The New York Angels is Investing to Get ROI from AI

This transcript has been AI generated. Please excuse any typos.

[David]
How are investors going to get ROI from AI? We hear a lot about the investment in artificial intelligence, but the New York Angels are working hard to find investable companies where artificial intelligence will produce real profits. We’ll meet NYA member Josh Powell and Eric Tao, the founder of Megaminds.

They’re using AI to scale their experiential learning model. Welcome back to this New York Angels edition of the Angel List. I’m David Hemingway.

I’m a five-time founder and a member of the New York Angels, where we fund and mentor great young companies. So, Eric, you’re a technology expert, came from a family of academics, so I guess it was inevitable that you would blend those two together. Tell us how Megaminds helps students learn.

[Eric]
So Megaminds is an AI-powered simulation platform, and essentially we have 3D environments where students enter. They’re very much akin to games like Roblox and Minecraft. But within these environments, we can populate AI characters, sometimes up to a dozen of them within a single virtual environment.

What that does is it creates a space for simulation. Students can practice career explorations. They can practice math.

They can practice academics. And that space gives them a safe space for practice and to fail and to try again. The research term is productive struggle.

And when you think about it as a game, what are they doing in a game, whether they’re shooting their friends in paintball? They’re practicing, right? They’re getting better every single time they play, and they’re improving their skills.It’s the same idea, is that the students enter a simulation. They practice their skills. They can fall on their face.

They get feedback from the AIs, immediate feedback from the AIs. The teacher gets the feedback from… We collect all the data that’s happening inside the 3D environment.

We have a separate AI agent that’s not student-facing, that’s serving up all of this data into actionable insights for teachers in near real-time. So a teacher can see within 30 seconds that Jimmy is having a problem with word problems and not understanding how to solve the problem. We’ll be able to see that in 30 seconds.

[David]
It saves everybody the embarrassment that I used to have when they would call on me in the middle of class and I didn’t have the answer.

[Eric]
Precisely. It’s called the effective filter. And each student’s effective filter is a different level.

When you have students, like say intervention students, that have grown up thinking they’re not good, they’re not smart, they’re not going to raise their hand in class. They’re not going to be active learners. But in our environments, as we’ve seen actually in intervention use cases, they’ll speak to our AIs for hours and they have no problem because they know it’s a safe space.

Non-judgmental, non-judgmental safe space.

[David]
And you have data, tell us about it, that shows that this actually enhances learning.

[Eric]
Sure. So this spring, we did an intervention study with a Title I, which is a low-income school in South Florida. So they had a population of students that 0% of them were at grade level and in fact 83% of them were two grade levels or more behind.

This is the unfortunate cohort, and this is a story that’s repeated across the country. This is the unfortunate cohort of students who were in the fourth and fifth grade when the pandemic started. They were supposed to learn their foundational math skills.

They never did. And those gaps followed them. Now they’re teenagers, now they’re eighth and ninth grade students and they still don’t know their foundational skills.

So the head of the school contacted me in March and they were like, we’re about to put these kids through an intervention. I had introduced her to the platform at an ed tech conference just a couple of months prior and she was like, I know you mentioned to me that you have middle school math modules. Would you want to participate in this?

I said, of course. So the students, we broke them up into two groups. We had a control group who didn’t use Megaminds and then a group that used Megaminds daily for 30 minutes for 10 weeks, not quite three months.

By the end of the semester, 67% of those students that used Megaminds had achieved grade level, whereas only 22% of the students in the control group were able to achieve the same. So I was able to do a two hour download or debrief with the students afterwards because I really wanted to, I mean, those numbers were shocking and I wanted to really get to the bottom of what happened and why they felt, how they felt Megaminds helped them. The word that kept coming back over and over and over again was confidence.

They said, working with the AIs allowed them to build their confidence, their math skills. And these are the students that I was talking about earlier where they’ve grown up thinking they are not good at math, not math kids. They were not going to be active learners and in these environments, they allow them a safe space to practice and most importantly, to fail and to learn from their failures and to try again and build, slowly build their confidence back.

[David]
It feels like you’re bringing almost one-to-one learning to the classroom.

[Eric]
It’s a version of that. Yeah, for sure. Because the AIs adapt to whatever the students say and how they’re feeling and how they’re behaving and all of that actually, all that information there is passed on to the teacher as well.

So we kind of hit them from two angles. So number one, the AI can tell sentiment. So if the student is frustrated, that that information goes to the teacher and then the teacher can see that Jimmy is being frustrated by word problems, unable to grasp word problems.

What the teacher told us after the intervention was she would open up her laptop and she would sit and she would just literally watch the students as they went through the modules. If she saw a student was having a problem, she could go walk, sit right next to the student, help them, help solve that learning gap on the spot. If she saw that it was a collective issue amongst the classrooms, that information is served to her as well.

She would prepare a mini lesson for the next day that helps solve that gap. Now traditionally, she would have to wait two weeks for the test or for the quiz to even spot the gap. And then she’d have to expend all the energy to go through and grade everything and figure out, okay, student A, student B, student C need this and then teach the lesson.

By then it’s already too late. So it really, it changes the approach of instruction and we call it a time shift. It allows the teachers to spend more time one-on-one with the students, which is where they’re most productive.

[David]
Yeah. Sounds amazing. Tell us about your experience of raising money from angel investors.

[Eric]
Well, it was remarkably smooth with the New York Angels. It was a fantastic experience. I found the organization to be extremely professional and everything was well organized.

I just followed what they told me and I was able to close a sizable pre-seed investment round through that. So I’m very grateful for the experience.

[David]
Well, and our ambition, of course, is to bring more than money to the table and Josh is certainly a very accomplished educator. So hopefully you’ll find that that’s also the case.

[Eric]
Yeah, absolutely. I think, unfortunately, I haven’t been able to meet all of my investors in person. I’ve met only maybe a third of them, but they all have a vested interest in education, whether it’s their grandchild who’s going to the new AI school or they all have some sort of connection to education that I found at least.

[David]
So tell us what the next year looks like for Megaminds and to the degree that you already know, what is the grand vision?

[Eric]
Well, I’ll start with where we’re seeing the most traction right now and that’s in career technical education. So again, setting up those simulations of a hospital where a student can pretend to be a nurse and work with AI patients or in an education use case, they can have their first parent teacher conference, but the parent is an AI, right? They can have their first student discipline meeting and the student is an AI and they can test their strategies, they can test their approaches, they can fail.

And the reactions are remarkably lifelike. So that’s where it seems at least in our sales calls that the connection is instantaneous. We’ve had calls where the CTE director within 30 seconds of seeing a demo was like, okay, I want to pilot this.

There just seems to be an instant connection. So that to me is a signal of early product market fit in this vertical. I think we have our eye on math, of course, because we have the results of that study.

But math is a bit of a bigger leap, right? You have to convince people, okay, why 3D for math? It’s not quite as simple as a connection as CTE, but the results are there, right?

And the results, we’re able to get the attention of the Michigan Education Corps who they handle intervention across the state of Michigan. So through them, we’re going to be able to create a series of pilots that we’re going to try to replicate what we did in Florida across a larger implementation. But I think that’s the key as well with math, is you have to have efficacy, you have to have research data that shows it.

But I think we might be onto something, and if we’re able to prove that across a larger base, then math is going to be one of our major verticals, I think, moving forward.

[David]
Just fascinating progress and an amazing application of AI. Josh Powell, you led Eric and Megaminds through the New York Angels investing process, and we know that EdTech could sometimes be a hard sell for investors, but seems like Megaminds has some real traction.

[Josh]
They do, David, and that was one of the things that really excited me about the opportunity. Obviously, at New York Angels, we see a lot of AI-driven opportunities, but few have established meaningful traction in the marketplace, and that’s what made Megaminds catch my eye. I’m also an EdTech entrepreneur for 20 years and have been an operator in that space, so I certainly understand something about the founder’s journey.

And while the AI-driven solutions have made it easier to build out products, even with a small team, one of the things that it hasn’t been as good at solving is that go-to-market and that uptake. And I think in the case of Megaminds, the fact that they have had multiple RFP wins, they’ve been recognized for the innovation in their solution and have multiple districts that have actually purchased, not just piloted, but purchased their solution and found value in it, was a real driver of interest. I also was able to interview some of those districts and early customers as part of the discovery process for Megaminds, and that certainly gave me confidence that this was a real solution driving real learning outcomes that stakeholders believed in.

[David]
And very impressive that they have data that shows that students do better with Megaminds than without, right? That’s real, real-life data from the classroom.

[Josh]
A hundred percent, David. And it’s compelling data, albeit, again, early in the process, in their journey, but to be able to capture that data and measure sort of pre-Megaminds versus post-program outcomes is something that’s very important. I consider myself a data guy.

My ed-tech platform was built on a foundation of data, and the mechanism that Megaminds has to capture student learning data throughout their journey on a very individualized and granular basis and make that information accessible and digestible to key stakeholders, including teachers, parents, and potentially administrators. It’s very, very powerful. And I think a clear indication that the model that they’ve established around interactions with AI-driven avatars that are part of a learning experience is not something that just exists in isolation.

It’s a mechanism to scale learning and data collection across an entire classroom, an entire grade level, or even an entire school or school district in a way that provides transparency and keeps the teacher in the loop. And to me, that is sort of a killer app that I’ve been looking to find for my entire career in ed-tech. Megaminds is making real strides in demonstrating that in an actual classroom setting, and it’s a key component.

[David]
Now, Josh, investors always want to moat. What does Megaminds moat, and how do you see that going forward?

[Josh]
Well, look, I think we can’t be under any self-delusion that there’s any moat that is not crossable, right? I mean, if you’re doing something good, interesting, and powerful, especially if it’s actually gaining traction and making money in the marketplace, there will be fast followers and competition. So just to be clear.

But I would let Eric speak to some of the specifics of his moat. But broadly speaking, as an investor, I think that there is a data moat here, first of all. So as you are gathering more information from more experiences that the students are participating in in the 3D worlds and the simulations that are designed to mirror realistic interactions in the actual physical world, translating into the digital world of Megaminds, that data set obviously makes the AIs more intelligent, more responsive, more targeted in terms of the kind of instructional delivery and intervention they can deliver to target particular student learning gaps.

And even if you’re not looking for gaps, per se, the ability to meet students where they are and extend that learning in a sort of real-time process of data analytics is very different than a traditional learning process, where you might have periodic assessments, whether it’s a state-sponsored, high-stakes, summative assessment, as they’re called, or whether it’s a more ongoing formative assessment process, where teachers may be periodically collecting data and then figuring out what to do with it. Megaminds has an interesting way of synthesizing that process so that the instructional delivery is intimately connected with the data and with the sort of next steps in that instructional delivery.

So it really takes a siloed, bifurcated process and unifies it in a way that scales. And that kind of delivery really wasn’t possible until recently at that kind of scale in a very coordinated way across an entire classroom of students simultaneously, because typically you only have one teacher in front of perhaps 30 students, and to individualize instruction using that model is a very, very difficult thing to do.

[David]
Excellent. Great. All right.

I’m sure we’re going to do many more episodes on AI, but this is a great one to learn how AI is actually helping children learn and bringing really a new dimension to New York Angels investing. So Eric Tao from Megaminds, Josh Powell, NYA member, thanks so much for joining us today.

[Eric and Josh]
Thanks. Thanks, David.

[David]
You can learn about joining our group or applying for funding for your startup at NewYorkAngels.com. And you can find contact information for our guests today or reach me with your comments or questions at TheAngelNest.com.

A reminder that we don’t make or recommend investments at The Angel Nest, and this program is for informational purposes only. We produce The Angel Nest from New York, the media capital of the world, and we get help from Rob Higley, he’s our producer, and Charles DiMontebello at the controls of CDM Studios here at the famous Art Deco Film Center building just west of Times Square. I’m David Hemenway, thanks for listening, and so long until next time.