Week in Edtech 7/15/26: NYC Pauses AI Procurement, Anthropic Launches Claude for Teachers, Norway Restricts AI in Schools, and More! Feat. Louise Baigelman of Storyshares & Ashish Bansal of StarSpark.AI

Join hosts Alex Sarlin and Ben Kornell as they discuss the latest developments in education technology, from growing AI skepticism in schools and major policy shifts to Anthropic's education strategy and conversations with founders tackling literacy and AI tutoring.

✨ Episode Highlights:
[00:05:49] New York City pauses AI procurement across its school system
[00:11:28] New study questions AI's impact on teaching quality and student engagement
[00:18:00] Norway introduces strict limits on generative AI for younger students
[00:21:41] Anthropic launches Claude for Teachers with major edtech partnerships
[00:23:27] Why partnerships and AI ecosystems may define the future of edtech
[00:30:52] How AI infrastructure could help smaller edtech companies compete
[00:33:31] Learning Commons and connected platforms point toward a more integrated edtech ecosystem

Plus, special guests:
[00:37:40] Louise Baigelman, Founder & CEO of Storyshares, on expanding adolescent literacy beyond the science of reading
[01:01:02] Ashish Bansal, Founder & CEO of StarSpark.AI, on building AI tutors that promote learning instead of answer-giving

😎 Stay updated with Edtech Insiders! 

Follow us on our podcast, newsletter & LinkedIn here.

🎉 Presenting Sponsor/s:

Innovation in preK to gray learning is powered by exceptional people. For over 15 years, EdTech companies of all sizes and stages have trusted HireEducation to find the talent that drives impact. When specific skills and experiences are mission-critical, HireEducation is a partner that delivers. Offering permanent, fractional, and executive recruitment, HireEducation knows the go-to-market talent you need. Learn more at HireEdu.com.

This season of Edtech Insiders is brought to you by Cooley LLP. Cooley is the go-to law firm for education and edtech innovators, offering industry-informed counsel across the 'pre-K to gray' spectrum. With a multidisciplinary approach and a powerful edtech ecosystem, Cooley helps shape the future of education.

Tuck Advisors was founded by entrepreneurs who built and sold their own companies, frustrated by other m and a firms, they created the one they wished they could have hired but couldn't find. One who understands what matters to founders and whose North Star KPI is the percentage of deals closed. If you're thinking of selling your ed tech company or buying one contact Tuck advisors now.

Every year, K-12 districts and higher ed institutions spend over half a trillion dollars—but most sales teams miss the signals. Starbridge tracks early signs like board minutes, budget drafts, and strategic plans, then helps you turn them into personalized outreach—fast. Win the deal before it hits the RFP stage. That’s how top edtech teams stay ahead.

Transcript

[00:00:00] Alex Sarlin: Innovation in pre-K to grade learning is powered by exceptional people. For over 15 years, edtech companies of all sizes and stages have trusted Higher Education to find the talent that drives impact. When specific skills and experiences are mission-critical, Higher Education is a partner that delivers.

Offering permanent, fractional, and executive recruitment, Higher Education knows the go-to-market talent you need. Learn more at hireedu.com. That's H-I-R-E-E-D-U.com. 

I think you have to be pretty deeply informed to be a critic of one of the edtech tools. Instead, people are just worried about AI, the letters, the concept being in classrooms.

And when you actually look at a lot of the proof points, and this is true for many, many of the-- especially the negative studies, but a lot of the studies generally, they're looking at what happens when students have access to AI, and that's usually ChatGPT because, to be fair, those are the most used

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. 

[00:01:16] Ben Kornell: Remember to subscribe to the pod, check out our newsletter and also our event calendar.

And to go deeper, 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.

Hello, Edtech Insider listeners. We're back, another Week in Edtech. It's Alex Sarlin, Ben Kornell here with you, and We're just having a great summer. Alex, how was your travel? You were in Michigan, is that right? 

[00:01:52] Alex Sarlin: I was in Michigan, yeah. Just got back from Detroit, Ann Arbor. I was there with my family. We got to see the amazing Detroit Zoo.

We got to hang out with some really close friends who live out there, so it was a blast. It was still hot, unfortunately. I was hoping there'd be some wind over the water or whatever it is up there, but it was still in the high 90s, even past 100 a couple days. But Michigan's beautiful. I love the Midwest.

[00:02:15] Ben Kornell: I was back home in Indiana this last week. Yeah. So just down the road from you, yes. Amazingly hot, it's a furnace. And got to visit the parents, and got to be in a red state where all the dialogue about education is choice, and what kind of voucher should we have, and Trump tax credit, and all of those things.

So definitely a chance to step out of California vibes for a bit. 

[00:02:38] Alex Sarlin: Yep. 

[00:02:38] Ben Kornell: So before we jump into all of our topics this week, what do we have going on with the pod? 

[00:02:42] Alex Sarlin: Yeah. So stay tuned. Both guests on this episode are totally amazing. We're speaking with Louise Baigelman from Storyshares, and Storyshares does this really interesting leveled reading for adolescents who are off grade level.

They've been sort of out front in adolescent reading remediation basically for many years, and they've now expanded into a full platform. It's a really interesting conversation. And then Ashish Bansal, who is a incredible founder. He's been at, I think, four of the FAANG companies, and he and his co-founder have been at every FAANG company literally.

He's worked at Amazon and like all of them, and they've built five of their own models to be able to create incredibly powerful tutors with their company StarSpark AI. So it's been amazing. And then I've just had some amazing conversations just that are coming out soon. Nir Zicherman, who has a company called Oboe , backed by Andreessen Horowitz.

Really, really interesting. They h- their goal is to hit a billion users through an edtech platform. Just talked to Vinitra Swamy, uh, Scholé AI. They've built this whole complex system with four agents that work together to create the pedagogy. It's been very, very inspiring to talk to so many amazing founders and all sorts of incredible people.

Dan Wang from Columbia, they're doing a voice discussion agent that works at higher ed. So stay tuned for all of these episodes and more. It's just, it's been so fun jumping back in and doing these great founder interviews because, God, the quality of the people, the deep thought about how are these products being developed, it's really inspiring.

Tell us about events. Where are you at? 

[00:04:11] Ben Kornell: I mean, the other thing too is how fast these things are going to market. The most recent event was ISTE, and there were a ton of teacher entrepreneurs on-site at ISTE pitching just with their laptop to their fellow colleagues. It makes me think about the entrepreneurial energy that teachers already have.

This beg, borrow, and steal has always been how we as teachers got our best curriculum and content, and now we have teacher creators building some pretty incredible alpha and beta products. And from an event standpoint, next week we have the Bridges AI and Government Conference. There's different tracks for different things, and there's an education track.

A big focus on this is around buyers and how do you purchase AI-enabled products. One, how do you understand the efficacy? Two, how do you understand the trajectory of where a product is going? When you make the purchasing decision, six months later, it's a totally different product. And then the third is really around implementation and long-term co-development and partnership.

So I'm speaking there. Come check out our panel. We're doing State of edtech. John Katzman's gonna be on it. Amit from El Ventures is gonna be on it, and Jen Womble, the kind of sales extraordinaire in education. Yeah. And my son Nico is also gonna be on a panel around community voice in edtech and AI and govtech in general.

So that's coming up. Hope to see you there in New York City Without further ado, let's talk a little bit about New York. Maybe we start with the backlash and then we go to Anthropic. Does that sound good? 

[00:05:47] Alex Sarlin: Yeah. Yeah, let's do it. 

[00:05:49] Ben Kornell: So New York kind of made headlines with pausing all AI purchases, which basically is like all edtech purchases, period, full stop.

There's a couple other headlines that have to do with the backlash, but really this New York City pause is rippling out where a bunch of other purchasing processes are really, really slowing down. And this means not just at the district level, but even at the site and classroom level, sales are tough, and we're heading into a really tough season.

What's your take on this one, in particular as a New Yorker? 

[00:06:23] Alex Sarlin: Yeah. I mean, there's a new mayoral administration in New York, as we know, over just nine months or whatever it's been, not very long. Six months, I think. And the new chancellor of schools who came in with that administration, I think is caught between a little bit of a rock and a hard place here.

I mean, I think he's trying to deal with the zigzagging politics of this, with parent backlash, with educator backlash, with vendor enthusiasm. Like, I think he is basically calling this pause as a way of trying to step back and say, "We need to really think harder about our AI stance." And I don't think this falls in the same category as the ban that New York did when, when AI first came out.

This is not a decision to ban it, but it's a sort of like, "Everybody take a breath. This is getting very heated," and it's involving a lot of different stakeholders, and I think he's saying, "Okay, let's stop." Now, we in the edtech field know that just saying, "Hey, let's pause until we have policy in place," is not a good thing for the field because it means that they're pausing purchases, they're not letting contracts go through.

I'm sure that it's affecting renewals. edtech companies and, of course, the New York DOE is the biggest system in the country, so, and one that a lot of edtech companies have used to pilot, have worked with to expand in lots of different ways. So it is definitely not good news for the edtech field. I don't think it's permanent at all.

I think it's a little bit of a everybody take a breath moment. That doesn't mean it's good, and it doesn't mean it'll necessarily fall in a pro-AI position, which I think would be the right way. But yeah, I mean, it's a sign of the zeitgeist backlash having real-world effects in a pretty real way. We saw LA go back and forth on screen time a couple of weeks ago, and now New York is in the spotlight and trying to figure out what to do.

What did you make of it? 

[00:08:01] Ben Kornell: Well, I mean, we covered Randi Weingarten's flip-flop on AI. Right. She famously was a leader with an open AI teacher coalition, and now she's saying, "Whoa, I thought this was a bad idea the entire time." So in that sense, from a political and social standpoint, it's not super surprising.

I think the observation that I have is that when there's uncertainty, things slow down. And- 

[00:08:27] Alex Sarlin: Yep ... 

[00:08:27] Ben Kornell: whether you think AI is good or whether you think AI is bad or somewhere in between, you don't get the clarity and you don't get the movement. All the works just slow down. And that seems particularly challenging in an era where the rate of change is so fast.

With schools slowing down to be more deliberate, are they going to get left behind? But I think the reason why this is probably an astute move for the chancellor, the new chancellor and the new administration, is it seems like the parent side of the backlash is also saying, "Hey, slow down. We wanna be more thoughtful."

So in this sense, I'm generally not a fan of schools slowing their roll when it comes to being innovative and transforming learning. But in this sense, if they had pushed it forward too far too fast, they were gonna lose parent support, and they were gonna ultimately lose teacher and student support. And maybe by doing this, this allows them to be much clearer about a framework of what's good AI-backed educational tooling and what's not.

[00:09:32] Alex Sarlin: Yeah. 

[00:09:33] Ben Kornell: The one worry that I have with this is there are infrastructure tools that do use AI, and things like Google Classroom, which are abundant in the district, that are already live and in use. So I get that it's a pause on procuring new things, but they should also realize, like Basically, most of everything you're using is already AI-enabled in the first place.

And so that's why it's a little bit of a conundrum. 

[00:09:56] Alex Sarlin: Agreed. And I mean, I think that it's gonna be interesting to see where they land because they do say that at the end of the summer there should be new AI guidance. There was a lot of pushback on the AI guidance they had in March, and I, I can't help feeling, and I know I, I'm a little bit of a broken record on this, but I can't help feeling like the edtech sector has not, as a community, no finger-pointing at all, but has not been able to come up with a cohesive, coherent response to this parent backlash, to this teacher backlash.

I mean, people are saying, "I see my kids and it feels like they're cheating. It feels like they're using AI to skip steps. I see my kids on the phone and I get worried that they're using tech too much. I think that AI is maybe collecting data on them," or various things. Like, there's a lot of inroads to the backlash.

There's a lot. And then the other side of the equation is still so hand-wavy where it just has not been a cohesive counter-narrative about why AI in all of these tools is really successful. There's a lot of great stuff that's theoretically possible in there, but it's not as clear as it should be. And look, I talk to all these founders who have these unbelievably exciting products.

They're thinking so hard about all the security. They're thinking so hard about a pedagogy. Like, they're thinking really hard about it, but none of that is getting out into the world and being able to counterbalance this incredibly clear, loud backlash that I just think AI and tech and all of these, the tech lash in general, edtech has just been swept away with it so far.

And I think it's really on all of us to be very clear about the difference between tech and 

[00:11:28] Ben Kornell: EdTech. Well, and also just to double down on that, it was student use cases of AI was where the majority of the concern was. Sure. Now there's questions around teacher use of AI. Yeah. A study that we got for this week in the Hechinger Report out of University of Pennsylvania- That's a good one

shows that generative AI can harm teaching. Now, I wanna caveat the-- So it's like a pre-publication in June from Professor Altungul, who has been a constant critic of AI in education And it hasn't been published in a journal which would validate the research methodology and practices. And it's about a 2025 study in Turkey with 2,800 middle and high school students and 193 teachers.

So I think a challenge I'm seeing too is like, which research should we pay attention to? Right. Justin Reich just posted on, on LinkedIn about research that we'd all been using as source of truth, and then a grad student went in and checked all the citations and was like, "Wait, this meta-analysis actually mischaracterized many of the studies that were part of it."

So- Right ... we're all being a little bit more careful about, okay, when we see research back, like what does that really mean? But Angela Duckworth is connected with this, of grit fame. Yep. So has a lot of weight. And basically what it claims is that student engagement goes down when teachers use more AI for efficiency.

And the efficiency value proposition has probably been the lead value proposition in the edtech space, especially on the teacher side. And 

[00:13:06] Alex Sarlin: the cleanest one for sure. 

[00:13:08] Ben Kornell: Yeah, the cleanest one, and okay, we get it, you're concerned about kids using AI and like cognitive offloading and so on. But with teachers, it saves time.

Surely that's good. And we wrote an article in Edtech Insiders about, well, we're saving time for teachers. What are the teachers doing with that time? Now, this study basically says even the efficiency itself just creates like more bland, less engaging worksheets, lesson plans, assignments, and they showed that especially the worst teachers fared the worst with the AI.

So one would imagine another theory of change is that, well, the best teachers should keep doing what they're doing, but the worst teachers should use AI, and that will raise the floor here. The study shows the exact opposite, which is basically like the best teachers are creative and they use the AI in creative ways.

The worst teachers, it just gets so mechanical that the class engagement and the class experience is not great. So I don't know, like- Yeah ... I don't wanna extrapolate too much, but if we get to a place where the consensus is that AI is problematic with kids and the AI efficiency claim is problematic with educators-

then we're really back to starting over. 

[00:14:20] Alex Sarlin: Yeah. I agree. It's also, I think, worth calling out that the actual tool that they used in this study is, they said, "A ChatGPT-based teaching assistant customized to Turkey's national curriculum," right? I haven't looked into exactly what that means. It may be an edtech company, it may just be a adapted GPT tool.

But I think it's worth noting that most of the actual backlash against AI in education still continues to be against the frontier models, the off-the-shelf general purpose tools. I think, again, edtech is getting caught up in it. There are definitely critics of Magic School. There are definitely critics of School AI or Brisk, but there aren't very many.

I think you have to be pretty deeply informed to be a critic of one of the edtech tools. Instead, people are just worried about AI, the letters, the concept being in classrooms. And when you actually look at a lot of the proof points, and this is true for many, many of the, especially the negative studies, but a lot of the studies generally, they're looking at what happens when students have access to AI, and that's usually ChatGPT, because to be fair, those are the most used.

There was a really interesting report out of Securly. Securly monitors student usage and technology use across thousands of schools, and it basically said that 60% of usage in classrooms was Gemini and ChatGPT, and that Magic School, Brisk, and School AI all combined were less than 10%. So I think in a lot of ways, there's the potential for the edtech world to say, "Yes, off-the-shelf general purpose AI tools really aren't very good for education."

They do a lot of things that you would never do as a teacher or tutor. And to your point, Ben, and the point of this paper, even the teacher efficiency, if you're using something that's basically a wrapper or if it's just straight one of the frontier models, it's not really designed with education as a core purpose.

But edtech is. Like edtech could be framed as the solution to this problem. Instead, it's being lumped in as part of the problem, and, ah, it's driving me crazy. 

[00:16:09] Ben Kornell: Yeah. There's three layers here. It's like the public perception layer, the actual efficacy of these products, and then underneath it is like business model questions.

And one thing that I will just note is that- Business model questions lead people to these product-led growth motions where getting quick wins on efficiency is really helpful in getting market traction. But actually getting efficacy means changing pedagogy and practice. It means improving instructional quality- Oh, yeah

which it requires the user, whether it's a teacher or a student, to have much more cognitive friction and do things that are more difficult. So we've created like a challenge where the perception of these things being low quality or harmful, plus the growth motion, which is find low-hanging fruit wins that don't require a lot of user change, means that the middle gets sandwiched, which is really what all of us who are in this space are about, transforming the learning outcomes for kids and students of all ages.

And this is where I think rightfully we've got some like core conflicts. Beth Havinga in Europe- Yeah ... she's been really, really diligent about bringing the edtech community together. She just published a book around edtech. It's in German, so I haven't read it, but they've really thought through how to build a consortium, and what they have is an advantage in that most of the buyers are ministries of education, where they have one single sourcing group- Right

that really becomes expert on the efficacy and the research and so on. So I think in our kind of Wild West fragmented purchasing market, it makes it really hard to have a singular voice and singular channel. 

[00:18:00] Alex Sarlin: No question. And, I mean, speaking of backlashes and national choices, we also saw Norway this week put basically a almost full ban on generative AI for younger students and a lot of guardrails around teacher use.

So basically, at the country level, you have these ministries of education basically deciding how much to buy into the AI positivity or the AI negativity, and they're trying to figure out exactly where it goes. One really interesting article that I recommend people look into this week was this article, "What If It's Not the Phones?"

from The Atlantic. And I mean, it's kind of a wacky headline, but what it's really about is a alternative theory. Basically, it's a social psychologist named Peter Gray, who's a colleague of Jonathan Haidt- And he is somebody who is basically, he has a book coming out really soon that basically says it is actually not the screen time and the phones that are responsible for the negative educational outcomes and the depression and so many of the things you're seeing.

He actually makes sort of an alternative claim that online engagement and the ability for kids to connect with each other through gaming or social media platforms actually is the place that preserves connectivity and play in a society that's increasingly isolating, and where many of the traditional places where students or people would collaborate and come together are being taken away.

And he has a totally different theory and model about why this is happening. Now, whether this will pick up and catch on, I have no idea, but it's just interesting to hear a sort of anti-Haidt narrative, because obviously Jonathan Haidt and Jean Twenge's narrative about phones and screen time have just undermined everything, has just been-- It's, at this point, I feel like it's become almost like universally accepted as an explanatory mechanism, and there's at least an alternative coming out that's like explicitly going in another direction.

So interesting article. We, you know, we'll check in in six months to see if people are talking about that book or if this is just a really sh- quick flash in the pan idea, but worth looking at because again, I think it's not just about the narrative. I mean, we need real evidence, but I just think we don't have a really coherent counter-narrative at all.

Even the companies that are putting their money where their mouth is and funding third-party independent studies. You know, we've talked a lot about Eedi. A lot of people have been really trying to figure this out, and a lot of edtech founders are going wildly out of their way to do things that are almost the exact opposite of what the frontier model off-the-shelf thing tools do.

So like the studies show that have access to GPT, and I just looked it up by the way, you know, the Duckworth, the study we were just talking about, it was a custom-built AI assistant, which is basically ChatGPT trained probably with like an artifact trained on the Turkish national curricula. So they're literally saying, "If you give teachers ChatGPT- It doesn't make the lesson plans more effective.

That's also exactly what every edtech company has been saying for years. That's why they're creating alternative companies. I don't know. I'll stop ranting about it, but it's just, I feel like this is a nuanced argument in a world that is hard to reach nuance, but I really encourage everybody out there listening to this in the edtech world, if you're in sales, if you're thinking about renewals and contracts, if you're thinking about how to frame it, say: "Yes, off-the-shelf version of the frontier models have shown that they are not effective at teaching.

Therefore, the edtech world has created all of these powerful alternatives, many of which are getting evidence of their effectiveness, and that's where we've got to turn." Anyway, we should talk Anthropic because that's obviously the biggest, I think, edtech news of the week. Do you have anything more you want to add?

I, I know that Randi Weingarten came out saying Newark public schools were part of the reason for her transition that she actually made. But, yeah, there's all sorts of stuff happening, but I don't know. What do you think, Ben? 

[00:21:27] Ben Kornell: I think this is the backdrop in which Anthropic made their big announcement. Yes, exactly.

Why don't you walk a little bit through it? I, I'm happy to kind of pipe in on what I think is game-changing here, but it's a b- major moment. 

[00:21:41] Alex Sarlin: So we had seen, you know, ChatGPT for Teachers launch over a year ago. We obviously have seen Google do many, many, many different launches of all sorts of different kinds for-- with Google Classroom and Gemini and Learn LM and all these things.

Anthropic, because it's been 18 and over, it has been involved, but usually behind the scenes in terms of edtech. It's been the underlying model behind many very successful edtech tools, including Magic School. If you've gone to the Anthropic site, you see lots of-- You see their education use cases, but they're always sort of through proxies, through edtech companies that they support.

They launched this week in a very visible way, Claude for Teachers, offering educators the ability to get paid Claude Pro, you know, services for free if they verify that they're a teacher. But also introducing a very deep connection with Learning Commons and their underlying infrastructure, and we've talked to Sandra and their team a lot on this show, so we should talk about this, and all of these really interesting edtech connectors with a bunch of edtech companies which will be very familiar to the listeners of this podcast.

Brisk Teaching, CoTeach, Diffit, Eedi, Magic School, Snorkel, TeachFX. I think we've had almost every founder of those on here. As well as Assistments, which is one of the most evidence-based edtech tools in the world by Neel Heffernan out of Worcester Polytechnic. And Canva Education, obviously an interesting crossover that is both educational and design-centered.

They have Teach for America building things. They have AFT involved. They have PlayLab involved. So it is a really interesting big push through the whole industry to say, "Hey, there's actually a really interesting role for edtech in classrooms and for Anthropic to actually be the core underlying model behind it."

But- Phil, and I'm sure I missed a lot of things in that description. What, what, what stood out to you? 

[00:23:27] Ben Kornell: I think the three things that I would take away is, one, Anthropic, Google, and OpenAI are gonna continue to play in the education space, and companies that are built with minimal competitive advantage are at risk for mainstream LLMs coming in and adding solution sets and capabilities.

The reason why that comes to pass is point number two, which is Anthropic is making a big partnership play here. Yes. And the partnership play, the last mile is the hard part in education. There's a, like, competitive advantage for all of these companies to be the product partner with Anthropic, even if it might undermine some of their, like, technical differentiation.

Being part of that platform or system gives you channel and delivery to way more users, gives you a lot of credibility, and also allows you to be that last mile winner in edtech. So the third one is we're moving to actually live out this model where our stack includes the kind of chips and AI and NVIDIA servers.

It includes cloud and so on, includes LLMs. But before you get to the app layer, there is another layer of AI enablement, the agentic layer, I think, that will basically define how all sectors engage with AI. And the reason why I think that the LLM layer has gotten commoditized already, I've been one-shotting with Fable, and I mean, it's incredible what's happening.

But you can imagine a year or two from now, our human regular user ability to differentiate between what the best is and the third best, it's gonna be minimal, and it's gonna come- Yeah ... down to cost. That layer is tough. So that's why all these companies are moving into agents, because that's where the actual meat of doing things with AI happens.

And so this idea that, like, creating a lesson plan is going to be so differentiated, well, it was already hard to differentiate. Now the LLM is gonna basically take care of that full stack, because creating a plan or creating content, that's like a core workflow for anything. Now, the differentiation will be, is that integrated with your content and curriculum, which- 

[00:25:49] Alex Sarlin: Right

[00:25:50] Ben Kornell: it lives on the app layer. So number one is this is here to stay. These people are all gonna be playing in the education space, like, be prepared to have your business disrupted. Two, like, last mile is where the value add is for most edtechs, which means that they're gonna be small businesses. I think the TAM of pure edtech, that's a big LLM- gonna be smaller.

And then third is the differentiating element is the integration with everything. It's both distribution, implementation, and integration. It's all those things. Yeah. But I will say this is probably the best received announcement because they came out with a partnership play. Learning Commons is also, I think we're gonna do something with Learning Commons to explain more about it.

Yeah. But Learning Commons is like embedding all of their open source tooling in there. So it's like really great infrastructure that's all wrapped up in a bow into this product, um, this product suite. And so I kind of, the educator innovator in me loves that this is happening. I just think it's gonna be a really tough time now to make the claim that your business is defensible given how well- Yeah

these players are coming into the space. 

[00:26:59] Alex Sarlin: Yeah, I agree with that. Although I am also of the mind that it is exciting to see companies that are, that are actually quite small, like the Snorkels and CoTeaches and, and Eedis here, working directly with Anthropic, being offered to basically every teacher in the country.

I mean, talk about a distribution partner. That, it's like pretty incredible. But I think you're right, that probably means that Snorkel by itself is not gonna be the next unicorn . I don't think that it ever was. But it's a great product, and it's created by two teachers, and it's designed for, w- to be really usable, and it's a really, really cool product.

I love Snorkel, so it makes me very happy to know that they're gonna be sort of integrated into a system and service that's gonna be used this widely. You mentioned the partnership play, and the only thing I would add to that is that you remember, you know, I don't know may- I don't know when this was, probably 18 months ago now, but OpenAI had this specific kind of partnership play very early on.

They said, "Okay, Khan Academy is our K-12 partner. ASU is our university partner." And those were good choices, right? Those are both very trusted brands. They're both very well-established groups. Very, very smart public figures behind them. But compare that strategy, that Khan and ASU, to the list that we just named.

I mean, Assistments, CoTeach, PlayLab, AFT on privacy and safety. They're gonna align to AFT's privacy practices. They're gonna align to Teach for America's fluency courses. They're gonna align to, they're doing a whole big pilot with Detroit to study that, whether educators are using this really well.

Anthropic is still for ages 18 and older, right? So this is a way to enter the school space that's incredibly It's not only partnership driven, it's almost like bringing a whole community with you. And then Learning Commons, of course, you know, I won't go into the details of what Learning Commons is doing because I think that most of the people on this, uh, listen to this podcast probably know the heart of it.

But they're able to say on day one, "We have all 50 state standards baked into this," because Learning Commons has done all the work to make that happen. We have all of these learning components, you know, all of the standards broken down into these tiny learning components so that you can ask for any particular thing from any state lens.

This is exactly the infrastructure work that Learning Commons has been working at, and Learning Commons had already been partnered with Anthropic in sort of a minor way in that Anthropic had, had been, had installed their, their module, their RAG model inside it, and you could access it directly from Anthropic.

But this is a buy-in to basically the entire philosophy of the edtech ecosystem at this moment around AI, which is if you get all the players together and people are working on each of their specific strengths and, and everybody's sort of combining forces, you can do something, like, much bigger than the sum of its parts.

Or at least that's the part of the AI ecosystem I really love. I hate the competitive part, I love the collaborative part, and they're really going hard at it. I think that's incredibly exciting. I mean, you can tell that we know Drew Bend, who's been at Anthropic for Education, go- moving up in the world for quite a while.

He's an ex-con person, ex-Khan Academy person. 

[00:30:01] Ben Kornell: Con, not 

[00:30:02] Alex Sarlin: ex-con. Yeah. They should call all their, their alumni ex-cons. That's- Ex-cons. But he knows the ecosystem. Let's just put it that way. He knows it, and the, the education people. He mentioned on LinkedIn that this was an onboarding project for their new education lead.

Can you imagine? Like, I mean, they are playing three-dimensional checkers or whatever, right? Three-dimensional chess when it comes to how to enter the edtech space in a really intelligent way. Whether or not teachers will buy into it, TBD, right? Some people might look at Anthropic and say, "Big tech trying to get our data, trying to get people to upload student data into unsafe places."

They already see a little bit of that, that just knee-jerk skepticism. But for people who follow this space, this is a very sophisticated go-to-market for a big tech company. They know the players. 

[00:30:52] Ben Kornell: Yeah. What is interesting to me is that this could also allow educational companies to focus on really what they need to focus on, which is differentiating the pedagogy, the product, the efficacy, and the context, and really allow for the technical expertise to live outside.

Um, the Learning Commons- Exactly ... infrastructure tools also I think are very exciting because it allows connections to assessment, data sets, learning graphs, knowledge graphs- Yeah ... and other tooling that, you know, individually any individual edtech company would not have access to. But being able to kind of build on top of that IQ is going to be really, really exciting accelerant.

I will say one thing that I'm thinking about is like, are micro companies going to win? Are large companies that have distribution you can fast follow, are they gonna win? Will big tech win? Like, how will this all play out? I think the micro company opportunity continues to be very compelling if you can build your like thin layer of like real value and expertise In consortium with this stack that they've created for you, it makes me think that a 10-person company really could scale.

One other thing before we go, our partners, so full disclosure, Starbridge is a partner of Edtech Insiders, and I'm speaking at the-- their conference next week. Ahead of the conference, they've announced a brand-new all-in-one distribution channel, essentially a public sector channel partner program where you sign up for Starbridge and they are your SDRs.

They make sure that you're in all the procurement platforms. They make sure that you're compliant on all the lists, and it's all using AI and kind of matchmaking based on their intel of who's looking for what things to purchase. That's another one that makes me think that this idea that distribution is the only advantage and that fast follow is going to win, I think that that falls away when you have these kind of infrastructures or distribution channels, in this case, where a small company doesn't have to have the sales force of, like, 50 people to be a national player.

They could use these channels and just have a really incredible product. And especially if, going back to our original theme, it's really about efficacy and research. You put all that together, and it could come out to be a golden era for building AI edtech. 

[00:33:31] Alex Sarlin: Yeah. And for putting together the pieces, right?

I mean, what's really interesting about this, oh, you know, we didn't even mention the HQIM, you know, because Learning Commons is already aligned to some of the high-quality instructional material curriculum, like Illustrative Math and OpenSciEd. You're starting to see a solution that isn't a point solution, that is actually an ecosystem of solutions that work together, right?

If you adopt a system like this, or if you're working with a Starbridge and you're looking to forgot-- government sales and you're looking to get, uh, verified for a whole state or something, right? You can come in and be part of suite of companies that actually all work together and integrate together through model context protocols.

I think that's a very exciting step in the right direction for edtech because it's such a fragmented market, right? You have all these people doing these very specific things, and instead of it be-- Like, I mean, TeachFX is a good example, right? TeachFX, such an interesting platform, uses-- gives feedback to teachers on everything that happens in the classroom.

It's really intelligent feedback that they can use as professional development, measures talk time, measures talk time by gender. I mean, just really, really interesting things, but it's such a specific solution. But the idea of being like I can use TeachFX. It's built in within Anthropic. It works with the curricula I already use.

I can tie it to my curricula. It already has my standards that I can connect to so I can... It starts to feel like a real solution. And the, this training that's i- involved in it, and because it's privacy vetted, we know that TeachFX already meets the privacy standards. Like, you're just starting to see, and Playlab will come in and teach us how to use it.

Like, it is really pretty cool to start seeing all of these really, really innovative edtech companies and founders sort of part of a team, rather than being all off on their own trying to beat down the door and explain what their specific edtech product does. So I agree with your point that smaller companies with really interesting distribution platforms can be really powerful, but I also think that ecosystem plays, the idea of a school being able to tie into an existing system is really powerful.

And we've seen that, you see that in finance, you see that in health, you see that in retail. Like, people don't usually come in and buy 100 different point solutions and try to weave them together, right? They get a system that works together, that already talks to, where everything already talks to each other, and we- we've never really had that fully working in edtech.

We've tried LTI and oh goodness, all sorts of dif- different things to standardize to various success, but that's what gets me excited about this. And who knows? You know, it's easy to be optimistic. This thing just came out. Nobody is brand new, but I, I like it for that talk. And I also, by the way, really like what Gemini and Google Classroom have been doing in that same vein, right?

They've been working with lots of partners, they've been working with lots of collaborators, and they've been trying to build an ecosystem. Build Google Classroom and Chromebooks into an ecosystem that is not just a Google product, it's a whole infrastructure. Uh, Instructure has done a little bit of that too.

I really, really like those moves. I feel like they're, they- 

[00:36:22] Ben Kornell: Yeah ... 

[00:36:22] Alex Sarlin: are promising for the space. 

[00:36:24] Ben Kornell: The work is too important, the impact is too important for us to all be doing redundant tasks- Yeah ... and like working at odds. So one, from an impact standpoint, collaboration wins. But I will also say Strategically as a business, collaboration wins too.

Because if you're trying to compete against these big tech players and crowd them out of our space, you're being shortsighted around where it's all headed. Yeah. And so, you know, to see Magic School and Brisk on there, there's potential disintermediation of some of their tooling and capabilities, but they know that the long-term value of being a collaborator and player where all these things connect together is so big that- 

[00:37:07] Alex Sarlin: Yeah

[00:37:07] Ben Kornell: I think we're biased Edtech Insiders is all about community and collaboration. Yeah. But that seems like one of the big takeaways, and it's also a great way for us to wrap up. We have two great guests coming, so please don't change that dial. Alex, you wanna take us out this time? 

[00:37:22] Alex Sarlin: Sure. If it happens in edtech, you'll hear about it on Edtech Insiders.

And remember, that subscriber community was abuzz about this Anthropic just- 

[00:37:32] Ben Kornell: Yeah. Become a member today. Join our WhatsApp. We'd love to have you in the conversation, and there's many more announcements to come. 

[00:37:39] Alex Sarlin: Absolutely. 

[00:37:40] Ben Kornell: Thanks everyone for joining. On to our guests.

Hello, Edtech Insiders. We have a special guest today, Louise Baigelman, founder and CEO of Storyshares.

Storyshares is a public benefit company dedicated to transforming adolescent literacy. Louise began her career as a middle school reading teacher at KIPP Lynn. I was also a middle school teacher, recovering middle school teacher. Louise earned her BA in English and Psychology from Cornell University.

Spelled the wrong way, but hey, go Big Red, and her master's in education from BU. Welcome to Edtech Insiders, Louise. 

[00:38:17] Louise Baigelman: Thank you. Great to be here. 

[00:38:19] Ben Kornell: So before we dive in, just tell us a little bit about your journey to start Storyshares. 

[00:38:24] Louise Baigelman: Yeah. So I started as a teacher. I graduated from college and did Teach For America, and I moved to the Boston area and was placed in a KIPP school in Lynn, north of Boston, and as a middle school teacher, focused on reading, writing, but also ESL, which is what we called it back then, supporting multilingual learners.

And I was a fresh teacher, and I entered my fifth grade classroom expecting to teach complex literary analysis, and I found that most of my students couldn't really read. And I was startled and horrified and obsessed with that piece of the puzzle because I watched how it impacted their ability to succeed in math and science, and I knew that it was going to impact everything that they needed to be able to do learning-wise, but also just in a daily life situation.

So tried to find resources to really win my students back to reading and give them the skills they need to succeed, but couldn't find very many. And so I always joke that I am an accidental entrepreneur because I was a teacher, and I was in the field of education, and I started Storyshares because I couldn't Stop thinking about the fact that we had to offer pathways for older kids to become readers.

[00:39:49] Ben Kornell: I mean, I love it. I resonate with it so much when you hear about science of reading and people engaging, deep diving in innovation and reading. It's really focused on early childhood up through third grade. Part of that, I think, is because there's a real ROI if you can solve reading challenges in that age.

But what inspired you to focus on 3 through 12, and specifically adolescent reading, and what systems do we need to help make that successful? And just for our average listener, they understand these challenges, but it's not like you can take a middle schooler who's reading at a third-grade reading level and give them a third-grade book and expect magic to happen.

So both in terms of the need and the solutions, what do you think is different about adolescent learning and reading? 

[00:40:40] Louise Baigelman: Yeah. Well, it's so interesting and powerful and endlessly complex, literacy in and of itself. But also what happened with the science of reading movement has been so powerful. So I was a teacher almost two decades ago, and that was well before science of reading, right?

And the kids were struggling with reading in middle school then. It was also before COVID. It was even before, like, smartphones were big in middle school. That was already a reality. It has expanded and morphed, but also science of reading entered the space, and it's such a beautiful, powerful thing what the science of reading movement is doing, which is identifying-- it's actually not one single philosophy or approach, it's just we want really evidence-based, research-based approaches to teaching reading.

So there's so much good in it. And what it uncovered, though, was a really challenging truth, that we have not been teaching reading correctly for decades, right? And so decades' worth of students had not been learning to read correctly. What's fascinating to me about the science of reading movement is everyone concluded in response to that truth that we therefore must make sure all K-3 students we see from here on out learn to read, and that will be the solution.

But the reality is that means that we've been passing along decades' worth of fourth through 12th graders ever since, right? And so the science of reading movement has to be expanded up to all the fourth through 12th graders today who were taught with approaches that didn't work, and therefore, and we know the data shows this, 65 to 70% of them aren't reading proficiently.

And so to me, the only thing I can think of as far as, like, why is that happening is, A, literacy is typically a K-3-- it's an early elementary thing. Mm-hmm. But B, we don't have the structures in place. And so to that question- How do schools have to look at it differently? K three schools are built for that's when students learn to read.

The teachers are trained in foundational literacy instruction. The schedules have ample time for reading. 

[00:42:45] Ben Kornell: But when- Yeah, it's classic learn to read and then read to learn, and then what happens is if you didn't learn to read, your read to learn era, so you're dealing with compounding gap there by the time you hit middle school.

[00:42:56] Louise Baigelman: Well, exactly. I mean, that's a great way to sum it up. Essentially, our system right now expects that you learn to read and now you can read to learn. And the reality is most of those kids, more than half, didn't learn to read, and we now want them to read to learn. And so everything devolves. And so there are structures which I could talk about on the positive side that will enable it, but we do have to think about it differently because you cannot take K three literacy solutions and give them to an 11th grader and expect them to work.

[00:43:25] Ben Kornell: And often remediation or intervention has like this negative stigma, but it seems like Story Shares is reframing that approach. And I also noticed that in addition to curriculum and learning materials, you have a big emphasis on teacher professional development. Can you tell me about the shift away from like an intervention negative stigmatized approach towards more of a learning opportunity approach and maybe a little bit of the role of educators in that?

[00:43:56] Louise Baigelman: Yeah. So I think one of the really challenging parts in the secondary space is that there is this false dichotomy. And we talk about reading wars, which is also a problematic term in my mind. But similarly, in secondary right now, rigor is at war with access. And so when you say remediation intervention, I like the word access because I think when we frame it that way, you recognize that access is required for rigor.

It is not preventing rigor. And I think though because like we want high expectations, right, and rigorous grade level knowledge and skills, we all want that, that there it's become a divide which ends up becoming pathways rather than a zoom out, which this requires because it's new stuff, right? And how do we enable access to rigor?

And how do we actually look at all the ways that literacy inherently plays in to the grade level more complex knowledge, skills, and texts and reinforce them both together? And so I know this is a short podcast. I won't 

[00:45:03] Alex Sarlin: It's 

[00:45:03] Ben Kornell: almost like concentric circles where if you start with the access and radiate out to more rigor and relevance, you've got this opportunity to be expansive in the learning zone.

But if you don't have that core access, you really can't get to any of those outer 

[00:45:20] Louise Baigelman: circles. Right. Exactly. Rigor is a lofty ideal that is impossible without access, right? And so rigor is a totally achievable ideal when paired with access? How do you make sure that, and I think that in the early grades it's more linked.

My kids are in first and third grade. The rigorous content for a first grader is pretty aligned with accessible content for a first grader, right? Some kids have learning challenges, but all in all, it's more paired. Once you get to eighth grade, or let alone 12th grade, but eighth grade, and you have to read To Kill a Mockingbird to analyze plot as characters develop, you can't read To Kill a Mockingbird independently or even with a little support, so you can't practice that skill.

But couldn't you practice that skill with a book that was more accessible to you, and then at least have a bridge? So It's a longer piece, but I think it is actually the thesis for what has to happen in secondary space, is we have to look coherently at literacy within core curriculum, and that might mean we have to reconsider tiers and standards and things like that.

But we're not doing anyone any favors if we're not providing students with the tools to independently continue to learn. 

[00:46:36] Ben Kornell: So just to put a clarifying point on that. So in like tier one, tier two, tier three interventions, often we have a fork in the road strategy where it's like you're either in the generalized or now you're getting these specialized services that live independently almost on a whole another strand.

And what you're saying is actually these need to be more nested in with each other so that I'm not making an either/or, I get access or I get rigor. It's access to get rigor. Is that what you're saying? 

[00:47:09] Louise Baigelman: Exactly. Exactly. Yeah. The way you said it sounds so simple, but it does require like so... Right? And if you think about curriculum companies and all of the players in the space, it's doable, but it is complicated.

[00:47:22] Ben Kornell: And so for Story Shares, in terms of the work that you do, are you embedding into a district's preexisting curriculum that they have already adopted and then raising the level of access? Or do you have your own independent curriculum and content? 

[00:47:40] Louise Baigelman: It's a really great question. So the short answer is we have done a light touch version where we did build a middle school literacy intervention program, and within that we did like the first level that we think of coherence, which is every foundational skills lesson.

If you're learning short vowels, the A says a, right? Also has a paired grade level standard. So when you're practicing that foundational skill, you're also getting that rigor at bat, we call it, right, at bat with rigor at a place where you can access it. That is a light touch version. We're excited about it.

We've seen a lot of results and just such positive feedback from educators because that doesn't exist yet, right? But then the second version, we are playing with how we wanna approach it. Our first strategy was to try and approach all of the core curriculum companies and say like, "Hey, your stuff is like not serving half the kids you sa- you wanna serve."

'Cause we do this at Story Shares, we build custom content for other companies. We could build an entire, like we could work together- Mm-hmm ... to build an intervention that would be perfect for your program. They are not yet, and I say yet very poignantly because they are not yet interested in that. I do think- Hopefully curriculum companies will start to recognize that to make an impact, but also to start seizing the new market they should.

So what we're doing instead is a lot of experiments and trying to hone in on with schools where we can pilot light touch, like field test, essentially versions of this before leaning in deeper. So 

[00:49:18] Ben Kornell: I mean, I think this is part of why I asked. It's a classic challenge that we have in this space, which is we build these artificial walls where our systems don't talk to each other.

So you're talking about coherence and, you know, I was a school board member, and I'm like I know all the education and technology companies, and yet the system itself and the incentives of the different players regularly work against coherence. Yeah. And so then you end up having a fragmented learner experience or teacher experience or schooling experience because people, the ability for IP to flow across different levels of skill or across different subject areas, I mean, it's all such walled off as a sector.

And, you know, my hope is that AI can provide some breakthroughs on that, but we're still seeing that resistance quite strongly even with adaptive AI tooling. I do wanna come back to the teacher side. So on the teacher side, there's two types of teachers. One type of teacher is, "I've got a classroom, and half my students aren't able to access this material, and the other half are, and I'm profoundly frustrated with differentiation."

And then in other situations, it's, you know, classrooms have been leveled, and so, "I've got an entire classroom of kids who can't access the information. How do I level up the rigor while providing support?" How do you support those different teacher personas, and how do you think teaching pedagogy needs to adapt given where we are from a literacy standpoint for adolescents?

[00:51:00] Louise Baigelman: Yeah. So I think the first thing, like on a high level, is that some of it is really a mindset shift where- 

[00:51:07] Ben Kornell: Hmm ... 

[00:51:08] Louise Baigelman: so far teacher education is focused on, right, K three teachers learn reading instruction. They are literacy teachers, and that is it. And for us, you know, literacy is not a finish line you cross at third grade.

It is a continuum that truly continues to evolve your whole life, even after you graduate high school. And so some of it is that mindset shift and structural shift where every single teacher should be trained, at least on a certain level, with some foundations of literacy. And that is partly just responsive to the, the reality of today's classrooms, that you will have students who can't access the core content you want to teach because they can't read.

And so you have to be able to do both of those things at once. And so all teachers should have at least that foundation knowledge. Story Shares, you know, started essentially as a book publisher, and we were trying to do-- we were just focused on engaging books for older kids that they could read and would wanna read so they could feel progress and feel seen and feel motivated to keep going.

But then we started realizing, okay, then there's the teacher piece. And many of the people on my team are middle-- former middle school teachers who had no idea that we were gonna be presented with kids who couldn't read. So that would be my number one thing. All teachers should have some baseline training.

That means middle school science teachers too 

[00:52:34] Ben Kornell: I mean, I'm K-8 credentialed, and I knew I was not gonna teach K through 3, and I did get some literacy training as part of that, but that was long, long ago, you know, basically as I was getting my credential, and then afterwards there was no ongoing support. So I think one of the most important things is just your statement around this is important.

Like, literacy is something everyone owns regardless of the age group that you're with, and that mindset could go a long, long way in terms of how people are teaching, and also when someone is struggling with access, is it my job to address that, or is it not my job? 

[00:53:16] Louise Baigelman: And also you're not alone 'cause I think so many middle school and high school teachers everywhere are feeling like, "Whoa, I'm not prepared for this, and I don't..."

And that is everywhere. And so how do we sort of start changing the messaging where it's not strange that you're struggling with this issue? It is actually the more common reality, and we as a, a system have to start supporting your ability to s- support those students. 

[00:53:41] Ben Kornell: So let's talk about the student side.

You've spoken a lot about the importance of reading engagement and self-efficacy. What strategies have you seen make the biggest difference in helping adolescents see themselves as confident readers? 

[00:53:55] Louise Baigelman: So this is another one where it is so different in the secondary space because we have to win them back.

And I think that's- Mm ... like something that's easy to just not think about, right? But at first grade, we're learning to read, doesn't have any baggage, n- negative associations. Typically, they're ready to rumble, let alone hormones, right? But by the time you get to fourth grade, eighth grade, and you've been struggling with reading and you feel bad at it, and you know that it's-- you feel embarrassed by it, you are very aware that this is not your thing by then.

And so for those students, before we can do any meaningful instruction, we actually have to win them back to the joy and purpose and frankly, their ability to conquer this thing. And so that's the first piece. Like I, I spend a lot of time talking about rigor versus access because to me, for older students especially, who are struggling with reading, all reading is good reading.

You need to find the right hook for that student that starts to put them on the pathway to changing their own relationship with reading. And so engagement, connection are-- I mean, first of all, things that every single one of us as readers relies on to choose what we read, but are so crucial for those older students who have been struggling for a long time and need to feel success and also connection.

[00:55:18] Ben Kornell: Yeah. I mean, I feel like that's part of the challenge of for middle school teachers specifically, and I know you're talking about upper elementary as well, but if you're a middle school teacher, you've got, you know, five sections of 30 kids each, 180 students. And if you're going to attack fundamental literacy, you're not just talking about the mechanics of it, which might feel very unfamiliar to you, but also the motivational elements of it, which are really, really intimidating.

Do you have any examples or anecdotes about people who are doing that particularly well? 

[00:55:57] Louise Baigelman: Yeah. Yes. So on the book side, and I just have to shout out the teachers here because I think so often, like, we hear endlessly from teachers who are desperate to find the resources to support their students in middle school and high school, and the funding's not there, right?

The policy's not there. The school leadership isn't yet focused on it. And so what I find is that the teachers have so much energy and ideas here to be leveraged, but they're working within structures that are, are challenging. Anyway, all that is to say, teachers themselves showing a student that they want them to love reading and that they see them as an individual and then giving them choice in what they read, real choice.

You know, not, "And I just pulled these out 'cause this is," right? Like, for a kid who's beginning reader, this is what it typically looks like. But something that makes them feel seen. That is one of the really crucial pieces. What we have started doing this year that we're so excited about is looking at when a school leader is on board and says, "Literacy is It 

[00:57:01] Ben Kornell: Everybody's job.

Mm-hmm. 

[00:57:03] Louise Baigelman: Yeah. And, and I recognize that the reason these kids are either not graduating or not succeeding when they graduate is 'cause they can't read, and therefore it doesn't matter if they learn, you know, the chemical compound argon if they can't read. And so when a principal or school leader or district leader says, "I'm gonna roll up my sleeves.

I'm willing to change the schedule. I'm willing to look at which blocks are in the day. I'm willing to invest in teacher training and coaching," to your point, where it's responsive and along the way, "And I'm willing to train all my teachers in how to communicate, look at the curriculum, identify connection points."

We just finished a pilot this year where we had an amazing school leader who did that, and this is going to be the proof point that we're, we're building on from here, so. 

[00:57:47] Ben Kornell: I would be remiss if I didn't ask you about AI and the future of adolescent literacy. What is getting you most excited, and what are your biggest concerns?

[00:57:58] Louise Baigelman: Yeah. I think AI and technology in general presents both a challenge when it comes to literacy and reading and, and amazing opportunities when it comes to adolescent literacy and reading, because some of our philosophy in general at Storyshares is meet them where they are, right, and grow from there. And teenagers love technology.

In fact, they're-- like, my third grader is already savvier than I am on a digital place. And so that is an asset that we can leverage to empower reading. It's also something we have to be careful about, because mastery never happens without true sustained practice. And so, you know, there is always the risk of attention when it comes to actually becoming masterful, proficient readers.

And I don't know if you're asking from like a business perspective from me. 

[00:58:52] Ben Kornell: All perspectives. I-- from an impact standpoint, though, I think we've definitely seen that like a reading assistant is helpful until you take it away, and then what have you got? Mm-hmm. And I think there's a broader worry that the urgency to teach literacy skills, especially post third grade, when AI can read it out loud for you, when AI can help make sense of it, where, you know, you can do a lot through voice, will we actually have a new wave of illiteracy coming because of perceived values of- 

[00:59:28] Louise Baigelman: Mm-hmm

[00:59:29] Ben Kornell: reading when it's still so foundational to access Any kind of content 

[00:59:35] Louise Baigelman: And I think like for that, at least for the next wave of the future, if you think about, I always challenge very literate, proficient readers to just go through one day and note all the times you encountered and had to interact with or rely on text.

And that can be everything from, you know, a lease to a prescription to a text to things that like we're-- it is to assessing whether something is true that you've just encountered, right? Like finding your own information. So I do think AI is an enabler of amazing things. I think it's also something we have to be careful about assuming will replace the need to read.

I also personally think I will continue to fight on the idea that AI will entirely replace literature. I think, you know, human art and creativity require a certain amount of emotion. I think it is a way to supercharge the types of content we're able- Mm-hmm ... to create. So my perspective on all of this is nuance.

No one thing is the thing, right? They-- Everything has-- is a double-edged sword, and so as long as we can treat it with nuance, we can leverage it and not have it, you know, have it supercharge what we're doing and not harm it in any way. 

[01:00:50] Ben Kornell: Well, we are going to leave our conversation with that. Louise Baigelman, thank you so much.

Storyshares founder and CEO. Thanks so much for joining Edtech Insiders today. 

[01:01:00] Louise Baigelman: Thank you so much for having me 

[01:01:02] Alex Sarlin: We have a very special guest this week with us. We are talking to Ashish Bansal. He's the founder and CEO of StarSpark.AI, an AI-powered platform reimagining how students learn. With decades of experience in AI, machine learning, and product innovation at leading tech companies, seriously leading tech companies like Google, Amazon, Twitch, and Twitter, Ashish brings a unique blend of technical expertise and a deep passion for transforming education.

His mission is to empower every student with a personal AI teacher and unlock a love of learning at scale. Ashish Bansal, welcome to Edtech Insiders. 

[01:01:45] Ashish Bansal: Thank you, Alex, for having me. I love your podcast and I subscribe to your weekly newsletters. I read them, and they're a source of excellent information for me.

I'm not from education background per se, and your information and your newsletter, et cetera, has helped me a lot in my journey in this edtech world. 

[01:02:04] Alex Sarlin: That is so nice to hear. I really appreciate hearing that, and especially with somebody like you who has this unbelievable background in machine learning, in AI, this is exactly-- We want you at edtech.

So it's fantastic that Edtech Insiders is able to help you spin up on the industry, and that's wonderful. So let's talk about your migration into education. We mentioned Google, Amazon, Twitch, and Twitter. These are some of the biggest, most successful tech companies in the world. What brought you into the education sphere as the place where AI could have the biggest impact on students and learning?

[01:02:38] Ashish Bansal: That's a fantastic question. Thanks, Alex, for asking that question. I'm also a dad of two kids. My daughter just graduated from college, and my son is gonna go into high school. And at the core, I was a li-- unhappy with how they were being taught math. Having experience and being at the top of machine learning field, which is inherently a fairly mathematical field, I felt that there has to be a better way for us to teach our kids math.

And after having worked on the Gemini team at Google, I felt that AI was at a point where

we can build a fairly capable AI tutor, which was affordable and accessible to everybody And I think language capabilities are very good because now language is no longer a barrier if you want to learn in Spanish or Urdu or Arabic or any other language, because in a way, math is a language in itself.

So after having tried many different alternatives from online, offline, in person for our kids, I think in a way I came to the conclusion that I have to solve this problem. I cannot expect others to solve this problem. So we started StarSpark with that mission. My co-founder is a longtime friend of mine.

He was working at Meta and Microsoft. Our standing joke, Alex, is that between the two of us, we have worked at all the FAANGs, so we had to start our own company. 

[01:03:58] Alex Sarlin: Those are amazing backgrounds, and I love that story of we looked for solutions for our own kids about how AI could be used to teach math in better ways, and we couldn't find one, so we built it ourselves.

That is exactly the kind of origin story we love to hear in edtech, and especially when you're somebody bringing all of this machine learning background into the space. And like we mentioned in your bio that you're interested in unlocking a love of learning at scale, and you're trying to not only influence learning, but a love of learning.

There's been a lot of complaints and hand-wringing over the last year or so about AI. What is the effect of AI on learning? Is it doing students' assignments for them, or is it actually able- Yes ... to increase how they learn and think and be useful for, for learning outcomes? How do you think about this question when you're trying to create a tutor worthy of your own kids and then obviously all the, everyone else?

[01:04:48] Ashish Bansal: Yeah, we have a very, very critical audience at home. I get daily feedback from my son on things that don't work. When we looked at the general purpose solutions, the chatbot that exists today, they're all answer givers. They're not actually trying to help students do critical thinking, and I think the challenge is they were built for support tools for professionals who know their job and who need assistance in doing their job faster.

This is a fundamentally different question than learning when a student is trying to learn something, because, A, the student doesn't know if the answer is incorrect. As a professional, if I'm trying to build a spreadsheet with the help of Claude, if it makes an error in the formula, I can actually debug it and make a fix.

But if a student-- if I get an incorrect answer, the student doesn't know. So I think it is not a fair answer for us to say, "Hey, this is Gemini's AI, it can make mistakes." Would you want to send your son or your daughter to a tutor who says, "Hey, I can make a mistake occasionally"? No. So the bar is very different for education.

I think that's one. And the second is, I spoke to hundreds of students prior to starting StarSpark, and they, in general, expressed frustration with having patience to get the answers and get the questions answered. I think the issue that we face is there is no space for students to be curious anymore. 

[01:06:17] Alex Sarlin: Mm.

[01:06:17] Ashish Bansal: In class, there is a fixed curriculum that needs to be covered in a fixed amount of time. Students are very wary of asking questions. A simple example is we always have TAs in college-level courses with professors, but how many students actually use the TA time? Because they are very concerned about how will they appear.

So I feel like AI, when-- if we have a purpose-built system whose focus is learning, not engagement, which is the other thing that I'll point out. A lot of systems are built for engagement, and there are many incorrect ways of going, doing engagement. Uh, for example, gamification done incorrectly does a lot of engagement, but no learning outcomes.

So how can we build something that is always available to the student when they need it? It fosters a curiosity, because I feel intrinsically kids are curious. They have many questions. Why is this formula? How did it come about? And all we have time in class is to say, "Hey, here's the Pythagoras formula. A squared plus B squared is equal to C squared."

And why? We don't have time to answer. Why the area of a circle is pi R squared or something, or what is the distributive property? So building a capable tutor that has infinite patience and infinite memory are the two capabilities of AI that we really liked. What we did not like about AI is its inability to do math, and that is fundamental, and I'm happy to go into some of the technical challenges in how the language models are trained.

I wrote a book on language models also. So we don't expect our AI to actually do math. We have a deterministic solver like MATLAB style solver. So our AI never makes mistakes because I cannot say that I'll give you a tutor that sometimes makes mistakes. So the key is we use the language models for explanation And we use a deterministic solver, which is very capable to do math for it and then support the AI.

So it's-- The important thing here is to think about purpose-built systems, not general purpose systems that we are trying to shoehorn into an educational use case by saying, "I have a study mode," which is, as we all know, is very easy to bypass and- 

[01:08:24] Alex Sarlin: Yes. I was gonna ask about exactly that. I think you mentioned the general purpose tools are answer givers.

They're designed for professionals for efficiency and productivity, and yet when you look at the statistics, they're still by far the most common tools used in education environments by both educators and students. Yeah. And so it's a purpose that is not only not aligned to learning, in many cases it's directly antithetical to learning- Correct

because getting the answer is not the point of, we have a practice set. The answer is- Yeah ... is a byproduct of the learning process. 

[01:08:54] Ashish Bansal: Absolutely. 

[01:08:55] Alex Sarlin: I totally agree with you. So the insight that you just named about large language models have training issues that keep it from being able to do mathematics at the level that you would want for the tutor, and so you have a deterministic module inside.

I mean, that's the type of technical insight that I think is really interesting and that is not always obvious to other founders. I'm curious, as you, you know, you say you've written a book on large language models, you were working on Google Gemini, you're very deep in this space. Like, what do you feel like the technical acumen brings to the tool that other founders might not know about as they develop their own version of an AI tutor or an AI teacher support tool?

[01:09:34] Ashish Bansal: I think understanding the consumer or understanding the user and their pain points definitely helps a lot. I'll give you some examples of insights that we learned in our observation of students. How do you enter an equation on a computer? It's a very simple question. Suppose I have to enter A square plus B square is equal to C square.

I think you all adults will also struggle with putting the square sign. How about a cube root? How do you even enter a cube root? I'm not taking the square root. So there is a gap in the human-computer interface. So we built our solution for writing. From day one, you can handwrite your things. Now, handwriting has two or three very important factors.

Any arbitrary equation I can very easily handwrite. Let's say I'm doing an integral. How do I input the integral sign with its lower and upper limits? How do I put limit extending to one X square minus one divided by X minus one? It's very hard to enter like, so math becomes hard. The second part is a lot of math is visual.

How do I do geometry? How do I do area of shapes? And I cannot do that. I cannot describe that there is a rectangle and there is a triangle on top, and then it is occluded by... Like, it's not possible to explain or how do I do angles, and these are inverse angles, and this is an angle bisector. So we fundamentally did two things.

We said, "Okay, you have to have a drawing interface which can interpret hand handwriting." And that actually has a neuroscience benefit that we all recognize that when you write something, you remember more. So we use neuroscience and learning principles to say handwriting is, is a nice thing, and it, it also allows you to access all parts of math, not just one of them.

And the other relative, uh, benefit from there is even younger students who have lower language capabilities, who can't type that fast, they find it very easy. We have students playing Pictionary with our AI agents because they're like, "Hey," and this is a grade three student, and they're like, "Hey, what's this shape?

What's that shape?" And that kind of open-ended curiosity-driven exploration is very helpful for the development of the mind. Another thing that I'll say is role of language is, I think, underreported and under thought of in this area. The way I would explain this is this. If you take a grade three math teacher and give them a grade nine math course to teach and vice versa, would they be successful?

And I don't think they will be successful, and not because they don't know the math. I think the way you speak to a grade three student is very different how you speak to a grade nine student. ChatGPT does not know this. Any question you give, it gives you a PhD level answer. And over time, we increase the sophistication of our answer.

When we say, let's say we are factoring quadratics. First, we have very simple, there's no coefficient to the X square term, and the roots are all integers and very nice, and there's squares. And then over time you get to complex roots, which are not no real solutions. And then you say, "Okay, now I will teach you complex numbers, and now you can have complex solutions."

So this is pedagogy because this is scope and sequence. So we built AI agents which have two things. We ingest the entire state curriculums of every US state and, and in fact, multiple countries now. So when a student comes from grade seven, I know what they are supposed to learn in grade seven. A humorous story is my son was in grade seven, he had a quadratic equation question.

I helped him solve it with a quadratic formula. His teacher gave him zero because they said that's something that we teach in grade eight, not in grade seven. Now, even though I'm a person who knows math decently well, I would like to say, but I don't know the pedagogy. I don't know the sequence in which it needs to be taught, and there is a reason you're taught things in a certain sequence.

These are some of the things that I'd like to point out to people. Think about the human who's going to use the system Because ultimately we are solving a human problem. The AI is just a way, a means to a goal, and the goal is still for the human to understand and explain. So we don't need to teach kids how to prompt.

That is, I think, the wrong experience, and I think this is something that is my pet peeve with a lot of the educator solutions is they-- when I look through the educator training from variety of companies, they're all, "Prompt like this, prompt like that." But the new model version prompting instructions change, so it's not a durable learning thing that I'm taking that I can transfer from one task to the other.

[01:13:49] Alex Sarlin: Yeah. 

[01:13:49] Ashish Bansal: So these are some of the challenges that I'll just, uh, point out, Alex. 

[01:13:52] Alex Sarlin: No, it's really interesting to hear you talk about combining user insights, like watching students having to do these equations and not being able to do them because of the interface restrictions with the technical solutions, right?

You say, "Oh, we'll just create a handwriting solution." That's something that is not always accessible to other founders, but it's be incredibly valuable for end users, and I feel like I'm hearing that in a few different ways. One thing I'd like to ask you about, you're mentioning the scope and sequence or alignment to state and even international standards as a sort of pedagogical backbone grounding the AI in curriculum.

That's a very powerful mechanism. But you're also talking about curiosity and the role of an AI tutor in allowing students to ask questions that may not be the next question on the textbook or- Yeah ... next question in the pedagogy. I'm curious how you are working on StarSpark being able to balance this structured scope and sequence level thinking that a school would require and the curiosity that an individual might say, "Hey, I wonder why Y squared and cubed works the way they do."

And yeah, I'm curious how you're balancing those. 

[01:14:53] Ashish Bansal: Yeah. So we actually have a learning plan, which is we are able to cover an entire year's worth of material in about thirty weeks, with about four to five hours of work a week, sometimes even less for the students. What we have realized is there is a neuroscience basis for a lot of our work.

Based on brain development, we learn different things. For example, spatial reasoning is something that develops around puberty. Now, puberty does not hit everybody at the same time, but i- if you are in school, you will get spatial reasoning at a fixed point in time. So some students will struggle with it, and some students will not, and then like a growth spurt, their brain will also have a spurt, and then suddenly those things will become obvious to them.

So having this self-paced curriculum allows students to take more time on certain topics. On an average, we have students who have completed three full grades in a year and a half because it averages out, I think, because they don't struggle on every topic, but y- they need to have an adapter. This is what, what we mean by personalized learning.

In addition to that, the love of learning comes when they have the ability to ask open-ended questions and relate concepts to real life. I think the big challenge in math education is a lot of students are like, "Why am I learning these formulas?" Now, if you say that you, if you want to build pyramids in Egypt, you better know the Pythagoras formula.

Um, and right? I mean, there, there are lots of documented uses of how they used mathematics to set up. And at that scale, how do you know if the li- si- the sides are lining up correctly and how many blocks to put and how do I find volume of solids and things like that. So we allow students to plug in their interests.

So I have a student who has plugged in One Piece, which is a very popular anime for kids these days, and we create math problems with Luffy and, uh, Devil Fruit. It's the same word problem. Tom has three strawberries. He gives two to Jerry, or Luffy has three Devil Fruit and he gives one to Zoro. Now, suddenly the kids are way more interested in the second problem than the first.

It's the same mathematical concept. We train our tutors to give a lot of analogies. Mm-hmm. So when we are trying to explain something, so we were trying to explain angles and our tutor came up with, uh, "Hey, when a soccer player takes a penalty kick, they're trying to compute this angle," because the student put, uh, soccer as their interest.

Somebody put baseball, and it was trying to do averages using RPA as an example. So when you start doing, using these analogies to- Connect the math concepts to real world, it becomes much more concrete. And this is a challenge with math, I think it-- as it grows, progresses in higher grades, it becomes more and more abstract, and people start to feel that this is not useful for real life.

But it is actually the science developed because we were trying to find better ways to explain what is happening in our world. And somewhere around the push to complete the curriculum and finish everything, we sort of lost, uh, this connection to the real world, and I think this is where AI does a really good job when trained with the proper guardrails in a purpose-built system like ours.

[01:17:54] Alex Sarlin: Yeah. That's-- It's really interesting. So you can concretize, provide analogies, uh, do interest-based adaptivity and say, you know, to change what the problems are to meet students' interests- Yeah ... and also allow those open-ended questions. I agree that the love of learning comes from both relevance and sort of being able to sometimes step away from a set curriculum and, and make sense of it in new ways or sort of ask about something that's off the beaten path.

It's a really interesting combination. I feel like I'm hearing a lot of different types of insights that you've baked into the design of StarSpark that I think are, especially in combination, can create an incredibly powerful experience, right? Combining that open-ended- Yeah ... and structured, the language, adapting the, the way that, that explanations are happening, adapting the way that the topic of the questions, you know, can be really powerful for engagement.

And as you said, you know, splitting the language model from the math model so that they're-- they have two different models sort of doing their own specialized work in conjunction. It's really intriguing. 

[01:18:51] Ashish Bansal: Alex, we had to build five different models. We had to build- Oh my God ... a separate computer vision model for handwriting and geo-geometric ship shapes.

We had to build a model for text to speech. So how do you pronounce equations in LaTeX? So we are all taught how to pronounce an equation, so we fine-tune our own text-to-speech model because all our models actually speak in multiple languages. We also had to build a speech-to-text model. So if you speak an equation, we directly transcribe an equation.

Wow. So there is fundamental work we had to do to make that, uh, interaction with the human much more easy and natural, so you can just speak to all of these things. So a lot of focus in today's world happens in the large language model era, but we forget there are so many other modalities that you have to build compensating systems.

So our, uh, main, uh, tutoring interface has incorporating five different agents that specialize in speech, in text, in, in drawing, and then an overarching orchestrate and a deterministic solver all working together. 

[01:19:53] Alex Sarlin: That's incredible. So when I hear you describe both the philosophy and sort of educational stance behind this, but also certainly the technology infrastructure, which is very sophisticated, it makes me wonder about what the plan is for StarSpark.

Are you looking to have the tutor be a complete standalone tool? Is it something that you anticipate that you might sort of work with other distribution platforms or companies? You know, would it be some large edtech company that uses the StarSpark system inside of it? I'm curious how you're thinking about distribution and getting this into as many hands as possible.

[01:20:27] Ashish Bansal: Distribution is the key challenge in edtech, as you very well know, and, uh, several episodes have covered this in the past in, in the Edtech Insiders. 

[01:20:36] Alex Sarlin: That's 

[01:20:36] Ashish Bansal: right. You see, our mission is to give every student in the world a capable personal AI tutor, and we believe that we've... Partnership is a strong aspect of this.

We would love to partner with companies to get this technology because we have the technical know-how, and as you'd mentioned earlier, sometimes top technical talent, if you're getting paid 100 million by Meta to build, uh, something, uh, do you want to go into edtech? I think it takes a little bit of craziness for people to do this.

So we, we are more than happy to work with it, and I think there's a lot of debate about will AI replace teachers, and I think the role changes. And we have found the most successful deployments where teachers, parents, and students, this trifecta works together. That is the most successful deployment. And I will not make a claim that says AI will do everything.

I don't think that is possible, at least in the near term. And we do need to make sure that all of these three work together. So we are very happy to work with as a supplemental tool, as a teacher assistance tool. You know, there are many challenges. One simple challenge I'll, I'll tell you is if you ask a teacher to frame a difficult question versus an easy question on a topic, the teacher knows what makes a question difficult.

AI does not. Even the best model. So we have to do a lot of work in what is... So then we relied on frameworks like Depth of Webs, Depth of Knowledge, or Bloom's Taxonomy. What is a level one question? What is a level two question? What is a level three question? How do you differentiate between them? Is more computation equal to more difficult questions?

Which is a lot of people think like that, but that's not true. It's just you can solve it with a calculator, so it's not really more complex from a critical thinking standpoint. 

[01:22:13] Alex Sarlin: Mm-hmm. 

[01:22:14] Ashish Bansal: So how do we create these things is very important and requires a lot of deliberate thinking in building. And to coming back to your question about distribution, it is a challenge.

I will not be, uh, I will not say it's easy or it has been easy for us, and we would love to partner with firms and bring our technology out. The important thing, again, for us is we need to help the students 

[01:22:37] Alex Sarlin: I think one of the complex dynamics in the edtech space is sometimes some of the most sophisticated and well-designed and pedagogically sound tools are hidden inside products that aren't-- that people have not heard about because they don't have big sales forces, and they don't have big install bases, and they haven't been around very long.

But, you know, something like this that's not only AI native, it's like very sophisticated AI native with many different models and lots of different deep thinking, feels like it could be a huge enhancement to virtually any math product that's out there. I would be surprised if you don't get a few calls and emails coming out of this interview.

I-- it-- because let me put it this way, I'm sure there are lots of math software products that have been trying to solve that. Even the question of how do you get an equation into the, you know, back and forth- Yeah ... into the computer. They've been- Yeah ... wrestling with that for years, and they build these sort of calculator interfaces.

The fact that you're going straight to handwriting is a very sophisticated answer, but it's also technically challenging. So the fact that you can do it is, is incredible. Looking ahead, you know, you are coming into this field from the AI space and coming with this amazing knowledge base. Where do you think we are going to go as a sector as AI and education get more aligned and more sophisticated?

We-- you know, I think we're a few years in. A lot of people are still using just off-the-shelf frontier models for education purposes. As we get a bigger and better suite of meaningful made-for-purpose education tools, what do you think might happen to the education space? 

[01:24:04] Ashish Bansal: Let me actually poke a little and ask a little more, uh, provocative question.

Are we ready for the era of personalized education? I think that is the fundamental question that we need to answer. And the challenge that I see is we have to cover a fixed curriculum in a fixed time in school. Irrespective of whether I have mastered the content or not, I will be moved on to the next grade.

And the challenge with math is that math is cumulative. It builds on the concepts you've learned in the previous grade. And if you do not master those concepts, my favorite is analogy, which is very, very crude, is that if I got an eighty percent, which would be a B, in grade seven, and then I moved to grade eight and I got an eighty percent again, I have point six four cumulatively.

That's right. Sixty-four percent knowledge gained. So grade nine is progressively harder. And if you look at the math anxiety studies, they show this very clearly where math anxiety starts to increase in middle school and peaks in high school, because high school is where all of your m-elementary and middle school concepts come together.

And by that time, you have lost that. So the question I think we have is: Can we have a class of ten students, twenty students, thirty students, where each student is doing a slightly different version of the curriculum based on where they're at? And I think that is the m-biggest friction that I see in our education system.

It is built for group education. It is not built for one-on-one education. And we need to really think about how can we empower One-on-one education. I think the model is upside down where we have teachers doing the education, the lecture part of it, and the AI is used for help. I think it should be the other way around.

Just, uh, for as a thought, uh, experiment that I want to propose for, as I said, I'm gonna be provocative a little bit. What if the AI did the lectures and the students help the students when they get stuck? So we completely flip the paradigm. We're saying the knowledge transfer piece is the lecture, which can be done one-on-one because we can then do it in Spanish and Mongolian and Vietnamese and all of these languages, which a teacher can't.

But then when the student gets stuck, that is where the teacher comes in because then they know what is going on with the student, what is their previous work, how do I help them? Did they have food this morning? There are so many challenges that teachers in classrooms face. Is the student paying attention?

Do they have any disabilities? So building all of those contexts into an AI model may be the place where I think the humans-- And the motivation aspect, you know. Nothing motivates more than a teacher, right? Because students do look for some validation. So maybe we need to rethink, are we ready for an era of personalized education?

And how can we empower our educators, our school districts, and our students to really benefit from one-on-one education? 

[01:26:59] Alex Sarlin: Yeah. That's a really fascinating thought experiment to think about, you know, in a world where personalized education, fully differentiated, individualized, is truly in place at scale.

Yeah. How does a school system have to adapt? What is the role of the educator? What is the role of a classroom or a grade, right? When people are, are able to sort of accelerate- Yeah ... or move at their own pace, like legitimately able to, not just within small constraints- Yeah ... as we often have now. It's a great question, and I think it's one that we should begin to be asking ourselves because I agree with you very wholeheartedly.

But I think we're in this era where we're, we're just a few years into this sort of generative AI being accessible to everyone through commercial tools, and people are using it, finding it incredible, but it's not really aligned to the education use case in so many ways. You've named- Yes ... at least like eight ways, you know, in this, in this conversation.

As we actually start getting it ready for education built for purpose, as we start understanding the, the pedagogy-- By the way, you've done a great job of explaining so many pedagogical ideas. You say you're not-- you don't have an education background, but you're picking it up pretty quickly. As we get it to be educationally sound, as we get it to be implementable in a variety of different environments, as the tools get better, as the structures make more and more sense, and as we actually move, we have to really rethink what the system is actually gonna look like.

'Cause at some point, that's gonna be the barrier to change. Right now, it's still the tech, I think. But when the tech is sophisticated enough, it's going to be the existing systems that actually keep things from moving as fast and, and, uh, effectively as it could. So I think your thought experiment there is really powerful.

I'm gonna mull on that myself. 

[01:28:33] Ashish Bansal: Yeah. 

[01:28:34] Alex Sarlin: Ashish Bansal is the founder and CEO of StarSpark.AI, building really sophisticated AI tutoring for math, including language adaptation, math, deterministic models, handwriting analysis, speech-to-text, text-to-speech, all in one tool. It's incredible, and they're thinking about go-to-market distribution and making a, creating a, a really better world for students.

I appreciate this conversation. Thank you so much for being here with us on Edtech Insiders. 

[01:29:02] Ashish Bansal: Thank you, Alex, for having me. 

[01:29:04] 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.

This season of Edtech Insiders is brought to you by Starbridge. Every year, K-12 districts and higher ed institutions spend over half a trillion dollars, but most sales teams miss the signals. Starbridge tracks early signs like board minutes, budget drafts, and strategic plans, and then helps you turn them into personalized outreach fast.

Win the deal before it hits the RFP stage. That's how top edtech teams stay ahead.