Hey, welcome to Top Shelf Tech. We're talking about Disruptive Technology today and as part of our Disruptive Technology series, we are going to unleashing AI. And who better to discuss it with than Asa Cox, Arcanum AI in Wellington New Zealand. So welcome along.

Thanks, Jeremy. It's great to be here. Always good to talk about improving the industry with some good tech.

Yeah, absolutely. And you guys, in particular, are in a very exciting area and one that's still really developing, and I think that's a big part of your business proposition is making it more accessible. Maybe just to kick us off do you wanna give us a bit of a rundown on what it is you guys do and how you are unleashing AI on the world?

Sure. Thanks. Yeah, so we're helping people unleash their awesome. I think COVID has kind of shown the world that good digital experiences are here to stay and are kind of a necessary part of business success. A lot of people have been investing in data and the cloud as you guys know well.

But what we've seen, there's a bit of a challenge to really get the exciting, intelligent features into people's software. And so our platform and our company are focused on getting people from an idea all the way through to the integration of intelligent features, powered by machine learning and AI.

And so how does that work? Do they pick up the phone, call you up and say, Hey, we've got this problem. This is what we want to solve. We've got a few ideas on it. How do you help?

Yeah. At the moment. Yeah. There's kind of that modality. We are working towards a pure platform play where out of the box, people will be able to build their own intelligent workflows, output from API and then kind of integrated into their software.
So that's what our platform is. Yeah, that's the roadmap that we're working towards, but at the moment, yeah, it's a bit more service-oriented upfront. And so we talk with companies about the kind of features that they're hoping to build or integrate, maybe they promise their shareholders or their brochure says powered by AI and hasn't quite gotten there yet.

And so, we're really there to help enable those. So we've got a team of data scientists. We've got expert machine learning engineers. And we kind of bundle that together into infrastructure as code and API and kind of people go into really releasing the intelligent features into their software.

Yeah. I mean, this is a big field, right. And artificial intelligence, and I think it means many different things to many different people. How would you break down what it is that you guys do and with some examples of the engagements you've had and the sort of results that you guys deliver?

Yeah. When we talk about artificial intelligence, you're right. There are the technical definitions and there's what the market understands it to be. And we look at it as processes and functions that humans could not do either all of it or part of it, either at the speed or the scale or the complexity that's needed for the software to really have that level of intelligence.

So we work across a number of different, or kind of all of the domains really of AI and ML. We work on computer vision. And so we were just talking before we came on air about the work that we'd done with New Zealand Rugby, and now we kind of do across multiple sports. And that's taking video and tracking all of the players throughout the field. So generating those X Y coordinate data for deeper analysis, but without having the little devices on the jerseys. We've done other computer vision work for Transpower, for example, identifying damaged power lines from images captured by drones. So kind of the, more than the industrial kind of space.

But then also then work in predictive analytics. Probably the broadest category where we also see the most demand at the moment. And that's been able to plug in a predictive element into an existing business intelligence process. We kind of argued that actually, most people are in business intelligence.

Then they try to get to business insights and actually whilst we call it business intelligence, not many people are actually at the intelligent part of that. And so we're plugging in predictive analytics, into existing BI pipelines. So stuff like customer churn and customer retention. The next best action, categorization, that kind of stuff.

So if we picked up on that example is a bit of a low bar at an entry point. Maybe you've got some data. You want to go get some results out of it. How much off the shelf is this? You know, you just plug it into the platform and it starts telling you insights, submissions. How much do you come up with a hypothesis in sit down and sort of go? We kind of think that these things might be happening in this way. What are the data points we maybe need to track and think about?

Yeah, it's kind of outcome-driven backwards. So when we started looking at, what do you need as a business? What kind of insight are you looking to gain that can help you make a better decision? And whether that be about investment, whether about project allocation, resource allocation, whatever it may be, may be part of your reporting process. We kind of worked backwards from that to go. Okay. Well, what would you like to be able to predict what prediction would have the most impact on your business right now?

And then we can work back from that to go. Okay. What data do you have available? What is your tech stack? What do your pipelines look like? And then our data science team will be able to build the model or we can work with the client and their data insights or data science team to build the model. And then our platform picks it up from there and productionizes it. So it kind of does all the automated scripting, the containerization, API. So then the integration into the BI dashboard or the reporting tool is then super simple to do.

This is still quite out there. You know, they've heard about AI, but they're not quite sure how that applies to their business. But also, how did it get started? Is it, they write a big cheque and they have to somehow make a business case and get ROI? Or they start small with an experiment. What kind of advice would you give to businesses that are just going? I think there might be something here, but how do I actually, from a business perspective approach this and demonstrate that it's something that's going to deliver value to the business.

Yeah, I think this is a really important shift in understanding and perception. I think historically when I say historically tools three years ago, I think including ourselves, we were all saying AI is really hard. You need a team of data scientists it's super complex and super expensive. You need a good budget and so on.

Whereas now actually in order to build momentum, which is kind of what we want as an organization and the industry should also be seeking as momentum for more and more use cases, more and more adoption of machine learning. So what we're doing with our platform is just lowering the bar right down to near zero. Right? So if somebody has already got a data science team, maybe they've done some work with one of the AWS ML services. They actually want to take that next step to move from the proof of concept into production, which is the bit where at the moment, 80% of projects fail at the point of IT integration.

So we basically lowering the bar to say, we will charge you nothing until it's in production. So we've tried to kind of remove all financial barriers and say, look actually with the right skills and the right tech, you can get value in days, not weeks or months.

Yeah. You mentioned cloud services, some of the AWS stuff, you know, Azure has kind of growing the cognitive services as well and other AI pieces, how important is the development to that ecosystem to what you guys are doing in the ability for customers to leap in and get a pretty quick return and not have to make sort of massive investments and building it out themselves?

Yeah, I think for all of us, it's huge. I mean, I think we all, we all see the power and the value of apps like Netflix and Spotify and Uber. We all want those user experiences. We all want the personalization and the adaptation of the smarts behind And so the ecosystem of capabilities is really important.

I think we're focused on trying to stitch that together and make integration into those software applications super easy. So we work with cognitive services and with the AWS ML services to build out those workflows. To enable them to be used really easily because as you kind of alluded to the building it is yourself option is a longer-term project. It takes different, more resources to be able to do that whilst the capability out of the boxes is there with those services. The actual training or configuring of those services all the way through to having the right compute all the way through to having the right kind of integration. You know, that takes a team, it takes a decent-sized budget and a decent amount of time. So I think the ecosystem has the components. And what we're trying to do is pull them all together to get that return on investments.

Yeah. Awesome. One of the probably more out there use cases that I've seen, you mentioned that you do a sport. It was a couple of years ago now at a conference in the US watching is, you know, basically, UFC fighters had this gamified sort of Mortal Kombat timer overlaid. It sort of showed they got hit. What was the damage? What was their fatigue levels like, and that kind of stuff? How far does this stuff go? I mean, how far does this go in terms of creating and packaging up new products and in fundamentally different ways that we're going to engage with content or with data and businesses?

I think massively. We've kind of all seen the demonstrations of VR and AR we kind of thought it'd be great, but how do we get to that point and where's the business value that kind of goes into it. But I think as we begin to pull together data with artificial intelligence, with some of that consumption layer, there's going to those different modes. And I think we will see the reinvention of consumption in some areas, and sport is definitely one of those yeah, through our sports division, Play in the Grey, we're fortunate that along with AWS, we are talking to a number of the major sporting bodies around the world and they are all looking kind of as COVID has impacted their revenue streams and changed the way people want to consume sport. They are looking at reinventing. So I think in the not too distant future, we will see in broadcast and near real-time analytics from computer vision, kind of we will see AR and VR kind of replays and simulations of the sports. So I think we are going to get pretty close to pulling together a more Netflix like experience for sports broadcasts and sports media.

And I think for some that are fantastically exciting, but others, it may be, you know, verging on destroying that love of sport, but in-stadium experience and in-home experience with a mobile experience, you know, we'll at least have the choices of how we want to interact with those kinds of sports.

Yeah. And look, you know, you mentioned Play in the Grey, do you want to just delve into a little bit more about Play in the Grey because I've seen a little bit of what you guys had done myself and, seeing a few awards, and the stuff that you guys have been up to and you know, it looks like some really impressive stuff that you're building from down here in New Zealand and sort of exporting out to the world.

Yeah, that's been super exciting since we started working with New Zealand Rugby for the Japan World Cup. Looking at that high-performance kind of view of competitor analysis. What we've worked on, what we've been building since then is being able to process broadcast video and get that same level of high density, high fidelity analytics data.

So with our AI models, we're able to generate a million points of data per minute. Which is obviously way beyond what humans can generate. And then we're using that to do our event detection. So when did a pass happen? When did a kick happen? When did a tackle happen? And that's being used in a number of different ways, but at the moment is typically somebody sat at a stadium or somebody watching the video, multiple humans, pushing buttons as these things happen.

And so we're automating that process so then teams of different sizes, not just the world leaders will have access to kind of video analytics using the AI that we built here in Wellington.

Yeah. That's awesome. And so, like you say you've been doing the consultant work you've been building the platform. What is the vision for the company and where do you want to take the platform ultimately, what is that there to enable?

We've been talking from the beginning about how AI and ML is a transformational technology. And I think what we've been frustrated about is the fact that people have been bolting it on, you know, around the edges of an organization, but no one has really gone in and said, right, we want to really have it as a foundational tech. We want to build the next generation of our user experiences, our products, our internal systems. And we understand that of course, it is hard to go all in. But we know that those companies that have the digital natives, the ones that I keep referring to are by far in a way, the global leaders, because they have machine learning and AI at the core.

And so for us, we want to and our platform to enable organizations to make that transition. To go on the journey from the kind of legacy siloed, disconnected, non-data driven or not intelligent application ecosystem. So to have something which is more closely resembles the cloud-native experience.

So, we're going to have a platform in which out of the box we'll have a whole bunch of pre-built workflows that are going to enable use cases across document processing, image processing, predictive analytics. We're going to have a very automated process of getting that deployed with the right level of computing. So the high performance is there and then we're going to work towards integrations with those end systems, whether it be Salesforce or other CRMs operational platforms, like monday.com and so on. So then you get the real end to end view of all of the value that machine learning can bring to your business.

So for us, it is about lowering the bar to make it accessible, get a return on investment really quick, and then go on the journey with these companies and support them as they build out machine learning across their organisation.

Awesome. I guess one other question in the back of my mind is this all possible with where we are at with cloud today? And how much has this underpinned by, you mentioned Salesforce and Monday as these kinds of SAS applications, you've mentioned AWS a couple of times, how much for businesses is this about cloud maturity, as we get more and more cloud-native applications, being able to easily plug in and exchange data and get more value out of it?

Important on that cloud adoption journey. Not just because of the data layer. But also when you start doing stuff up, web-scale when you actually push some of these things outside of the organization, into the hands of customers, you need the compute capability, the flexibility of the on-demand nature, the serverless nature of what the cloud can offer.

And so I think that's an incredibly important momentum shift as cloud adoption continues. And I think if you look even further forwards, into the world of 5G and what that's going to end And that means actually having AI at the edge, is then going to become something which is accessible to organizations of all shapes and sizes.

And so what we're trying to do is get people to think about horizon two, horizon three, and instead of thinking that's five years away and millions of dollars. Actually, how can you start the journey right now? Because most of the pieces of the puzzle are already in place. We kind of say, look, all of the world's problems, pretty much solved by technology. It's just a case of the humans getting in the way of getting them in the hands of everybody that needs them.

Awesome. Hey, look, thanks a lot for joining us today. It's been really interesting to hear about all the work that you guys are doing and I guess democratizing the technology, putting it in the hands of many, to have a huge impact.

So. Thanks for joining us and also thanks to Brent Colbert, who's recently joined your team for connecting us up for this slot, for this conversation today.

Really appreciate the opportunity to chat, really interesting. Thanks for the questions, and look forward to working with you guys on bringing some more awesome to New Zealand.

Awesome. Cheers mate. Thanks a lot.