continue to evolve very quickly in alignment with other things. AI itself with the underlying models. We know that all the big foundation models are working on video already, and they're all sort of racing to get it out. OpenAI announced Sora a while back. It's still in beta.

Meta has a video one that they're working on. They've all worked on voice, and they're doing better and better voice interactions from both sides. Voice recognition, so you can ask things, and voice generation, so you can make incredibly realistic and compelling voices. They're working on avatar creation.

We've seen, uh, you know, Hagen and Colossian and Synthesia and Prof Jim and all of these other companies sort of say, Oh, well, we can actually create people. We can create avatar teachers that can continue to present content. There's just so many different things happening in the underlying tech space.

And what excites me, you know, one of the many things that excites me about Novo Dia is that, you know, You have all of these different formats of content that can be created. Anytime one of these evolutions, you know, comes out in the world, you're perfectly poised to put it directly into districts and into classrooms.

And that, that speed of translation from, you know, high tech You know, AI evolution to actually making it into the classroom games. I mean, really involved games are going to come very soon or music videos, you know, where you literally the song and the whole video is generated by AI, but you can make it about anything you'd like, including educational content.

The mind is just reels with all the things that are sort of right around the corner. And nobody is just right there, sort of right there on the cusp of it, able to make this and put it right in. So, I mean, as somebody who's been, you know, looking for precision learning, personalization, you know, pedagogical rigor for a long time.

It's one of the tools that just makes me most, most, most excited about AI in the classroom. So I'd love to hear how you think about this fast moving AI world. You know, already you use a lot of cutting edge AI in the platform and you're able to create, you know, whole curricula with lots of super engaging interactive material very, very quickly.

But as you look ahead, what formats and types of educational material are you most excited about as the field continues to evolve? 

[00:35:22] Yair Shapira: Yeah, I think that the secret in what Novodea does, because the multimedia is something that happens in AI regardless of AI not only for education. It happens in AI for everything and anything.

And we are seeing it. It's amazing that we are seeing it as it moves. I mean, it's like seeing a child grow. But I think the secret in what we do is that we have a backbone that is very solid, pedagogical. And then the translation into multimedia is at the click of a button. We have the information, the pages of the book that we create in a very structured way, in a very hyper connected way.

Think Wikipedia. Everything is mapped in our backbone like Wikipedia. So on a textual basis, we know everything very precisely, and we know how to convert it to a different standard and how to convert it for different students. Then when it's manifested as an activity, as a lesson, it can be manifested in so many different ways.

That are very easy for us to create the difference between creating a book and creating a video or creating a song or creating a memory game or creating an interactive game is one line of code, right? Because we have the knowledge itself in a very structured taxonomy. You know what it belongs to, you know, what needs to come before and what needs to go after, you know, what happens if they succeed in this or if they do not, because all of that is already structured, just converting it into an activity.

It's no secret. There are so many companies who do that, but not only in education. I think Alex and I, uh, dare to make the correlation here. You remember, in the 1990s, when cell phones started, they started as a wireless version of a wired phone, right? The phone function in my phone now, in my smartphone, is one of the tiniest features.

I hardly use my cell phone as a phone. I think in publishing it will be the same. I think we will start from automating manual publishing. And what Novodea does today is a version, better in many ways, but it's a version of traditional publishing. Fast forward three, four, five years with the tech velocity, I think it will distance itself from the traditional publishing.

It will serve the same need. Teaching students in schools, but I think it will look in a very different way than what we can today imagine. 

[00:37:55] Alex Sarlin: Yeah, I totally agree. And I think that's a great metaphor about the sort of phone, you know, the evolution of cell phones where it starts saying, Oh, well, what if you could, you know, make phone calls on the go?

I mean, my, my father was a very early adapter of cell phones, and he had a little like a briefcase sized phone that would keep in the trunk, you know? Yeah, exactly. And it was like, Well, yes, it would make a call. And if you're stuck on the side of the road, or if there's all these reasons why you might need to make a call in from the go, but the idea of the, you know, when you have our phones in our pockets now, not only are they phones and computers and cameras and recording devices and, you know, GPS, like the expansion is just endless.

And I see a similar metaphor of like, you know, you think of what constitutes educational content today. It's often. Readings, it's worksheets, you know, it's, uh, prompts for assignments, right? I mean, these are sort of just the building blocks of a traditional curriculum. They will continue to exist, but they will be one choice among, you know, dozens.

You know, you'll be able to make apps about a particular, you say, oh, you know what? I'm gonna make an app for my math class this year so that all the kids on their phones can access, you know, Mrs. Johnson's math app, and it has all these things in it, and it'll push notification. Why not? It's increasingly possible, especially because coding is yet another thing that AI is getting better and better at and agentic work as well.

I mean, there's sort of no limits, or could you create Roblox content directly out of a curriculum? Yes, you can, 

[00:39:22] Yair Shapira: you know, 

[00:39:22] Alex Sarlin: no question. 

[00:39:23] Yair Shapira: Could you, 

[00:39:24] Alex Sarlin: there's no limits. 

[00:39:26] Yair Shapira: When I entered the industry about a decade ago, people told me it's impossible. It's very slow. Nothing can change and so on. And I think it has been static until March 2020, but I think since March 2020, I think it's one of the fastest moving industries.

I mean, there's never a dull moment, right? First it was COVID, then it was post COVID, learning loss, then it was Essercliff, then it was AI, and surprisingly to all those naysayers, the heavy systems find their agility, you know. And things change in the middle of the school year. So I think we are at a singularity moment in K 12, which is very surprising to everyone.

[00:40:04] Alex Sarlin: Yeah, I totally agree. And I mean, and coming out of a, you know, this pandemic and learning loss and the sort of catching everybody off guard globally, most people off guard globally about how to even, Continue schooling in that environment, you know, there's been so many end of reduction in investment in it.

I mean, there's so many things that sort of have looked and felt like headwinds in education and education technology and yet the agility. I think it's sort of the necessity is the mother of invention, right? The combination of the need to change and this incredibly powerful set of technologies that are suddenly.

In our grasp is just creating an amazing moment. I know I keep saying this, but like seeing Novo Diaz platform is really one of these moments where you really realize how it can all come together. I mean, there are hundreds of many, very exciting AI education movements and companies right now. Novo Diaz I think stands in very.

Small few where you look at it and you say, Oh, this is clearly where education is going because it just embraces so many of the capabilities of AI. It doesn't sort of apologize for them or hide them or have it be like, you know, a small addition, a chat bot on the side of something or a slight improvement in an existing process.

It says, let's go AI first. You know, what if you could do an educational publisher that is centered with AI? And what could you do? And it turns out you can do a whole lot faster, cheaper, and much, much more engaging. And then, and then the last piece to put together, and I know you're working really hard on it with all your subject matter experts, is pedagogically evolving.

You know, making sure that people often think of engagement as a trade off. You say, oh, we're going to make it more fun, but maybe we won't be quite as, you know, quite as intense or rigorous. But, you know, great educators know that that doesn't have to be a trade off. You can make it more fun And just as rigorous at the same time.

So that'll be my last question. I know we're coming up at the end of our time here. You're doing all sorts of work with all sorts of different kinds of subject matter experts to ensure that the pedagogical rigor, that you're not compromising on learning by introducing all of these different elements and this customization and the engagement.

Tell us about how you're doing and how you plan to continue to do that as Novodea continues to grow. 

[00:42:15] Yair Shapira: Sure. So first of all, everything goes through. through our SMEs. AI is way, way, way behind the scenes, right? It's part of the editorial. By the way, I'm sure that every publisher today has an editorial team that also uses AI.

I'm sure they go to GPT as well. So everything passes by our SMEs and then through the district. Through this process, We learn verbally, we speak with them, and we learn what we didn't do or what the system hasn't perfected. The system learns, because the way that the SMEs interact with the content is through a chatbot within our system.

The system! Oh, the objectives are not defined in the right way. The order is not correct. I don't see this as a next step, but rather this. And so on. And through this, the system itself improves itself. So it continuously improves. Think about it this way. It's an AI system that sees a curriculum as a molecule.

This is the molecule. Built of atoms that eventually boil down to specific activities, even specific images and phrases. Thank you. Very, very long taxonomy, and while building it, because it's a system, because it's a computerized system, it ensures the consistency and the accuracy. There is no way that you'd find what you find often in textbooks, that page 74 is contradictory to page 26.

Because it was built by a different, it was written by a different person, maybe in a different year. Okay, in our case, everything is always built within a system, within a database, everything speaks to everything. So, by definition, it's already more accurate and more pedagogic. And additional pedagogy is injected from the SMEs.

[00:44:04] Alex Sarlin: Yeah, it's a really interesting model. So we've all talked in AI in education about the human in the loop, quote unquote, which, you know, and it feels like you're really, you know, humans are driving the quality bar of the curriculum. And you often have a lot of professor SMEs or, you know, the types of curriculum creators that are very much aligned to the curriculum creators that publishers would use traditionally, but you're super powering them, right?

You're powering them with. The ability to make changes very quickly to have really consistent underlying knowledge graphs and taxonomy and the ability to make the system itself better and better in reaction to their comments. So it's not that they're just it's different than just making a single correction.

It's actually correcting and improving the entire underlying model, which ideally speeds things up even more and makes it even easier to have high quality materials. 

[00:44:55] Yair Shapira: Exactly that. So the system is super powering the SMEs and the SMEs are super powering back. 

[00:45:01] Alex Sarlin: Right. Well put. Well put. Um, it's incredibly exciting.

So just we try to end the podcast. We've been not as consistent recently with two questions. And let's do I know we're over time. So let's do a lightning round of this. I know that you are obviously deep, deep in AI. But what is the most exciting trend that you see coming in the EdTech landscape that you feel like our listeners should keep an eye on?

And ideally, this is something that, you know, you sort of see in particular from your perspective at Novodea. What's coming our way that we should all sort of get our mind around? 

[00:45:31] Yair Shapira: Yes, so I think that two things are happening in AI that we are using in Novodea and are really looking forward. One we touched, which is the multimedia.

The other is a collaborative AI work. People often think of AI systems as a cheap GPT bought maybe with some pre prepared bones and some systems are actually most tools are, but the really heavy systems use a multi agent. Architecture, where every AI atom, every AI element mimics the behavior of an expert that would sit around an editorial table.

One who is a graphic artist, and one who is a lawyer, and one who is a subject matter expert, and another one who is a curriculum designer. And the collaboration between them, As you can imagine, it's by far smoother than what happens between people, but it's getting more and more. The fundamental backbone of AI is becoming better and better and we make use of it and it's really a power multiplier.

[00:46:27] Alex Sarlin: Yeah, that's a great point that multi agent different AI sort of personas, so to speak, working together behind the scenes rather than One really, really interesting. And what is a resource you would recommend? So if anybody wants to go deeper, obviously we'll put Novodea's website on the show notes for this episode.

That's N O V O D I A. But is there a resource that sort of influenced your thinking in building Novodea or in sort of some of the realizations you've had about AI and education that you would recommend for people who want to dive any deeper into the topics? 

[00:46:59] Yair Shapira: So things are moving very fast, so you really have to stay up to date.

Of course, I love Tech Insiders because you really bring a variety of views. I also read GSV's AI and Education newsletter because it gives a very concrete overview of things that happened and things that are happening even before they are out there. There isn't always an opportunity to click and learn more, but it gives you a daily or weekly overview.

Love that. 

[00:47:26] Alex Sarlin: Yeah, absolutely. And actually, yeah, you mentioned a book to me once that I'd love you to name because I think it's really meaningful. You know, there was a book about sort of how to create a new field rather than you sort of continue. I think it's a really interesting one. I have it on my shelf now and I think others might benefit from it as well.

[00:47:43] Yair Shapira: Yeah, I love the book play bigger, how pirates dreamers and innovators create and dominate markets. It's a great book that explains how you build a new category. And how you lead the category, especially when it's a winner takes all situation. 

[00:47:59] Alex Sarlin: Fantastic. So as always, we will put the links to all of these resources in the show notes for the episode.

And yes, the GSB newsletter, Claire Zhao's incredible AI and education newsletter is one of our go tos as well. That's a great one. And we'll put a link to Play Bigger as well. Thank you so much for being here with us on EdTech Insiders. This is Yair Shapira, CEO of NovoDia. Thanks for being here. 

[00:48:22] Yair Shapira: Thank you, Alex. This is wonderful.

[00:48:24] Alex Sarlin: Thanks for listening to this episode of EdTech Insiders. If you like the podcast, remember to rate it and share it with others in the EdTech community. For those who want even more EdTech Insider, subscribe to the free EdTech Insiders newsletter on Substack.

People on this episode