[Voiceover]: I saw one retailer describe Black AI as it's like Google for cctv. My problem working from home by myself all the time. Just ChatGPT is my only friend and
Justin Williams: as I walked into this business and yes, they're engaged with Walmart. Who in their right mind would beat a path to the world's largest retailer and say, ha, I'm going to start here.
[Voiceover]: Welcome to ADD to Cart, Australia's leading e commerce podcast that express delivers all you need to know in the fast moving world of online online retail. Here's your host, Bushy.
Nathan Bush: Hello and welcome back to ADD to cart. My name is Nathan Bush, otherwise known as Bushy, joining you from the land of the terrible people here in Brisbane, Australia. Now, as you would have seen from
[Voiceover]: recent episodes of Add to Cart, there is definitely a trend towards D2C brands moving in store.
Nathan Bush: So for physical retail is becoming really
[Voiceover]: important for all of us in e commerce.
Nathan Bush: But there are so many parts of
[Voiceover]: physical retail that we might not just
Nathan Bush: think about even as we're expanding into that area. And one area that I'd never thought
[Voiceover]: too much about was security cameras. I mean, we see them everywhere, but how do they actually work? What data can we get from them? What's the future of security cameras? We're going to dive into that today. Not an area that I thought we
Nathan Bush: would, but this is a fascinating conversation to understand how modern AI technology can
[Voiceover]: transform and the types of intelligence we can get out of security cameras in store.
Nathan Bush: Now imagine if your store security cameras
[Voiceover]: could be used not only to catch thieves, which is what they were built
Nathan Bush: to do, but also unlock the secrets
[Voiceover]: of how customers navigate and engage with your space in real time.
Nathan Bush: Sound like Black magic? Well, that's the magic of Black AI and today we have Justin Williams here
[Voiceover]: to tell us all about it.
Nathan Bush: Black AI is an Australian spatial AI startup that is transforming how retailers use
[Voiceover]: in store camera systems by integrating AI with real time video that already exists in most stores. Black AI is helping Aussie retailers and some big international ones analyze human behavior
Nathan Bush: patterns, prevent theft, reduce losses and make
[Voiceover]: smarter decisions that boost both customer experience and revenue.
Nathan Bush: Our guest today is Justin Williams and
[Voiceover]: I am lucky enough to have known
Nathan Bush: Justin for a very long time. He is a multinational corporate tech leader who crossed over to early stage Aussie
[Voiceover]: businesses in 2016 in the startup space
Nathan Bush: and is currently steering the ship with Black AI. With years of experience working with major retailers such as Woolworths and Walmart, Justin is now working as Chief Customer Officer at Black AI, harnessing AI to reshape
[Voiceover]: this world of security.
Nathan Bush: We chatted about how AI driven journey mapping and analytics can turn old CCTV
[Voiceover]: systems into powerful customer insight tools.
Nathan Bush: We talk about the role AI is playing in loss preventions and all the ethical implications and privacy implications that come with that.
[Voiceover]: We also discuss why physical stores are
Nathan Bush: making a comeback and we draw upon
[Voiceover]: Justin's knowledge from from years and years of retail experience.
Nathan Bush: And we talk about the tech that
[Voiceover]: can help you capitalize on this opportunity.
Nathan Bush: If you haven't already, I would love to invite you to sign up for
[Voiceover]: my weekly e commerce newsletter.
Nathan Bush: Every Tuesday morning I send an opinion piece on something that's happening in the industry as well as five must read
[Voiceover]: e commerce stories, research pieces or case studies from the past week.
Nathan Bush: It's a great way to stay informed and inspired in e commerce.
[Voiceover]: You can sign up for free free
Nathan Bush: at add to cart.com subscribe as always, a huge shout out to our partners Shopify plus and deliver in person for
[Voiceover]: their continued support of Add to cart. We really appreciate it.
Nathan Bush: Now let's jump straight into our chat
[Voiceover]: with Justin Williams, Chief Customer Officer at Black AI.
Nathan Bush: Justin, welcome to ADD to cart.
Justin Williams: Nathan, thank you very much. I'm so excited to be here.
[Voiceover]: It is nice to see you. It's been a long time since we've actually sat down and had a face to face. We're on the phone a little bit. But before we get into our backstory and our true romance together, I want to talk about Black AI because when you went over there, I had no idea what you were doing. It was this, this mysterious company that I had no idea what was going on. And I think in my research I saw one retailer describe Black AI as it's like Google for cctv. What did they mean by that?
Justin Williams: So this particular retailer, they're a fairly large national retailer in Australia. The problem that they and many retailers have is that there's a lot of CCTV cameras, there's a lot of CCTV footage, but to be able to search that footage for a specific event or a specific thing is almost impossible. And so the way this retailer sees our platform, it enables for the first time the ability to conduct targeted search.
[Voiceover]: Okay, so set it up for us here because we will have a lot of e commerce listeners here, some who may have physical stores or are thinking about physical stores. Obviously, security cameras aren't the first thing you think about when you're expanding into retail. How are stores normally set up from a camera perspective and a security perspective even before the Black AI solution was around?
Justin Williams: I'll look, security cameras, as your listeners would, Nathan, they are not new, right? Like all the way back to the old analog cameras. And talk about stores changing, you know, I mean for me now it's like every month the interior of a store changes. But those cameras, you know, some of them very old, have been fixed for a long time. So the cameras are the, they are owned by the security or usually the loss prevention team. There's, there's profit protection or some label of that team. And those cameras may be IP enabled or they may be analog interfacing to an encoder that, you know, outputs an IP signal. Many and varied would be my response to that, Nathan. And often suboptimal.
[Voiceover]: Okay, cool. All right. We are going to unpack the Black AI solution shortly because from what you've told me about it so far, it is fascinating. There's so much to go into from both a technical but also a customer analytics to also ethical and safety concerns. So we're going to unpack all of that with you. But I want to take us back a bit because you're not new into the retail game. We met when we were at Super Retail Group. Well, I was at Super Retail Group, you're at SAP. Life looks very different for us both now. We're sitting here both in T shirts, video conferencing from our homes. We're a long way from that corporate world. Tell us about your journey in retail.
Justin Williams: Yeah, sure, Nathan. I'm originally a Sydney boy with my family and I got asked by SAP to take on the role of what was then a retail regional general manager role and to move my family down to Melbourne. Melbourne, of course, home to many of Australia's leading retailers including a number of the Westfarmers group. So yeah, Bunnings, Kmart and you know, we've had likes of Coles once upon a time in the Westfarmers stable, you know, so too Meyer and others. So for me at SAP, I didn't intentionally get close to retail, but I was the executive sponsor on a number of key projects for large retailers. And then since that time, quite by accident, not by design, whether it was running a last mile delivery platform back in the day and then working with Australia Post to launch Shipster in 2017, 2018. I've always been in this retail community on the technology side of it. So yeah, I'd love to claim credit for there being some brilliant design that I followed but yeah, it's just turned out this way and it's just so fascinating like the evolution of E Com and the Dominance that E Comm showed to in store retail and then now post Covid to see the in store retailers and their resurgence and redevelopment, it's really exciting.
[Voiceover]: Yeah, it is a resurgence for sure. We're seeing, you know, how powerful that is. We had announcements for on the ASX for last fin year and there was definitely a pattern that those with stores were in a better place from a profitability perspective. So that resurgence is absolutely real. But from your perspective, I really enjoyed watching your movements post SAP because you really immersed yourself in the startup world and for a long time there I was like, what are you doing? Like you were touching so many different spaces. Like you said, fulfillment spaces, tech spaces, you're, you know, having a look around at everything. And I was like, there's something good that's going to come out of this. There's something big. You are going to discover something. So when you put up the post saying that you joined Black AI and your quote was, I did not expect to be here. Tell me about the moment that from all the startups that you were speaking to in that space that you went, this is the one for me.
Justin Williams: Yeah, I got the referral September last year from a former colleague of mine at SAP. And you know, as I usually do, I go and meet the CEO, I meet the team, I get close to what their challenge is and during that I figure out, well, you know, is there compatibility here or not? And in particular have a look at, you know, what their platform is. Right. It's technology and it may do a number of things. And for me, I stepped into Black AI, it was the middle of December and I committed to the CEO that I do four days a week out until the end of March. That was, that was the commitment. And in January, I worked four days a week with one of my colleagues at Black AI and Walmart in the US and a lot of early morning conference calls, a lot of IT native on the platform itself. And I've got to say, it was before the end of January. And I emailed the CEO and I said, look, you've passed my due diligence, I'm good to go. I don't need to wait until the 31st of March. I've actually seen the tech work and solve really, really complex problems. So yeah, for me, the performance and the application of the platform is truly remarkable.
[Voiceover]: And was it what you saw in the technology that got you excited or the reaction that you were seeing from some very large retailers that got you excited?
Justin Williams: Yeah, that's a really good question for me. Having seen quite a bit in the eight years or so since I left SAP, I've tried to not listen to myself. Getting excited about this. A bit like my wife, she does not like to listen to me.
[Voiceover]: My problem, working from home by myself, myself all the time. JPT is my only friend.
Justin Williams: Yeah, correct, yeah. For me, I, I listen very closely to the Walmart stakeholders and their feedback on the platform. And I have connected with a number of people, you know, Woolworths here in Australia who have engaged with black AI and people I know who are experts in this and without pump priming them, I've listened to their feedback and their questions and yeah, this is one that's got serious legs. So I'm, I'm excited.
[Voiceover]: I can tell your eyes just lit up. So let's dive into the solution because I think most people will be like, all right, give us the details. How does it work? So can you talk us through? We kind of set it up at the start with most retailers have security cameras. Some are analog, some are digital. They're all in totally different states. You're dealing with large retailers both here and globally, who could I imagine even have maybe different setups per store, per state, per territory. There would be nothing uniform about it. How does black AI come in and help those retailers with such diverse setups?
Justin Williams: The first starting point of all of it, like a good design thinking assignment, is what is the problem that you're setting out to solve and being really clear on what that is, that's fundamental. But then taking a small step to test the feasibility of, you know, applying the platform to helping solve that problem, that is key. And so, for example, the use case that is the subject of interest of a retailer, we get deployed into their site, an appliance, it's a special GPU based appliance that's actually quite expensive compared to your ordinary CPU server. And this appliance is connected to the video streams of the retailer and those video streams need to be RTSP format. So for the nerds out there, that's real time streaming protocol and there is a minimum resolution that is required for the platform to accurately measure.
[Voiceover]: And are you installing these per site or is it one installation for, for an organization?
Justin Williams: This is per site. So it is, you would call it quite infrastructure heavy because it's a dedicated appliance. We call it per site and they
[Voiceover]: have to be on site because it's dealing with real time information and a lot of information.
Justin Williams: Correct. We are talking metadata per frame, per second, per camera. And so that is a really heavy data load. And normally the broadband at the store is not set up for that. But also there is the anonymization and the protection of privacy that should occur within the store's physical site before that data is exported for further processing off site.
[Voiceover]: Okay, and so you've installed the hardware from a retailer's perspective, then what kind of data are they getting? How do they access it? What does that look like?
Justin Williams: So we've installed the appliance, we've checked all the fields of view of the cameras to make sure that those zones are relevant for the use case measurement. And all of this data, as you've rightly said, it, gets put up in a web portal that is directly accessible by. By. We hand over user logins for the retailers key staff, we would call it, who are engaged in this exercise. And what they have is a dashboard. And that dashboard has been designed to measure the particular features and criteria of that use case. And in addition to that, there is a. What's called a segment builder. And the segment builder is finding those videos that match the specification of the use case. So that enables the retailer to click on those videos and see not just, you know, the customer compliant with that use case measurement, but the customer from when they were first picked up through the entrance store camera, anonymously and safely. But as they travel through the store and they meet the criteria of that use case. So both data points, the dashboard, firstly for how many people per second per
[Voiceover]: day to typical store counter stuff.
Justin Williams: Yeah, really detailed store count though, and I'm happy to go into that. The ability to measure is much more complex or much more. Sophisticated is a better word than simple store counters.
[Voiceover]: When you say sophisticated, are you breaking down into demos, into profiles of the types of customers coming into store?
Justin Williams: So we don't look at the types of customers coming into store. Our platform does measure human behavior, right. Anonymously, safely, no publicly identifiable information recorded, etc. But what we do is we go back to that use case. The retailer has said they're really interested in us measuring for them. And we are able to say how many people, yes, entered that product zone, how many people actually dwelt to read that sign or look at that advertisement? And then of those people, how many actually interacted with the product and more importantly, where they then went on from there. So in the product zone itself, now they've been exposed to that retail media or whatever other stimuli is there. What did they actually do when they were in the product zone itself? Did they reach for a competitor item or did they choose that particular product that was advertised to them. So there's a lot more detail than is normally provided by store counters and things.
[Voiceover]: And what are the most common use cases there? You gave some great examples of. Hey, when you're in this section or you're looking at this signage, what are retailers briefing you on that they really want to uncover about their in store customers?
Justin Williams: Initially, the focus for us has been shrink, right? Otherwise known as loss prevention, otherwise known as shoplifting, theft, etc. That is a burning or fraud. Bastards, those two.
[Voiceover]: Yeah.
Justin Williams: So since the cost of living crisis going back a couple of years now, this has become an increasing problem for retailers. In addition to that, it's coincided with the express lane for shoplifters, otherwise known as the self service checkout. And that has become a real source of concern for retailers. So that theft, and it's not just, you know, we need to stop people stealing stuff, but they actually need to research and understand the behaviors of that theft itself in order to redesign to prevent. Because prevention is the ultimate cure. And you know, going and stopping people for a bag check, it's a high friction exercise that may on occasion put at risk both the customer as well as the team member. So we want to avoid all of that and understand the behavior better. And that's what our platform helps the retailer do. In addition to that, Nathan, we are able to serve up a notification in real time to the retailer once those use case criteria have been satisfied. And that's what I've been involved in this year working around and just remarkable results is all that I'll say.
[Voiceover]: And so when you talk about serving up real time notifications to the retailer, is this like in my head? And this is me being backwards Queenslander, in my head this is like a bunker, like in the nrl. It's like you've got all these screens up and you're getting notifications of what's coming up. Is it like that from a retail perspective that it all comes back to one central source or are you feeding it directly to the team that are on the floor in those stores?
Justin Williams: So we're in the process of currently architecting how that alert or notification is served and to whom to make sure that we respect the privacy and protect the safety of customers and team members like that. That's, that's your starting point. But as you say, currently all of those security cameras feed into a room and there are multiple monitors that surface that camera event. It's very difficult for someone to follow all of them at one time. And none of the cameras show a continuous stream of you know, this is Justin Williams journey through the store. So now that we've got the ability to satisfy the retailer that we can validate the use case in real time and we can flag a notification for them, we really carefully thinking through how we surface that ultimately you know for me prevention is the cure on this one right? And the safety and privacy of team members and the customers is absolutely paramount.
Nathan Bush: When you think of your last E commerce experience that went wrong, how did it go wrong? Was it because there wasn't enough product images? Was it an out of whack mobile experience? Was it that one rogue negative review?
[Voiceover]: I doubt it.
Nathan Bush: I bet it was because of your last mile experience and whether your product turned up on time and undamaged. That's what we remember. Delivered in full on time or die fot for the sexy shortened version is a key metric for retailers and customers.
[Voiceover]: Why?
Nathan Bush: Because it's a critical indicator of reliable on time delivery and the post purchase customer experience. However, many Australian retailers are settling for 70 to 80% delivered in full on time results. That's a lot of unhappy customers. That's why brands such as July, Samsonite and the Party People turn to deliver in person. With an average dyfot score of 99.6% you are delivering experiences to remember. Do not settle for less. Find out more@deliverinperson.com
[Voiceover]: look, it's obviously a really hot space at the moment. You mentioned the self checkout cameras earlier and obviously a lot of contention here in Australia and I'm imagining imagining globally as well around, you know it's pretty confronting when you're checking out and you're trying to work out is this an Idaho potato or is this a. I don't even know other names of potatoes but you, you know I find them so stressful. But then you've got yourself looking at you, you looking stressful and you're like I don't look like that. But then you obviously do when you're trying to work out what a potato looks like. Tell me about that because that would be a very contentious issue right now, especially for customers. That whole surveillance culture and being under the microscope. I know you said being anonymous and making sure that the data isn't personally identifiable is really important. How do you do that?
Justin Williams: Sure it's a fantastic question and there are a number of players in what we call Computer Vision AI and I will just speak to Black AI's approach to it. As I said that appliance installed in the store that conducts the transformation of the video content to pixelate your publicly identifiable information. And so black AI for those listening,
[Voiceover]: you just circled your face. So obviously identifying facial features, it's really important to blur them because you don't want to capture them. Because that is. Yeah.
Justin Williams: And you're blurring that in the retailer's environment before that is exported into the cloud for further processing. That's the first and foremost thing. Secondly, the platform black AI focuses on measuring what we call soft biometrics. So for example, like you and I wearing our black T shirts today, that's a thing that we are, you know, Caucasian and light skinned, both dark hair. I think yours is gelled, mine is not.
Nathan Bush: I think what you're saying here is
[Voiceover]: we're stereotypical middle aged dudes in retail tech.
Justin Williams: Well, I've got the glasses on that your listeners can't see, but I've got the mustache. So take that, I'll raise you one. So moving on, soft biometrics, is it. So it's the color of the shirt, it's, you know, are you wearing shorts or long pants? And if so, what color are those? And we assign a number to that unique mix that represents, let's call it the Nathan Bush look. And Nathan Bush is that way separated from everybody else per frame per second in that camera view. And then as you go to a new camera again, the same soft biometric measurement is conducted. And if it is found to be a close enough match to a previous one, highly unlikely that it would not be somebody else. We pair those up and that's how we stitch together the unique journey that is Nathan Bush's store journey as well. The additional safety element to this is that you are assigned a unique anonymous alphanumetric just for you for that session. So once you leave the store and it's passing through what's called an end zone, that session is like deleted, effectively archived only. And if you re enter the store, that will commence with a new session and a new ID is applied. And there is no correlation between those IDs even though you are the same, same Nathan Bush.
[Voiceover]: That must be tempting though, to go down the path of a return customer. Lifetime value. Stitch that together, wouldn't it?
Justin Williams: No doubt there is, you know, significant interest and retailers have spoken to us about. Could we for example, measure that Nathan Bush visits one of our stores in a shopping center and then goes into another one. And for the reasons of privacy and security at the moment, our platform is not designed to do that. Could it go there? Possibly, yes. Technically it's Feasible. But at the moment, as it's been designed, we're all about privacy and safety. And this actually comes from the CEO co founder and the CTO co founder of the business. They're a lot younger than you and I and they back in 2018, very, very conscious of privacy and security long before the 2022 expose of choice of a couple of major national retailers. And so it's well set into the architecture of the platform to protect the privacy of the individual and to protect the safety of customers and staff.
[Voiceover]: And when you are challenged by retailers around potential new feature sets or even internally thinking up new use cases for this technology that you've got, I can imagine you come up against a number of ethical dilemmas or decisions that you have to make around privacy. Because privacy in this space is changing so quickly and because you are operating across borders is different in the US as it is in the uk, as it is in Europe as it is in Australia. How do you set a framework around how far you will go with privacy and security?
Justin Williams: For us, engaging with Walmart, it required the business to become ISO certified. 27001 is the security certification. And that took the best part of, I want to say, six to nine months. And I recall this from 2223 before I joined the company. But that was a prerequisite that Walmart had for us to successfully contract and participate with them. And those sorts of international standards which are commonly referred to, that for me is just that's the table stakes for it. So there are a number of privacy legislations that are nationally specific and in some cases to states. And so it is that, you know, we know that the Privacy Commission has recommended drafts to the Australian privacy legislation that will be released or is about to be released or is under consideration at the moment. And we expect those to be a lot more stringent and more strict. So first eye for us is to be ahead of the game, as in be able to easily satisfy the requirements of ISO 27001, which we have done, but constantly on the watch out because for us the enforcement of privacy will only become more problematic, which it should right to protect the identity of individuals.
[Voiceover]: I want to take a slightly different direction now. You mentioned the use cases there around fraud prevention, which makes a lot of sense because that's the original use case of security cameras essentially is fraud prevention. You've just superpowered that to be able to make it easily accessible and intelligent. Your retailers who come on board with fraud prevention, are they finding any other interesting pieces of information, especially around customer behavior in store.
Justin Williams: Absolutely. Like, if I call it cx, right? Like you'll feel customers would know what CX is. And that was more or less a curiosity of a particular retailer. And they have, you know, what we know as NCAP displays that they have specials, promotional areas. And there's a famous video moment where, for example, this, like a full family, and at this display, there happened to be, I think, table and chairs. And this family decided to seat themselves at this retail display and well and truly tested the product. So, yeah, there are some really unusual insights that we reveal as part of this. And typically, what has been observed by a store manager or is observable by store team, you know, teammates or associates, those things often take a life of their own where they are generalized and they become, you know, a bit of an urban myth. Like, this is a thing. And for example, we had. We conducted measurement for the use case for this particular customer, and we said to them, well, actually, your use case is not a theme, because in fact, what these people are doing is not coming here and queuing. They are coming, checking on the availability at the counter, and then they're going away and shopping and coming back and then getting served. Whereas there was this perception that there was a queuing behavior that was measurable, et cetera. So there's a few, what I'd say, tightly held urban myths within a business. And those businesses are very surprised when they see the actual data of what is.
[Voiceover]: Can you go and install this, maybe a free trial just at Kmart, and just prove to them that that the middle of the store checkout is just a shit experience? Can you do that for me, please?
Justin Williams: Thank you for the suggestion, Nathan. I'll take that under advisement, and, yeah, I'll see if we can help with that one. But truly, there is, I think that the problem that we solve, and I think you asked this originally, that the problem that we solve is the current methods of measurement, be it direct observation, be it lasers, be it beacons, whatever they are, those things only give part of the story, right? And what I've seen, at best it'd be 20, 25%. And some of those things fuel the perception that a particular phenomenon is like, you know, nationwide, and it's a massive problem. And when actually you deal with the facts and you rely on the data and you see the videos that support that data, you're like, oh, wow. Well, actually, it is either maybe it's bigger than we thought or maybe it's not a thing. So it's wonderful to Be a seeker of the truth and just present the data as it is.
[Voiceover]: Yes, exactly. And on that, have you had any insights or is there any connection, any way to connect the, the in store insights that you are gathering for retailers to the online or the digital experience?
Justin Williams: We've thought about this and I would say it's early stage discussion, but thinking about the behaviors and thinking about the insights of the in store and how that can be applied to online and that's a really good hill to climb I think because E commerce has certainly led the way as far as user analytics are concerned. Knowing the page by page view, following the, you know, the, the mouse or cursor of the individual and, and doing that. Whereas stores have been relatively blind. Right. They've had a store surveyor with a clicker or there's been some kind of measurement device that beacons were meant to be.
[Voiceover]: Our savior. The ultimate attribution tool.
Justin Williams: Several hours of data that is then generalized to a full year that leads to a design that actually creates more friction than it was set out to solve for. So for Ecom in reverse to leverage some of the insights of behaviors and to either equip the retailer to query those behaviors online better or new tools that may be created that allow the measurement of that behavior. I think it's a very interesting future ahead for Ecom because technically, obviously it leads the in store massively. And in stores have legacy pause. Legacy cameras, legacy, all sorts. And that's what they're overhauling. So I think the lessons learned for Ecom will be very exciting in the next few years.
[Voiceover]: Yeah, speaking of legacy, just looping back to the start of our conversation question that I didn't get to there. Can you plug into any camera system?
Nathan Bush: One of my core childhood memories was buying my first Tac Discman. I freaking loved that thing. That is until all my friends got the Sony Discman. Then my old Skipman, well, it didn't seem so cool. And so it is with e commerce platforms. Did you know that well over 100,000 Australian businesses are now on Shopify and they power over 25% of all E commerce here. All the cool kids are there too. JB hi Fi High Smile, Tiger Lily, lskd, Patagonia, Princess Polly, July Age Culture Kings. They've got Shopify plus and it's easy to see why. The Shopify plus checkout converts 36% better than other checkouts. It's customizable. It has D2C and B2B capability. It has international expansion stores. And that's just the Start. Don't settle for a platform that skips. Check out shopify.com auplus and get in touch with the Aussie Shopify plus team to see how Shopify plus can power your business without missing a beat.
Justin Williams: I think one of the failures of the computer vision AI market generally is not setting out the prerequisites or the compatibility requirements to enable this magical tool to measure literally everything. So as I said, the analog cameras, they're not giving us IP based RTSP feeds and those are not feasible even if they're interfacing with an encoder. We need to test that specific manufacturer's encoder model and to determine the resolution that encoder produces, etc. And so that is problematic. Another camera type which presents some challenges for us, what we call pan tilt zoom cameras. These are the ones that are joystick controlled and so they can obviously pan for a different field of view and they can zoom in and out. And if that is used, that kind of breaches the original design that we use for measurement within the platform. So they can be used, but they present real challenges. So what you're saying to someone is actually that pan tilt zoom camera, it's no longer a pan tilt zoom, so there are some constraints and you know, I think there's just a misconception that, oh, you can plug this magical box in and every camera will suddenly light up and we'll be able to see everything. That's not often the case.
[Voiceover]: No, that makes a lot of sense from a commercial perspective, obviously rolling out some big deals. So I can imagine that there's not some sort of website where you can select the small, medium or large package. What's the commercial model for black AI?
Justin Williams: Well, it's strange you said small, medium, large package because the pricing framework actually has has that and it is designed to enable someone who only needs, let's say, a small number of cameras to be able to step into our platform. However, I'm yet to see a retailer say, oh, I only have these small number of cameras and they're the only ones I want to use. Like so often is the case, they've gone, ah, fantastic, and we'll add this to it and we'll add that to it. So multiple cameras tends to be the default position. As I said, the GPU enabled appliance per store is quite expensive and the prices are only getting more expensive on that. So that is a point of friction and that means going back to the problem and making sure the problem is really valuable and worth solving. And Then there is, after the installation of the appliance, there is a period and this is for all Computer Vision AI players where they need to calibrate their platform. So it's not like E Commerce where you've got a Shopify plugin and boom, away we go. This thing of Computer Vision AI, it does take quite some time to configure and tune the platform to measure exactly what it is the retailer wants.
[Voiceover]: Okay, so in other words, contact you
Justin Williams: for a quote that is best. And like the simple starting position for me on this is make sure that the problem that you want help solving is really valuable. Because if you're just after a simple store counter, go get a laser and do that. But if you're after something which can down to the second, follow that person through the store through various use cases and have those use cases linked together. If that data is helpful to you, then yeah, we perhaps are a platform that can help solve that problem.
[Voiceover]: As you were talking, then my mind went to another use case which I'd be interested to test with you here, which we've in our conversations never tested.
Justin Williams: But would it be, I'm always up for that.
[Voiceover]: Would it be, have you ever implemented it in warehouse or fulfillment centers to test productivity and processes?
Justin Williams: So we have not, but we have had retailers ask us, can we please deploy this platform in those environments? And the requests have really come from safety and compliance, to be honest. That's the first and foremost. And then secondarily to understand the movements that occur within those spaces because obviously, you know, on the shop floor and all of the measures designed to prevent theft are quite mature, questionable as to how successful they are. But in the back of house, where safety is number one for employees, there is also quite a different dynamic for somebody who has a different objective. I'll say. So we have not deployed in that back of house environment. It is on our roadmap for deployment. We have a lot of stores in front of us that we need to just roll out to. But the back of house is definitely unquestioned for them. I mean, simple things like supply chain docks, availability. Is there a truck there currently? How long was that truck there for? Yeah, all of these things that are vital for the replenishment of the retailer, store inventory to make sure that, you know, the successful customer experience is supported.
[Voiceover]: Lots of use cases there, you know, from like you said, safety. But even tracking stock from the moment it arrives. How long does it sit there? Where does it go? When will it hit the floor? Yeah, you mentioned there, obviously this, this you mentioned the founders and that they are young founders. This sounds like a capital intensive, capital hungry business, especially with the physical hardware element and the global nature of it and hiring people like yourself. You know, I've seen your paychecks. I read that you do have some really strong investor backing from some angel investors and VCs including the likes of Blackbird, which are, you know, they generally know how to pick a winner. What are some of the unique perspectives or even pressures maybe that you've experienced in your time so far, working with investors with what you're rolling out with black AI?
Justin Williams: I'll start first here by saying to talk to investors, the CEO is required. So I will abstract my experience separate to black AIs and I will reflect on the challenge that is deep tech, right? And I regard black AI as one of my first deep tech experiences. First things first. I mean this business has been going now for seven, eight years, right? And we just engaged with major retailers like Walmart and Woolworths. And that's why I got the call up, right? Like I've got some history with big retailers and can I help these folk that have been successful in getting engaged with these businesses. But back to, you know, the deep tech thing, it's building something that has not been built before and it's difficult to put comparisons on it. So the power of the platform is so significant to know when to pivot from research into fit for market, et cetera. That expertise, the patience, the support, all of that, I think I'm safe to say, is what's essential within those investors because it's not as simple as just rolling out a bit of code that someone is likely to replicate or copy within 12 months. This is very, very particular deep technology.
[Voiceover]: There's no playbook to what you're doing. And I could imagine that you would need investors and generally in this space that are on board with the direction. So I could imagine there'd be a big fair bit of consultation that goes into some of those bigger decisions when there is no use case or no playbook in front of you essentially to follow.
Justin Williams: Yeah, 100%. And I'd say it goes back to the founders. They are remarkable truly for having the foresight to create this platform that does what no other platform does. So it's truly remarkable for that kind of thing to be pop out of Australia, right? It's up there with Atlassian afterpay, those types of things that you're really surprised. I mean even zero out of New Zealand, who would have thought, right?
[Voiceover]: We've claimed them now they're an Australian company because they're on the. So tell me about the next 12 months or do you even think in 12 month increments, what, what's next for black AI?
Justin Williams: So for me, yeah, I usually chunk it down into 90 day pieces and at the moment I'm staring at Christmas and crossing that threshold and then into 2025 and growth of the business is, that's something that we're having to deal with right now and forward planning on the structure, on those resources to be just in time meeting the demands of the business and customers. It's, it's really exciting. So I think what is ahead for us and we see 2025 clearly is, you know, moving well into engaging with customers on a commercial footing to solve problems and for us to track the data of the impact of the platform so people would know that as building a case study. But for me it's much more to understand the economic impact of the software and use cases and then to be able to use that data to support the go to market plan that we've got which says, you know, let's stay the course with retail for the time being because retail in Australia is very similar to how it is in the US which is very similar to Canada and UK and other places where there are people in store. So we've got quite a long way to go in retail and there are other industries that we have our sights on but we are preparing to tackle those in due course once we're satisfied that we've made big progress in retail.
[Voiceover]: Are there many competitors in your space that you have to keep a sideways eye on?
Justin Williams: Yeah, I think for me I never discount somebody who says they're a competitor. I mean there are other local Australians based websites that I've seen that say they do this too. There's a really big global specialist that deals in that self checkout, you know, eerie kind of camera stuff that you talked about. And there are others, North American, Europe, European and always emerging. Like there's one from South America which I came across, I was like really? Like what the heck. And if you read, you know, the websites, they speak of very similar things, they use very similar language. But our experience in working with you know, very large retailers and talking to their teams that scan the market and constantly research and test this tech, they're saying wow, you know the one label computer vision AI, that's the thing. But your black AI platform is very different to the way others work and it delivers very different results. So I think it's, it's an emerging space and a really exciting space and that's why it's just fantastic to be a part of it.
[Voiceover]: I love that you acknowledge that there are competitors out there because there are a lot of people who would go, we just focus on our own game. We run ahead and we don't worry about what else is out there. But it does come down to strategic investment decisions for you, doesn't it, in terms of what markets do we approach, do we stay in retail, do we expand? So you've actually got to know what else is playing out there and where your differentiators are so that you can put the groundwork down under what you need to put the groundwork down and then expand from there. So, yeah, interesting point.
Justin Williams: 100%. Like for me it's all about critical mass and as I walked into this business and yes, they're engaged with Walmart. Who in their right mind would beat a path to the world's largest retailer and say, I'm going to start here, you know, just like I've said, that's called the index verse. Go to market methodology. And you know, to be engaged with one of Australia's largest everyday load pricing retailers, it's a real honor and opportunity. But with that comes the obligation to see it through. You either succeed or you fail and you go from there. And yeah, it's definitely tough. It's always about prioritizing what can we afford to do versus what we'd like to do versus what we have to do.
[Voiceover]: Yeah, well, you'll know you've made it once those Kmart registers move from the middle of the store to the front again. That's when you've done your job.
Justin Williams: All right, yeah, I'm hearing you. I'll put that one on the roadmap. Thank you.
Nathan Bush: Thank you, Justin.
[Voiceover]: Thank you for the chat. It's just been fascinating. It's not our typical E commerce scenario because we're talking a lot about physical stores, but to your point, we're talking retail here and physical retail is so important. We're seeing so many D2C brands establish physical presences, we're seeing so many traditional retailers bringing their physical and online worlds together as one entity. So I think understanding what customer data we have available to us is so important across both online and physical retail. So thank you so much for that insight because it's an area that I hadn't thought about a lot.
Justin Williams: Yeah, listen, Nathan, you know, just watching Adore Beauties news of late to opening select sites. It's fantastic to see that pure play E comm step into real store retail as well. It's fascinating and just the industry is terrific and it's been a real pleasure to join you on ADD to cart here today. Thank you.
[Voiceover]: Thank you Matt. Great to catch up.
Nathan Bush: Now, if people want to know more about Black AI or get in touch
[Voiceover]: with you directly, what's the best way for them to do that?
Justin Williams: Yeah, look, I don't know if the website has a call to action on it, but just use my email jwilliamsblackai.com and yeah, I'd be very happy to have a discussion and see what we can do to help but really exciting and you know I love your work with the E Comm community. Like in Australia, it's pioneering and we are a really advanced retail market and some of the things that we do here other markets learn from and you know, good on you for leading the charge.
[Voiceover]: Thank you mate.
Nathan Bush: I'm in the privileged position of hearing amazing stories like yours so I appreciate you taking the time to share it with us.
Justin Williams: All good, pleasure. Talk soon.
[Voiceover]: There you go.
Nathan Bush: I bet you hadn't thought about security cameras for a while and I think that's just fascinating to dive into that one area where you think about the technology that's coming down the pipe pipeline and the kinds of analytics we will
[Voiceover]: be working with and expecting over time.
Nathan Bush: What got me thinking was how can this play not just in our physical stores but in other activations around us to understand how customers are really interacting with us, whether that be physically or through screens. There is so much data and behavioral insight that can come with video analysis analysis and I always love catching up with Justin. His journey from the world of corporate retail and consultancy into the startup space is fascinating and I love how he's taken the leap on Black AI, which is a really, really exciting startup in Australia. But as you saw on the world stage, here are my three main takeaways from our conversation. Real Time Data for real Time Impact Black AI is ability to transform traditional CCTV that's always existed into a tool for real time. Decision making is game changing by not having to change out a lot of the physical components and using an AI overlay to analyze what's coming through modern security systems. It's not just preventing theft, it's giving retailers detailed insights into customer behavior. From prevention preventing theft to optimizing store layouts, the physical world of retail can now be almost as real time as the online one. This kind of data empowers retailers, especially those who are used to moving quickly to make smarter decisions that boost both security and the customer experience. Number two, the ethical innovation dilemma. What really stood out to me, and you could hear Justin stressing upon it all the way through our conversation, was Black AI's commitment to protecting customer privacy while delivering valuable real time insights. It's a difficult balance to get right, but as Justin said throughout that conversation, Black AI was built with this in mind. It's not something that was tagged on to the end. Their use of soft biometrics like color clothing and anonymized behavioral patterns rather than personal data or face recognition ensures that retailers get the insights they need without compromising customer identification. If you take anything from that conversation, understanding what kind of questions you should be asking around privacy when it comes to video analysis and recognition with AI is a really important one for your business. It's going to open up a whole range of data that you never thought
[Voiceover]: would be possible or that you'd be liable for.
Nathan Bush: And number three, bridging the gap between physical and digital. We all know it's a trend that many e commerce brands are now launching into physical retail and tools like Black AI will be coming onto their radar for seamless and data driven connection between the two worlds. Understanding how customers can move and behave in store can help brands optimize bricks and mortar spaces, but also take those lessons into the online world. Just like we've been taking online lessons into the physical world. Once we're in an omnichannel world, we unlock a huge amount of data and analysis that we have to work out how to use between both spaces. Thanks for tuning into today's episode of Add to Cart. If you enjoyed the conversation, please share it.
[Voiceover]: Leave us a review on Spotify or
Nathan Bush: Apple Podcasts and don't forget to support our amazing partners Shopify plus plus and Deliver in Person who help bring this podcast to life. Before you go, we'd love to invite you to join our free e commerce learning platform Add to Cart Campus. Meet other professionals and learn from e commerce experts to take your business and your e commerce career to the next level. Register to join campus@addtocart.com au e commerce canvas now. If you enjoyed today's episode, make sure you share it with a friend or a colleague. Or even better, leave us a review on Spotify or Apple. It really makes a difference.
