192. Idea: Industrial Maintenance Solutions – Spot AI Transcript

00:00 Chris: 

Welcome to EECO Asks Why. Today we have an idea of episode and we’ll be exploring the human side of technology. And to do that, I have Tanuj Thapliyal. And he’s the CEO and co-founder at Spot AI. So welcome Tanuj. How are you doing? 

00:18 Tanuj: 

Good. I’m doing great. Thank you for having me. 

00:21 Chris: Oh, we’re excited to have you and very excited to have you. I know when we connected, we talked, you have so many cool things going on at spot AI. And I really am looking forward to this conversation and maybe just set the stage for us. You know, we have a lot of industrial listeners out there that are tuning in EECO Asks Why so, talk to them. When you refer to the human side of technology, how does this apply directly to them?

00:44 Tanuj: 

Sure. So Spot AI was founded on an observation that lots and lots of people at work are changing their relationship with the cameras that they’ve installed in the workplace. So specifically in the industrial workplace in, you know, the factories in America, there are a lot of security cameras installed that were designed and built for security. So when you think about security, you know, you think about, like one time in six months emergency. And what we’re seeing is that those systems, they were never designed for daily use. They were never designed for a line manager or, a plant manager or, a frontline worker, a laith the operator. To actually be able to use to make better decisions with how they do their jobs. What we saw was a gap to build really, really easy to use technology that anybody in that factory could pick up and use within a few minutes without training. And we felt that the data off cameras is far, far more useful than just those security emergencies.

And, you know, it could help all kinds of people in the factory make all sorts of decisions and we can get into a lot of specifics and stories and use cases as well. That’s kind of, you know what I mean, you know, on the human side of technology, that it’s great to have the latest and greatest, gadgets and technologies and, AI and all these fancy buzzwords, but it’s kind of, it doesn’t mean anything if somebody can’t pick it up and start using it and benefiting from it really, really quickly with really low friction. 

02:26 Chris: 

I’m with you. Give me a quick rundown though. So Spot AI, what’s up with that name? 

02:33 Tanuj: 

So there’s a story behind the name. And basic idea is that, we felt that people want to use our cameras for more and people are looking to spot something in a screen or spot something in a video and if you actually look at the logo, there’s one blue kind of colored you know, spot. And the idea is that in the whole image, there’s going to be one area of the image that’s relevant for the workflow or use case. Right? For example, you might have a camera looking at, you know, rows of shelving in the OQC and the outgoing quality control section of a factory.

And perhaps there’s a section on those roads where there’s a stack of boxes, that’s a skew and wasn’t stacked properly. And it’s a safety hazard because it might tip over, tilt over, for example. So the idea with Spot is it’s kind of like a verb, like you’re trying to spot something or you’re trying to see what’s happening or see what’s going on around you.

That’s pretty cool. Yeah. Yeah. So that was kind of the idea behind that. 

03:41 Chris: 

I got you. So basically, and it’s also, I’m assuming, so there’s a platform that you built to tie some of these devices together to be able to spot those instances and, and make a call in a moment. 

03:52 Tanuj: 

Exactly. Yeah. So whether it’s for operations or safety or even security in those rare emergencies we want to make it easy for them to spot what’s happening now and what’s actually going on in the factory. 

04:06 Chris: 

No doubt, no doubt. I’ve been to so many manufacturing facilities Tanuj. It is unreal. And I think about it and I’m curious, you know, so many of them are late adopters to technology. It can be baffling. I’ll rack my brain. I’ll try to figure this out sometimes because you know, we’ll have a smartphone in our pocket that’s smarter than the rocket that took people to the moon, but we go the factory floor and it’s still like that technology is still there sometimes. Right? So why are they late adopters? 

04:36 Tanuj: 

There are a couple of things that are happening. I think one of them is that switching costs are high whenever you’re trying to change a process and you need a lot of conviction and proof that that switching is going to be worth it. I think that’s one of the, one of the big ones. Why certain, you know, factories might take longer to adopt a technology versus another. There’s nothing wrong with that, by the way. Right? For example, you might have set a certain output rate per hour, on a particular line that you’ve spun up, you know, your spinning up that line, you’re allocating resource to that line. And, any blockage or slowdown in production on that line. It’s, it’s a very expensive thing. So the training associated with shifting a process, it’s actually kind of expensive and it brings a risk into the factory to do it.

So, you know, the first thing that I would say is that you have to bring down the barrier to be able to make those switching costs lower. And then two, you have to make that change in whatever you’re doing in a process, worth it. And when you bring in technology, you are, you know, for better or for worse, you are changing a process and changing the way that people are used to doing their jobs.

So my view on it is you have to start with the smallest kernel and the simplest kernel that’s useful for that factory. And rather than changing that workflow you really want to supercharge a workflow that they’re already trying to do. And I think if you do those things, I can be easier to bring new technologies into the factory. Now that’s just on the implementation side. There’s also other areas where you can bring down friction. For example, when you’re researching and discovering new technologies that you can bring into the factory, that’s also a very critical area. You need to make sure that. You can provide very rich and easy to read and easy to understand content, right?

Lots and lots of content. You need to make it very easy for people to sample or try or see a similar example of it in action. And then again, you, you don’t want the first bite-sized chunk to be the size of a XL pizza. You want something very tiny and small, you know, in kind of that first baby step towards a digital transformation. And like, you got to prove value and earn the trust of the customer. That’s kind of with the buying experience and then with the product, and then after a customer’s bought that this to me is arguably the most considerable sides, which is you got to support and help the customer every step of the way, once they’ve brought in the technology.

So you gotta be available for whatever they might need, you know? You know, adopting the technology and the org, but the human side, it allows you to focus on the user and it allows you to focus on all the different people, you know, that work in that customer’s particular factory and really cater to their needs and make sure you’re being really responsive to their needs. And you make sure that you’re, you’re bringing down friction for them to be able to use the technology. 

07:51 Chris: 

And I mean, it sounds like as well, Spot AI, you guys are focused on the video component of the technology for an industrial manufacturer. So maybe speak to that, the impact back in a half as well as the cost barrier because I’m assuming there is a cost barrier as well from the industrial standpoint, when are you looking at the video technology. 

08:11 Tanuj: 

What we do is, you know, we make these easy to use AI camera systems. And the idea here is that a operations, you know, plant manager, right? If a conveyor belt stops, we make it very easy for them to go into their camera systems and figure out why did the conveyor belt stop? Did a nail snag. Was there a power issue and you’re not seeing any, you know, like LEDs light up at the bottom of the belt. Is there a tear in the belt? Is there an issue where, you know, the operator of the belt had actually been relocated to a different part of the plant, And there’s nobody there operating. And, video is kind of that ultimate ground truth to know what’s happening. So, you know, that’s kind of a, like an operations use case where you can have some real impact because it’s really useful to have that visual context so that you can, you can operate kind of the next steps of your process. Safety is another huge one where you can have impact in a factory. Huge, huge, right. 

09:18 Chris: 

How would you do that then from a safety standpoint? Are you talking like the EHS departments of factories? So speak to that a little bit more. I’m curious on the safety component. 

09:28 Tanuj: 

Yeah. I mean, I think this is significant and extremely important. And the way I think about it is that you have these EHS professionals working really hard in all these factories to keep people safe and, with the complexity of COVID I mean, we’ve gone two years with this thing now. With safety, you want to basically be able to provide constant training and education of your workforce on how they can keep themselves and their coworkers and peers safe. And video’s a very good way to do that too.

So for example, you can very easily use video to build training videos and best practice videos for new hires. So if there’s particular types of machinery you’re operating right now, look like someone who’s hired into a factory. I mean, you are going to be going to some school before, and you’re going to be learning some basic safety practices there, but once you’re actually in a production environment, making sure that becomes a muscle memory and it becomes automatic. Video plays a huge role, like for example, forklift operation. There’s going to be a set of standard operating procedures, making sure you’re wearing a safety vest. You’re clipping the vest appropriately to the equipment. That’s a clear one, right?

10:46 Chris: 

I mean from a safety standpoint, When I’m thinking about the video component of it, workforce attrition, skills gap, all these things that he’s manufactured started getting hammered with every day. It sounds like what you’re saying is the video that human side of technology with that safety component could really impact that greatly, because here I put right in front of you, Mr. New employee, who I’ve been trying to get to come to my plant forever. Finally, I got you to come here. So here’s some great video to keep you safe and to teach you, and that you can reference back because you’re not going to remember all this stuff, you know, day one.

11:21 Tanuj: 

No. So our customers do this is, you know, they put up signage right on the factory floor in like actual, like flat screens and videos screens and then they’re able to see around the corner of exactly what’s happening in an industrial process, for example. And then there, you know, you can have training videos so that you can coach and train your workforce anytime there’s a safety slip and the way I think about it as this is that, you know, the best people are going to want to work at places that really take care of them. And one way you can really, really take care of your people is make sure they’re being kept safe and make sure it’s a continuous improvement program that, they’re always being kept safe.

And this is kind of one way to accomplish that is you can use video to basically set up alerts when there is a safety moment. And then the next morning stand up you can bring that up and then use it to coach and train everybody. So it’s something which I think is really, really crucial, particularly now, more than ever when there’s some very significant train labor shortages. I think education is a way to solve that and two I think treating your people really well is a pretty obvious way to solve that, right? 

12:33 Chris: 

No doubt. You’re all over it, but I’m also, I’m thinking about Mr. Maintenance manager or E&I tech, who has a small budget, and they’re trying to figure out how they can start implementing some of this with the constraints they have from a budget. So to speak to the investment side of this, you know, for a manufacturer making these investments, what did they need to consider here? 

12:53 Tanuj: 

Yeah. So the way we approach it, Chris. And this, again, ties a bit to the human side of technology is like, you want to make it as low friction as possible for the end user and the buyer to see value. And it doesn’t have to be a big first chunk. So we do a couple of things there one, we never require a customer to buy before they can try it. So we actually do a lot of free trials. So anybody who wants to try the technology out can do it absolutely free of charge. Now, the next thing is time, right? The time investment also ends up being pretty expensive. So if you take the dollar side aside, right. From a time perspective, we’ve designed the technology. So that it’s really, really easy to install. And what I mean by this is it’s like installing an apple TV in your home. It’s something like, it’s something you just plug in and you’re done. Like my mom can install the system. 

13:54 Chris: 

I’m not sure if my mom can, but maybe. 

14:00 Tanuj: 

That, that becomes a design vision. The system is really easy to install and then say. You do a lot of really generous free trials. And third, you make the first purchasable, you know, the smallest chunk that you can buy, you only make it a few grand and a couple thousand bucks. So if you do all those things, then you can enable somebody to adopt the technology at whatever is the lowest friction point for them that they’re comfortable with. Now, if they’re not getting value out of the systems, they shouldn’t be our customer. So at that point, you know, like they can stick with what they have. If it’s on a free trial, they can ship it back, no harm, no foul. If they are getting value, what we see happening is customers do add more cameras and they do increase their coverage over time. And rather than trying to pre-sell the whole thing up front. We’d rather start with the smallest amount that’s actually what they determined to be useful for them. And then let them decide if they want to expand, but we make it very easy. So like, you know, we only charge couple bucks per camera feed per month, and then if they need hardware, we actually give it away for free. So if they need cameras, we’ll give them, you know, high definition cameras for free, and then they can just keep the cameras, you know, even if they don’t renew the software. 

15:18 Chris: 

I love that approach because you’re building advocates. First of all, you prove it out. You build the advocates internally. I go back to an episode that we did a long time ago with an executive at Rockwell was she told me think big act small, this let’s those employees really, you know, have that greater vision, but take those first little baby steps, I think is what you said earlier. And then once you get that, then it, then the snowball starts rolling because they see value and they can start adding it and things like that. And I’m curious, have you seen any headwinds? So if you’re an industrial right now, you know, and things that, you know, I’m trying to, I’m trying to think for Mr. Maintenance Manager out here, that’s going to, you know, wants to move forward video technology. What are some headwinds they may have that you could help them understand right now? 

16:06 Tanuj: 

From a video intelligence perspective, you know, there are a few headwinds that you’re going to face. I think one of the first is figuring out, you know, what workflows and what use cases are the right starting use cases where you can get value right out the gate, right? What’s the largest with the least friction possible because there’s a bunch that you can do with your cameras. And there, I mean, there could be stuff around site visibility, some of these operations workflows, workforce training, safety security, and what I’ve learned and what we see do address the headwind of like, what is that first organizational use case you start with the best ones are look and see what is your organization already using cameras for?

And actually start there because. If your organization doesn’t, you know, actively try to use video off the security camera systems, for example, then, it will take a little bit more time to build that case, but if they’re already trying to access video off the security cameras, to be able to detect, you know, slowdowns in production, for example, that’s a great candidate first because somebody in that org is already trying to use the systems and hitting pain.

17:29 Chris: 

That’s wonderful. That’s wonderful. Now you also, you know how it is when you, when you have budgets and you’re in a manufacturing facility and industrial setting, you have to have a return. You have to show some type of metrics on how this is working. You know, even though this is the human side of technology, and we’re talking about video, what, where do you point people to show this is making an impact. 

17:51 Tanuj: 

Yeah. Yeah. So on the safety side, you can actually measure from an accident basis, right. Near misses and, you know, accidents that have been yeah. That had been prevented. It gets a bit more serious with worker’s comp, but you know, you can actually also root cause, you know, worker’s comp claims. Where my head is generally at is more on the accident prevention side and the near miss side, because that’s when you’ve done a win and a solid for everybody, you know, once someone gets hurt, it just kind of sucks. That’s a situation you never want to have happen, but it does. So, you know, measuring and quantifying the near misses, it’s important. 

I think on the operation side, looking at throughput is something that’s really interesting. For example, like there, there are certain sectors of the economy and of our supply chains that have had a lot of growth compressed because of COVID. You have situations where you have a lot of factories that have such a backlog of orders that they don’t have much time for recovery and their throughput and how much they expect to be able to produce off of that line. Like it’s, it’s extremely crucial that that number stays high because they don’t have the time to be able to catch up if you have some yield loss, for instance. So another really clear way you can measure is by looking at throughput of a line and then making sure that, you know, whatever pass-fail criteria you’ve set, right. That you’re yielding at the correct amount, you know, for whatever finished good that you’re producing or intermediate goods that you’re producing.

So video intelligence can help you there as well by detecting the slowdowns, you know, visually. Is there a cosmetic defect on the part, is there a actual mechanical defect or a fit check issue? You know, did the line actually stop, right? And you had a stoppage of production and you’re trying to figure out what happened. All these things drive very quantifiable ROI. That’s a more top line and kind of revenue associated item. 

19:55 Chris: 

Any examples, specific to manufacturing that you can think of that would really resonate or maybe I don’t know a line stops or, or, or something falls out of you know, off a conveyor belt where, where this technology made a significant impact.

20:10 Tanuj: 

Yeah. Yeah. So let’s so there’s, there’s a really good one. So, you know, let’s say you’re a, you know, injection mold facility. So you’re producing like clamshell enclosures, basically. So you have some, you know, pellets, some resin, Plastic, they go into this big hopper, right. You’re dumping them in, you’ve got this big machine that’s heating and melting it and then you’re compressing it into a mold. And then, you know, the mold comes apart. You’re ejecting parts off the mold and they’re falling onto the belt and then they’re rolling on the belt right. Into some sort of box.

This is an example of something that I witnessed firsthand, which is that somebody had decided to leave the machine on overnight unattended. So imagine cameras looking at rows and rows at these injection molding machines and you have no lights on any of the machines except a red light on one of the machines. And that happens to be the machine that you know, is going to be doing a production run the next day. And what happens to injection molding machines. If you just leave them on, is the thermal and heating profiles get all messed up. So, I mean, you’re keeping the thing hot the whole night, so you’re coming the next morning.

You hit go on that job. And because the machine has been left on the whole night and it didn’t do a very good job at, you know, regulating its own temperature, all the parts come off, the machine warped and now you have to scrap a hundred percent of that run. There’s a cost side to that. There’s a lost productivity side to that. The customer of the factory, they’re seeing a slip in schedule. Right. Because it’s going to take another day now to do the run. And video and video intelligence. It’s a very easy way to solve that problem because, you know, you can auto detect, you know, if a is on or off visually using the cameras you already have, or you could even mount cameras, you know, and look at the machines and, you know, be able to tell if they’re on or off. 

22:18 Chris: 

That could have been a simple fix that with the video? 

22:23 Tanuj: 

Exactly. Exactly. And it’s using kind of, you know, one of the most ubiquitous senses that we have which is vision.

22:31 Chris: 

Absolutely. So think about, maybe speak to the industrial manufacturer out there right now. Specifically to the person inside the, the facilities that you work with, you know, how these programs get started, these initiatives, is it engineering? Is it safety? You mentioned safety already. I’m just curious. Where do you find the biggest impact and the biggest willingness to try new technology like this?

22:55 Tanuj: 

So interestingly, it starts a lot with IT. Oftentimes IT, so a lot of our conversations actually start with IT. 

23:08 Chris: 

Is that where the need comes from? Like the request comes from IT?

23:12 Tanuj: 

No. So this is the interesting thing is if the factory or the organization is already trying to use our cameras for more, IT is generally the one hearing about it and getting, They’re becoming very, very aware of the problem. And the reason is because that EHS professional is going to it and saying, Hey, the cameras are down. Or, you know, Hey, the software here is really clunky. The plant manager is going to it and saying, Hey, I want visibility. I want more coverage. I want to be able to see more on the factory floor.

And that’s when IT goes and tries to make the systems better and faces a lot of pain and friction. So when we talk to customers, generally having, you know, IT being aware of the pain is usually a good indicator that that organization has some video workflows that they’re trying to get to. And they’re, they’re trying to do more and more. Now that being said, operations people like plant managers tend to be a really, really good, you know, like center of excellence to work on. Safety is another good one, but, but we hear each of them kind of, you know, equally intensely. 

24:25 Chris: 

Well you know, but when I’m thinking about an industrial facility and you mentioned IT, but the, the impact you’re making is on the OT. So a lot of times I’ve heard a lot of friction, you know, there could be friction because there’s not all the IT and OT don’t always align. So are you finding, do you have to fight in both corners or?

24:46 Tanuj: 

No, we see if you make it really, really easy to install the systems and use them. You are supercharging IT’s job because you’re taking one thing off their plate. Like it’s so easy to install, set up configure. And then it’s a self-driving system where it’s fully managed. So, you know, we’re making sure the camera’s stay up. We’re making sure the system just works. It’s getting better every two to four weeks because it sat it’s software as a service. So we’re constantly improving the system with upgrades from IT’s perspective they don’t have to worry about this thing, it’s installed and that prevents there being, you know, some sort of friction with them, but at the end of the day, IT exists in organizations to also serve their end users. So to me, it’s really about, super powering the operator and you have to align to what problems the operator wants to be able to solve.

And then you need to make life easy for them and for everybody else in that company as a result. So we don’t see that same tension because we focus so much on simplicity, ease of install and ease of use. 

25:59 Chris: 

So is it safe to assume that I requirement needs to be a robust industrial network out the beginning of the gate?

26:08 Tanuj: 

It’s a good question. Our technology is pretty robust. So similar to this call we’re on now. Sometimes it’ll get a bit grainy the video, and sometimes it’ll get really sharp. It’s always recording at the highest resolution there’s technology at work with Riverside.fm that enables reliable streaming of the video, regardless of bandwidth conditions. We have very similar technology. What that means is that regardless of the network, we will fit within the pipe that that customer has and if they have really good bandwidth, that’s great. If they don’t have that good bandwidth, we’ll still work. 

26:49 Chris: 

So it sounds like you could throttle up and throttle down as depending on the performance of the network at hand.

26:55 Tanuj: 

Yeah, and it’s invisible. So basically it just works. Now we’re going to be recording down at the highest resolution we possibly can, but then the actual delivery and consumption of video we’ll, we’ll auto tune it. And the one area where we’d recommend, you know, being a bit cautious is if customers use a lot of Wi-Fi. And the reason for that is if you’re in an industrial facility and you have 50 cameras, like even the best Wi-Fi on the planet, isn’t going to be enough to be able to serve video over all of that, because you’re going to have other business traffic. Like you might have some, you know, ERP or shop floor system. Like they’re going to be other things that you need to use that Wi-Fi for. And that’s where we’d steer away. As long as it’s like Ethernet.

27:43 Chris: 

You’re golden there. That definitely cleared it up a lot for me, I guess. And the last area I was thinking about on a topic like this, I don’t think we can leave without talking about cyber security. So I mean, how, what is the approach there, you know, so far is keeping these, this videos, technology safe and secure for the industrial user. 

28:04 Tanuj: 

Yeah. So we employ a zero trust security model and what this means is that only authenticated users in authenticated sessions have access to the video. And then it’s up to the customer to decide who they want to authenticate as a user. That’s extremely crucial. And then we, you know, we employ kind of the bleeding edge technologies there to make sure that stays maintained and we’re constantly making it better and better. So we’re constantly adding more and more from a cybersecurity perspective. This second point, which is, I think really important point, which is the customer’s videos are their own video.

So we provide customers actually hard drives included in the subscription and all the video is recorded privately on the actual factory floor for the customer 24/7 and then 30, 60 or 90 days guaranteed. And in that case, that video is behind the customer’s firewall. It’s secure in their own building and if they don’t want to send it up to the cloud, they don’t have to, it’s entirely up to them.

So if they only want a local set up, think of almost like an airplane mode, right? Like you want a local setup where you’re trying to use video to see around the corner, see if the trucks have shown up in the loading bay. See if that conveyor belt has stopped see if a stack of boxes is too high, our system fully works in that mode. None of the video will actually go and touch. So the video will actually go straight from the camera straight to whatever the end user device is. It could be a desktop, it could be a laptop, it could be a tablet, it could even be a phone. So from a cybersecurity perspective, you’re enabling them to keep the video private and onsite only if they want remote viewing, then we’ll enable remote viewing. And there, you know, we do full encryption at rest and transit. We make sure that there’s like total encryption of all the video. 

29:56 Chris: 

It sounds like you’ve addressed that head on and that was the big question I had. So a lot of it stays local when it needs to have that remote access. You have the, the models in place to be able to keep it secure.

30:08 Tanuj: 

And it’s also fully encrypted even locally. So were pretty aggressive about working on this and you know, this particular area of cybersecurity and making sure we’re going to be best in class.

30:19 Chris: I love it. I love it. Well Tanuj. This has been great. Now we call it EECO Asks Why. I always wrap up with the why, my friend. Speak to the industrial listener out there. You know, why is embracing technologies like video so important to the success of manufacturing in the future? 

30:33 Tanuj: 

Our view is that video is a data source can absolutely maximize human potential and it can supercharge every single person on that factory floor to crush their jobs. It can supercharge the ability of that new hire to learn how to operate a new machine or new equipment. It can supercharge the ability of our plant manager to make sure throughput stays high.

You know, these factories be able to attract and retain the absolute best people because you’re helping everyone on their individual missions using video and using cameras to get visual context of what’s going on around them instantly, and then be able to take corrective action rather than asking the question, what happened? You know, why did this happen? You know, you can ask why is this happening in the present tense and pulling out a very rich video feed on your phone? It lets you answer that question. 

31:28 Chris: 

Well, I love the answer. I love how you said, you know, video can maximize human potential that’s it. So Tanuj, thank you, my friend, this has been wonderful for the listeners out there that want to learn more about Spot AI, go to the show notes, check out that we’ll have all the links there that you can connect directly with Tanuj, his company, the wonderful things that they’re doing. So thank you again so much for sharing all this great insight today.

31:51 Tanuj: 

Thank you. Thank you for having me, Chris.