Pitchfest Tech + Data Winner: On3, Pitch by Paul Cardis

This series highlights our Pitchfest finalists. At the 2023 Housing Innovation Summit, we introduced a new program, coined: Pitchfest in order to bring new ideas to light, offer feedback that startups can use to advance their solutions, and introduce these innovators to potential advisors, partners, investors, and clients. A dozen companies made it to the final round, and four came out on top.

In this episode, you'll hear Paul Cardis from On3 give his pitch.

Visit the Summit website.

Transcript

Haley (00:00):

Hello and welcome to the Housing Innovation Alliance podcast. This series highlights our pitchfest finalists at the 2023 Housing Innovation Summit. We introduced a new program coined Pitchfest in order to bring new ideas to light, offer feedback that startups can use to enhance their solutions and introduce these innovators to potential advisors, partners, investors, and clients. A dozen companies made it to the final round and four came out on top. In this episode, you'll hear Paul Curtis from on three give their pitch. They were the winners of the tech and data category to get in touch with them, visit their website and find ways to connect in the description.

Paul Cardis (00:38):

Thank you, Dennis. Great to be here. We're a leading learning and task support technology concerning frontline workers. Our idea is those who work on their feet or work out of a truck are really underserved by technology in terms of what they do, their job tasks. All of that really isn't well documented. In fact, we're still doing things the way we did a hundred plus years ago, which is basically apprenticeship training and live training. Our executive team is made up of folks throughout the industry and so we come with many decades of experience, understanding, home building, understanding construction, and really put together a platform here to solve a major problem. We were named the 2022 most innovative product at the International Builders Show. So we're very honored with that award. And I'm gonna be sharing with you our V2 iteration of this product. Basically our learning model is how do we develop frontline learning solutions to tie into what Bronco was saying, the huge need we have for training, but also how do we train people on the job?

Paul Cardis (01:32):

How do we support them in the job? And we take a micro-learning approach. How do we capture the 30 seconds of how to do this task or the 45 seconds of how to do another task and stack up hundreds to thousands of these little mini essential modules and creating a platform to serve that. And we use AI to serve it up. So basically we can pulse employees on a Monday morning. They get three questions regarding safety and for those who get all the questions right, they're able to go about their day fine for those who didn't get 'em. All right. The AI system will detect what areas of weakness that employee needs some remediation. So it'll program, you need to watch this video or watch that video and it can do this at scale. So we can manage thousands to tens of thousands to millions of employees effectively.

Paul Cardis (02:15):

AI creating a customized learning path for each employee. This is very important because it's one thing to do training, hang a certificate on the wall, but most people, in fact the data shows we, we forget 90% of what we learned. So there's a big problem there and that they're not retaining it. So we need to process continual improvement as a, as a buzzword. Our system actually makes that happen. By leveraging ai, we also use AI for our V2 iteration, which is utilizing this device here to provide field assistance so they can get live support. So they can say, Hey on three, how many anchor bolts are needed in this location on the foundation? Or what's the nail pattern on the sheathing? And be able to get answers real time that is exact to their location and to the job. So this is where we're headed.

Paul Cardis (02:57):

We're utilizing a variety of different hardware. In fact, we've got now two hardware providers that give us this capability to do kind of like the Avengers. This is actually a monitor where I can actually see it's also a camera. So we're using AI vision to detect what the employee is seeing. So this is our new iteration that we're raising funds for to complete the model. So we're doing everything in terms of object recognition of sinks, object recognition of drywall, sheathing, hardware components, was talking to Simpson, strong-tie and MiTek, all these components. If we can get this to recognize, then in my eye piece, I could see, ah, would you like instructions on how to attach the hurricane clip? Yes. And I can watch a quick video on how to install the hurricane clip. We're also engaging in safety as well, which is a big deal because safety's a big deal in our industry.

Paul Cardis (03:43):

And so we are also gonna recognize danger zones. For example, if there's a trench over here, red right light will go in my eye indicating there is a trench and I need to be careful. Or I could maybe watch a video on trench safety on how to properly work inside a trench with proper safety cage. We're also getting into inspections, which is really exciting. So this is where we can record an entire inspection process through the camera ai recognizing and identifying potential quality flaws. But moreover, providing assist to the employee to identify and follow a process of inspection. What's also great is we're gonna record the whole thing so that if a manager wanted to come back later, could replay the tape and watch the actual inspection as it was happening in the house. So we're really integrating a number of technologies to really change the way we build homes today.

Paul Cardis (04:30):

Also, procedures are gonna be very important following order of operations. Step one, step two, step three. We can present each step inside the lens and provide that instructional support. And we're able to do client-driven content as well. So clients can come to us and say, Hey, I want to do this specific job. This is the order. Make it happen. And I want you to recognize when these things are happening in the environment and we can make that happen. Our roadmap is to continue to develop. We've been building basically a base of knowledge using mobile based applications that run on people's cell phones. And we've been amassing lots of modules. We're gonna be adding the AI vision later this year. We're already starting on those models right now and it's going very well in terms of what it can recognize. And then lastly, we're gonna be adding in our NLP chat bot. So we've heard a lot about chat. G P T, we'll likely plug into it or one of its competitors to do the final delivery. We're, I'll literally be able to talk to this device and it'll be able to serve up the various different assets. We've grown quite a bit. We raised 1.4 million to date filing with the S E C for a crowdfund and expecting to close in June, 2023. So I appreciate your time and I'll take your questions.

Speaker 3 (05:42):

I have a question for you. It's very interesting how you can serve things up through the AI functionality and we all, I think when you're, when you're doing something and you search around online and you find some valuable assets in the moment is great, but what are you doing to then deliver those assets to other people on the team who might not be directly involved in that process?

Paul Cardis (06:01):

Yeah, so we have a program called Trade Share because one of the biggest challenges is not just educating a builder's employees but actually penetrating the trade base, which is 80% of the workforce. So we have trade share where a project superintendent can forward either airdrop, do a text message or email any asset to that trade contractor in the moment. Follow

Speaker 3 (06:21):

Up to that. So when you were saying you'd get things prompted to you mm-hmm. <Affirmative> and you'd be able to capture that and share the video mm-hmm. <Affirmative>, do you also then automatically share the assets that came up there? Because that's still a lot of work for somebody to go back and track what are the things that they wanna push?

Paul Cardis (06:36):

We're already setting up packages that are going inside ERPs. So we're working with hyphen to insert that into a schedule. So whenever they're doing a order in, it will come an on three package of various different key moments that they want associated with that order. So we're now opening that door and we're also working with ECI to also do a similar integration with them. How

Speaker 4 (06:56):

Many headsets do you have in the field? And I guess the follow up would be, is hardware the revenue model or is it software content? All

Paul Cardis (07:02):

Software. Okay. Yeah, we're working with two providers. This is View six is one provider. We just signed a deal with Epson and and we're gonna be adding a third one as well. So we're really hardware agnostic. It's the software that runs inside that you talk to on three, it's an app and just like you would on your phone, it's, it's just another app. Just to

Speaker 5 (07:19):

Clarify and then I have a question, but the, so the, the hardware is not on three, it's hardware. So you would basically sell on three as a software and they would purchase then the helmet? Correct.

Paul Cardis (07:27):

And now we're becoming agnostic to hardware. We're expecting Apple to release their headset this summer. That's the expectation. So it'll be interesting to see what Apple comes out with. It's gonna be likely glasses, but it will be lower cost than some of these ruggedized units.

Speaker 5 (07:41):

My other question was on the content development of the training. Who are the developers and what kind of, how do you get all those assets?

Paul Cardis (07:46):

Our clients are actually the, the developers as well as we also have a development team that's making courses and we're partnering with building product manufacturers to create courses and building science experts as well to add courses to the library. But our clients today, actually the app itself captures just as much as it delivers. So I can say, hey, on three record and it starts recording a process in the field, which then is automatically sent up to our servers and then can be vetted by the company and then turned into a learning program. So we see learning as not just a one-time event. It's very continuous cuz as they see problems, they capture those problems in video, we develop learning and, and a questionnaire and a full assessment around that and now we can roll it out to the entire team, make sure that everybody's on the same page to not do that going forward.

Speaker 5 (08:30):

On the pricing model for the consumer, does that vary based on the training that is provided? I, obviously you have expenses for developing the training, but whatever there is on their menu, is it like you buy up on your subscription based on what's included in the training or you just subscribe to the on three which,

Paul Cardis (08:45):

So with our clients, we try to make it very simple. They buy an enterprise package and that allows them a ton of flexibility within the package. So they can create content, they can pull from our library. They're not nickled and dimed. They come all in and we see it as, like I said, a real continuous improvement program that they're constantly adding new programs, they're constantly engaging, they're using the AI to make sure all their employees are at competency level cuz they do dip and then we're able to make sure they're operating at 80% or better knowledge retention, which is very powerful cuz we're finding that the industry is very low in retention. So we see this as an ongoing platform and and platform fee.

Speaker 6 (09:19):

Well thank you for the presentation. So I think this idea with the integrating with the AI models is a very smart one. So you could plug in with like philanthropic or chat g, pt, open AI or hugging face one of those. Right, right. So I was wondering like can you please talk more about your engineering team? Who's your CTO and who are engineers on your team?

Paul Cardis (09:36):

Yes, so we are developed, we're working with a number of partners around the globe. So we have a Canadian team, we have a European team, we have a Philippine team and we're leveraging different groups to different parts of our software. So right now AI is being driven by our European team. So I have a developer in France, he's acting as our cto, but post-funding, we're actually gonna hire cto. I'm playing CTO as it is right now, but that's one of our objectives to fill that role. Well it's insource, we're outsourcing it, but it's also, these are gonna be FTEs with us. So we, we take up 80 to 90% of their time today post-funding. We plan on bringing 'em on board full-time. We have five engineers right now.

Speaker 3 (10:10):

One of the biggest challenges we face is communications and language barriers. So what do you guys do on that front?

Paul Cardis (10:17):

We are bilingual right now everything is is English, Spanish and we're adding 64 other languages for any asset goes through an auto translation process and we'll be available in native language. So we're very excited to expand to the larger language base cuz we've had requests for Russian and Portuguese, Korean. Those are some of the big requests right now that we have to fulfill. What's your biggest pain point right now? Biggest pain point? Interesting. I think it is getting adoption right now. I think the industry is not used to capturing things in video. So actually getting our clients to follow a process. Normally they just let a problem happen and they don't document it in video. So that's a whole new thing we have to rewire in the industry, whether it's using your cell phone or if you happen to have one of the headsets, it's great. But right now just getting people to crack out that phone, take a picture, take a video when they see something isn't right so that we can develop a program on it. That flywheel's just starting to turn right now. Right now we're targeting residential production builders, but we've also expanded into commercial and we're gonna continue to expand out throughout the remodeling and custom builder space. We're very active there and we've actually been even outside of construction now in the medical and hospitality space to do a very similar thing. Is there a live

Speaker 7 (11:22):

Person interacting with the person wearing the augmented reality?

Paul Cardis (11:25):

We do have a live mode and so we have several partners that can work through our headset and our software to do on three live, which is in beta right now. So they can dial up an SME and be able to have 'em live, connect in their transaction, whether it be their cell phone or through the headset. This lasts for about five hours as is without the extended battery. You

Speaker 5 (11:43):

Have a few customers that have been using on three. Do you have any data on how it's been impacting their workforce?

Paul Cardis (11:48):

Yes, we do. Yeah, we're working with several of the large public builders today and in fact we just are in in reviews with both of them. They're all expanding contracts with us right now. Thank you.

Haley (12:03):

Thank you for listening to the Housing Innovation Alliance podcast. We invite you to learn more about pitchfest and let us know if you are interested in participating in the next cohort@housinginnovationsummit.com.