130: Mastering Skills through Video: A CEO's EdTech Odyssey with Josh Kamrath of Bongo

Holly Owens (00:02):

Hello everyone and welcome to another amazing episode of Edup Ed Tech. My name is Holly Owens

DaNadia Johnson (00:10):

And my name is Nadia Johnson, and we're your hosts

Holly Owens (00:14):

And we're excited today. We have a wonderful guest with us who's going to share so much about learning and credentialing and all the different things in the world that we want to hear about. So we have Josh Kath, and he is the CEO of Bongo. Welcome.

Josh Kamrath (00:30):

Yeah, thank you for having me. Well,

Holly Owens (00:32):

We're excited to get to talk to you in here about all things bongo, but before we jump into all that, we want to know a little bit about you. So tell us your story. How did you become a CEO of an EdTech company and get into this space?

Josh Kamrath (00:47):

Sure. Yeah. I actually have a little bit of a interesting kind of non-traditional path to becoming the CEO. So I guess almost about 12 years ago is been my tenure here at Bongo, and I was actually the first employee of the company. That's so cool. Really,

(01:07):

It was founded by a tenured professor, so a university professor of mine. I had worked at a different company right after I graduated from school, a middleware company, and then my girlfriend at the time was from San Diego. She was a year behind me at university, and when she graduated or was coming up to graduation, she wanted to move to San Diego. So I put that first job and put my feelers out. And Dr. Jeff Lewis, again, professor of mine was one of the people I reached out to and just asked if you knew anywhere in southern California I could work. And he said, well, I have a company that's coming to minimum viable product stage. Why don't you try to go sell this tool into universities and colleges? So actually moved in with my girlfriend's parents' basement or into their basement. We weren't married yet, and I thought it was going to be temporary. It ended up being about two and a half years living there, but it was awesome. Isn't that how it always is? You're like, this is going to be

Holly Owens (02:08):

Like a year.

Josh Kamrath (02:10):

Yeah, no, it was awesome. I got to develop a really good relationship with them and ultimately married my wife Josie. And that's the kind of origin story of how Bongo started to become commercialized. Fast forward a little bit to, I guess five and a half years ago, the university basically was putting pressure on Jeff to say, they were saying, Hey, either come back and keep your tenure or no more sabbaticals trying to run this company. So at that point, he went back to being a tenured professor, and I took over as CEO and we've just continued to grow the company. So that again, was about five and a half years ago.

Holly Owens (02:58):

That's an amazing story. I do too.

Josh Kamrath (03:02):

I

Holly Owens (03:02):

Really love that you were the first employee, you got connected through a professor in education, and that's really neat. Like, Hey, just come do this. It's like you're in the right place at the right time. It Was

Josh Kamrath (03:15):

The right place, right time, and it was also cool. He originally created Bongo or the first iteration of Bongo to solve his own problem. He was a business communications professor, and that's right when online courses were just coming of age and I wanted a tool to basically facilitate face-to-face presentations, sync up visual aids, and get feedback and coaching to learners and couldn't find a tool like that in the market. So he is set out to solve his own problem and developed bongo. But yeah, it's been an awesome journey. Still have a super good relationship with Jeff, and like I mentioned, he's back to teaching and also can't fully eliminate the entrepreneurial bugs. No,

Holly Owens (04:08):

That's hard. He's Operating.

Josh Kamrath (04:10):

Yeah.

DaNadia Johnson (04:12):

That's awesome. I think that's probably one of the more unique stories we've heard on edup ed tech. So at Edup Ed Tech, we are an ed tech podcast and we know the EdTech landscape is evolving. So what emerging trends or innovations within EdTech do you find most promising and how do you see them shaping future of education before we get into bongo, because I'm very excited to learn more there.

Josh Kamrath (04:39):

Yeah, totally. I would say my opinion of course is going to be a little biased because at Bongo we're trying to develop our own niche or a needle mover in the universe, but

DaNadia Johnson (04:54):

Right as you should. Exactly. But of course, I'm super excited about just the different capabilities and enabling properties of AI

Josh Kamrath (05:04):

Or generative ai. So there's been some AI my entire career in academics or in the academic landscape, and it's a little bit of an ebbs and flows kind of thing, but I do really both see what directly is involved with bongo, but also just some of the other products out there in terms of truly adaptive, truly personalized learning. It's actually obtainable now, I feel like compared to nine or 10 years ago when adaptive learning was a really big thing and I think kind of fizzled away. It's just some of the things you hear about, I suppose, or the cutting edge buzzwords, if you will. I feel like we're actually a lot more attainable now compared to previously.

Holly Owens (06:03):

Yeah, absolutely. I feel like AI and adaptable learning has been around for a while. Teachers do adaptable learning all the time

DaNadia Johnson (06:12):

To cater to their different audiences,

Holly Owens (06:15):

But I really feel like what you're saying now, there's the technology base to support that instead of having to finagle the technology to do what it should be doing, and you have to pull the pieces together to actually make it do that. So

Josh Kamrath (06:32):

I

Holly Owens (06:32):

Really think about a lot of that stuff too.

Josh Kamrath (06:35):

And I think too, some of the things, I think the majority of vendors out there, how they've been leveraging generative AI is to democratize content creation, make it really easy and accessible to just adaptive. For instance, 10 years ago, you'd have to create a massive framework with huge decision trees to make it work.

Holly Owens (06:57):

And

Josh Kamrath (06:58):

Now generative AI can figure all those paths, learning paths out almost instantly. So having tools like that that enable super quick access to personalized learning or to, I guess what bongo is working towards is more validating that somebody truly authentically knows something. So that's always has been our mantra, but now with AI as more of a centerpiece, it's actually scalable as opposed to in small cohorts how it was previously facilitated. So again, kind of an example of how AI is just an enabling type of technology, not like a silver bullet.

Holly Owens (07:50):

Yeah, absolutely. It still needs the human piece of it. So let's talk about it. I typically stay away from doing too much research. I want to be genuinely excited and when you start talking about it, and I'm still going to be excited, but I went way deeper than I anticipated bongo. So cool to me. So tell the audience, what is bongo? What do you do for education institutions, corporations? Tell us all about it. What's it about?

Josh Kamrath (08:19):

Sure. Yeah. So bongo, what it is is a video assessment tool that has learners demonstrates skills or competencies through recording a video of themselves doing something. So that could be an open-ended question on curriculum or content that they just learned. I was an economics major in university, and all of my testing or evaluation was either multiple choice or open-ended.

Holly Owens (08:46):

Exactly. That's exactly

Josh Kamrath (08:49):

The visceral reaction.

(08:52):

But it's ubiquitous, right? And why it's ubiquitous is it's scalable. And so what bongo brings to the table is when we were in university, I'm sure you guys recall when you're called out in a classroom setting in a one-on-one basis. So if you're fiddling with your phone or passing notes or something and a professor or teacher calls you out and asks you a question on the curriculum, one of the reasons they do that is to put you see how you think on your feet and whether you're paying attention. And it's a very authentic representation of do you know what you're talking about? Or can you demonstrate the application of knowledge as opposed to just having that knowledge. That's what bongo validates. So we have learners record a video of themselves doing something, and then we have a bunch of feedback workflows. Historically they've been human feedback workflows where the professor or peer could be self-evaluation, goes in and watches that video recording and critiques provides feedback and coaching around that.

(10:03):

So as I mentioned Dr. Jeff Lewis when he started bongo, that was for business communication courses or entrepreneurial courses like business as the college of business type of courses. Now bongos in several hundred universities and colleges all around the world, most of our user bases in North America or Europe, but it's also touches most departments. So there's lots of different areas where showing you and demonstrating you can articulate your knowledge is what's most important when getting a job, not just whether you can recall facts or figures, but actually showing you can apply knowledge. That's what bongo essentially facilitates. And as I was mentioning historically, it's been human evaluation. Now we certainly still have human evaluation as a core foundational piece, but we also have a bunch of AI capability that will give feedback both on the delivery. So how confident or nervous was that learner? Were they using a bunch of filler words?

(11:13):

What was their speaking rate? Things like that. As well as what I think is insane is contextualized feedback. So an instructional designer or faculty member can upload source material. So let's say chapter seven of a textbook, our tool will ingest that, come up with the learning objectives, what are the key aspects of chapter seven, and then validate that the learner, so the student when they're giving their presentation actually hits on those learning objectives. And it can be, like I said, it's contextualized, right? So it can be use metaphors or you don't have to say something specifically. So an example, maybe a little more in the corporate sector, but if you're familiar with the concept of the parking lot, you can have the Objective. Yes. Lots of parking

Holly Owens (12:02):

Lot ideas.

Josh Kamrath (12:03):

Exactly. You can have in our tool, you can have the tool look for did that student demonstrate the parking lot concept? And then the student could say something like, Hey, great question. Let's take that offline. So the student never says parking lot, and the AI tool will identify that that student is saying and representing that concept. It's remarkably accurate and cool. I guess the last little tidbit with it is, as I was kind of alluding to, just from a scalability standpoint, observational assessment is the type of assessment or evaluation that has the most rigor and gives the most authenticity of does that person know that thing? But like I said, it only scales one-to-one like pure linearly scales. So with bongo, you can have, let's say 500 students complete an assignment and you might have a professor and probably a professor and a few TAs, watch all those videos and critique off of a rubric.

(13:10):

You can have the rubric uploaded as a source material in bongo and have the AI evaluate off that same rubric. Generally, the AI will come within about 2% variance compared to that human evaluator. Wow. So in a large capacity indistinguishable from human evaluation, of course, we still our philosophies, we want the human to be in the loop. We still think it's incredibly important to have the professor validate and basically approve, yes, the AI scoring them accurately. But what this can do, in that case the professor is massively scale them, really improve just the efficiencies of grading and make them able to actually have that rigorous type of evaluation or assessment, but do so at a very large scale. I

DaNadia Johnson (14:09):

Love that.

Josh Kamrath (14:10):

That's a little bit on my soapbox. So sorry for the long explanation.

Holly Owens (14:13):

No, I was listening the whole time.

Josh Kamrath (14:16):

Oh my gosh,

DaNadia Johnson (14:17):

I love

Holly Owens (14:17):

It. Something that's so needed.

DaNadia Johnson (14:20):

Yeah, different ways to show mastery. I think that's where education goes wrong sometimes is we just have that same traditional method of showing mastery, and that doesn't really show that sometimes it shows that you know something, but it doesn't show that you can do something. And I think that's the biggest thing that I is big in instructional design is most of the time we want to make sure that people are doing what they need to be doing the performance, not just the knowledge, not having it in your brain. So I absolutely love that, and I think that was one of the biggest things when I was a former educator within the personalized learning realm is different ways to show mastery. So I love that. I do see on your website that you talk a little bit about data privacy and security. So I'm interested to know a little bit more about what steps do you guys take and maybe what steps can ED tech tools or companies take to safeguard their data and protect the privacy when it comes to technology and learning?

Josh Kamrath (15:36):

Yeah, I think I'll answer that in two parts. So one, just from a straight data privacy, security protecting personal identifiable information there, it is pretty much in my opinion, like a prerequisite now to be GDPR and CCPA compliance. So some of the things we do are encrypting information both in transit and at rest, having a bunch of systems and processes internally on both who can access different information on the backend, but also keeping logs and records of that. So I think that's been a little bit of an evolution since I started at Bongo to now is now there's a bunch of basically governmental regulation that requires you to take data privacy very seriously. And as I mentioned, Europe really led the forefront with GDPR and then California, and now there's a bunch of states like Texas and New York that are kind of following suit and here in North America.

(16:46):

So I think that's just going to continue to become more of important components to any software company. And like I mentioned, it's in my mind pretty much a prerequisite just to even operate keeping information safe. But I guess the second part that I would add is protecting not necessarily personal information, but also proprietary information. So as companies use AI and bongo included, putting safeguards in place so that things that are contributed to tools like bongo don't go train the larger models, the larger LLMs, and I think that's a little bit more of the wild west. I can maybe speak to how bongo our approach is. It's called a zero shot approach. So you can kind of think of it as RAM compared to a hard drive. So you an instructional designer or professor submit source material. Again, that chapter seven example, we're using that to calibrate the AI and give it context, but then as soon as you move on, that information is effectively for forgotten. It's not stored or used to train a model or anything. There are some organizations, and I think the problem, to be honest, is a little bit heavier of a problem or more of a complicated, convoluted problem in the corporate learning sector where people are uploading trade secrets, basically.

Holly Owens (18:29):

We're not allowed to use, I work at Amazon, we're not allowed to use, we're not allowed to use chat GPT.

Josh Kamrath (18:34):

Exactly. Yeah. Aren't that's a

Holly Owens (18:39):

Huge no-no, Exactly.

Josh Kamrath (18:42):

That's exactly right. So whether it's GPT or any of the models, our tool, we built it as a microservice so we can actually switch in and out different generative models very quickly because right now, GPT is the most advanced in a lot of capacities, but we've actually worked with the Amazon AI team around some of their models. And with Bard and Gemini, it's a total arms race. And so there's having the flexibility for companies to have the flexibility to swap in and out of different models, but also putting safeguards in place to protect their end user's information against being used in a nefarious way, I think is again, pretty much critical. Yeah,

Holly Owens (19:41):

Absolutely. And also too, I am seeing from your site, and I already started thinking about this when you started talking, is how this helps companies hire people with the skills and proof of that. Anybody can put anything on a resume.

Josh Kamrath (20:00):

I always say

Holly Owens (20:00):

This, I have enormous amounts of student loan debt and I have four different degrees hanging up here on my wall, but not one time has one person at any company asked me

DaNadia Johnson (20:14):

To

Holly Owens (20:14):

See those. They have never asked me. They may have asked at one point in higher education a while back to see transcripts, but they've never asked to see my actual degrees.

Josh Kamrath (20:25):

So when

Holly Owens (20:27):

You're thinking about that in relationship to the skill building and what people can do, how does this tool, because on the corporate side of things, how does that help companies hire people who are truly skilled or sift through what people actually know? Well,

Josh Kamrath (20:44):

So that's essentially bongo's vision is to be the measuring stick of skills. And I've a little bit, again, kind of on my soapbox, but the reason people don't ask it is a little, I've had the same thought of, it's been surprising that people don't ask for validation of a degree, but maybe it's because it's very broad. If you can get very granular on a specific skill or specific cluster of skills and maybe having different levels of that skill and maybe even having different levels of that skill within certain contexts. Being empathetic in a nursing setting is totally different than being empathetic as an IT professional at Amazon.

(21:33):

So same skill empathy or demonstrating empathy, but within context is also very important. And I guess I and bongo as a company, we really think that if we can get really granular and be a kind of final validator for skills where it's more of a currency skills, then become a currency to be used to better your career. Or like you mentioned in the hiring process, making better, more well-informed decisions on hiring someone as opposed to another person that's actionable data that either businesses or higher education institutions can use. And I think that'll push, that could be a pretty big dent in the universe. That

Holly Owens (22:25):

Gap right there, which you're talking about between higher education and the application in the real world

Josh Kamrath (22:31):

Is huge. It's

Holly Owens (22:33):

Still huge, even with all the technology. I can just see, Nadia, I want you to talk about this too. When you're sitting down and you're planning a higher education degree, what are the skills you have the overall program outcomes, and then you have the specific learning objectives of each course. But I know a lot more people nowadays are aligning it with skills and this can come in and help with that.

DaNadia Johnson (23:00):

I think that's why you're seeing unfortunately, lower enrollment because people are going to trade school and focusing more so on specialized skills just because that gap is so wide. I mean essentially. But yeah.

Josh Kamrath (23:17):

Yeah, I think, I guess my hope is both that organizations can make better, more well-informed decisions on how to operate. But I also hope that universities can better, more accurately fine tune their curriculum or their program paths to be more applicable for the real world. Sometimes I wouldn't change my journey, but economics, a lot of it as a degree that was mine. It's super high level, a lot of it, and it's not super tactical. You had mentioned trade schools or vocational, you're taking something super specific and making something tangible. When you learn about economics, it's super macro and high level and not necessarily as obviously applicable to fixing a water heater or something. And having, again, a measuring stick of skills or a tool that can be used to facilitate experiential exercises, IE bongo, I think is just going to lead towards more well refined learning paths that are ultimately more efficient and I think will lead to a lot better learning outcomes.

DaNadia Johnson (24:57):

Yeah, I love that. I think, yeah, I think I'm hoping that that's what happens. Totally. So now we're coming to the end of the episode, so I just want to kind of go back and think about is there anything that we missed? Is there anything else you want to share about bongo, any upcoming things you've got going on? Anything that we missed where we're at now?

Josh Kamrath (25:28):

Yeah, I would say there's lots of really cool upcoming things within long, I don't know how much I should divulge publicly right now. Yeah, I

Holly Owens (25:38):

Was going to say don't violate any, Don. Yeah. You don't have to

Josh Kamrath (25:41):

Give

Holly Owens (25:41):

Us

DaNadia Johnson (25:41):

Much

Holly Owens (25:43):

Us enough to entice us in the audience. Absolutely.

Josh Kamrath (25:46):

Oh, Sure. So

Holly Owens (25:47):

I

Josh Kamrath (25:47):

Think as I mentioned, we had just in October had released this contextualized evaluation and having that continued one right now, it's incredibly capable and powerful, but continuing to have that be more well refined is certainly something we're working on right now incorporating different visual components. So right now most of the AI is what's being verbalized. So organizations are using the teachback methodology, for instance, to have people both demonstrate they have that knowledge, but also allow the AI to pick up on it at scale. So a quickly example there, one, a large institution in Canada has a bunch of nurses go through an evaluation and they basically have to come to that assessment or they're doing it at their homes. They bring a cold chicken breast to the assessments, they see the instructions, they have to make an incision and then demonstrate they can provide or do stitches on that chicken breast,

Holly Owens (27:01):

Oh My goodness, Application. And they

Josh Kamrath (27:04):

Also have to go talk through or articulate the five steps of what they're doing. So the tool both is capturing that they have that skill and it's validating through what's being verbalized at a limited scale. I think they have something like 20,000, 22,000 students a year going through this type of assessment. So at very high scale, it's essentially grading them and checking for accuracy. Still has the human in the loop to come in and validate it as a final marking, but again, just speeds up the process. So we have that as a use case right now in the coming months or quarters. Being able to have some image recognition to identify that and have the evaluation have further validities is something that we we're looking into heavily. Being able to have that type of assessment for a screen capture, we have screen capture. Again, right now the teachback method is the way to enable the ai, but being able to have the AI see where the learners clicking around is going to be a really cool enhancement.

(28:18):

Somebody alluded to it a minute ago, but being able to incorporate or map up to different skills ontologies, there's 30,000 different skills that need to be learned in this domain. Being able to map up our products to those skills to have an organization be able to make internal, maybe not hiring but moving up the corporate ladder decisions or, so there's lots of cool things I would say coming. I would say I've a little bit been bouncing around between a corporate learning use case and a higher ed. As I mentioned, we got our start in higher education. We still have more than 50% of our user population, so millions of students worldwide using bongo to go through an experiential exercise or give a public speaking presentation or practice English as a second language. Again, lots of use cases. So just I would say maybe in closing, if you will, super excited about both what we've been able to accomplish so far, but also again, what AI is really unlocking and enabling from touching as many people as possible standpoint. Yeah.

Holly Owens (29:37):

You're answering the last thing I wanted to ask. So we like to think about the future here at the end and wrap it up the future. So what's the next big thing in EdTech? What do you think is going to revolutionize education? Maybe higher, you can speak to higher education or corporate, or

Josh Kamrath (29:53):

Where do you see stuff going? Yeah, I think with corporate, I think it's having more accountability to what's being taught and being able to map that back to ROI is what's going to be paramount in the next two years. A couple years ago, budgets were, the economy was roaring and budgets were wild. That's snapped back and everybody's, yeah,

Holly Owens (30:23):

Really fast Questioning every dollar spent. So being able to

Josh Kamrath (30:27):

Make a business case and be able to give attribution to tools on an ROI standpoint I think is what's in the immediate future for that side of the market. I think in the education space, I think that you had mentioned student debt. That's just a crazy, it's crazy how much debt student debt there is in especially North America. And I think that universities, maybe it's going to be a little bit of a controversial comment next, but I think the universities are going to be held a little bit more accountable too, to what's being taught and

Holly Owens (31:10):

How much real world applicability

Josh Kamrath (31:13):

There is. And for that domain, and I guess it's an exciting time for bongo. Effectively what we do is help people take knowledge and put it into action, or at least real world application. But I just think that there's going to be more accountability and people are going to question that more. And the universities that can answer those questions and point to results are the ones that are going to rise to the top essentially or continue to propagate.

DaNadia Johnson (31:48):

That's such a good point. I think so. I mean, we spend thousands and thousands of dollars on degrees. We need an ROI just like corporate companies do. Where's my return on investment? So that's such a good point.

Holly Owens (32:02):

Yeah, absolutely. I know there's a lot of people in the audience too who reflect back on their programs. Obviously when have got some positions without the degrees I have, but really the real world is where I learn most of my skills. I'm

DaNadia Johnson (32:17):

Actually talking about that in class right now

Holly Owens (32:20):

Versus formal learning. Very relevant.

DaNadia Johnson (32:22):

We learn in our environment more so than we learn in a formal environment, but should be, they should be together. We should be doing both.

Holly Owens (32:32):

Absolutely. And hopefully there's tools like Bongo that kind of bridge that,

Josh Kamrath (32:35):

Right?

DaNadia Johnson (32:35):

Yes.

Holly Owens (32:35):

Yeah. We need tools like Bongo to bridge those gaps. For sure.

DaNadia Johnson (32:39):

Yeah, for sure. So

Holly Owens (32:40):

Thank you so much Josh, for coming on ed Tech and talking about bongo and sharing your experiences and all your expertise. We really appreciate it.

Josh Kamrath (32:48):

Absolutely, guys. Thank you a ton for having me.

Holly Owens (32:51):

Of course.