Daniel Spahr with Stream Analyze

SUMMARY KEYWORDS

data, device, cloud, industrial, asset, analytics, iot solutions, ai, good, Spahr, daniel, edge, people, industry, business, install, stream, sherpa, technology, provide

00:03

Welcome to the Industrial Talk Podcast with Scott MacKenzie. Scott is a passionate industry professional dedicated to transferring cutting-edge industry focused innovations and trends while highlighting the men and women who keep the world moving. So put on your hard hat, grab your work boots, and let's get

00:21

right welcome to Industrial Talk. Thank you very much for joining and for your support. We are broadcasting on site right now, right this very moment at IoT Solutions World Congress, it is a an event that brings technology, people, companies, solving problems, wanting to collaborate, you need to put this on your calendar for next year if you are not here. But do it because it's great to get people like Daniel is in the hot seat. And this platform, as you know, is dedicated to industrial professionals all around the world because you are bold, brave, you dare greatly. You collaborate to solve problems. You're making the world a better place. That's why we celebrate you here on this podcast. Daniel Spahr is in the hot seat. How are you doing?

01:09

I'm doing fine, Scott. Thanks for having me again. I was here last time too. That was fun. It was. It was it's a great show. It's a good show. Great show.

01:18

Yeah. Because the conference is it's bigger this time.

01:21

Yeah. Well, I think, you know, people gotten out of their seats from COVID. And now they're actually getting on site.

01:27

It's good. Yes. You know, it's interesting, even though that COVID whole that pandemic thing going on? It didn't slow down. The the innovation, the technology,

01:37

no, it actually picked things up for us. COVID helped us a lot. Funny enough. Close things. No. But to open things up. Definitely. It was really strange. But we thought you know, this, is it, but no, not at all?

01:48

Well, no, because they realized that, okay, well, that was sort of a gut punch. But I want to stay in business. So what do I need to do? I need to think differently. And so yeah, I agree with you. 100%,

02:01

well, actually saw a lot of companies actually spending more time in innovation. Because they're not doing they're not as much in their business as usual. Not as much in meetings. So now they go, Okay, how do we innovate? How do we innovate ourselves out of this, and that helped. So people started to listen more.

02:16

Yeah, but that's true. It is. I never thought of that. Because you don't just go back to the salt mine, which a lot of people will at this conference, they're busy, they'll go back to the salt mine, and they just do what they need to do. But however, it is a it's a great way of being able to meet. I want to say trusted individuals, trusted people who know what's going on. So that's what

02:38

that's you know, that's what business is about trust.

02:41

That's it? Yeah. That it? I mean, it's great that you got this technology, this innovation, you got all that wonderful stuff. But the reality is, is that it gets down to trusting people,

02:50

trusting people, but also trusting in technology. So I think it's a combination of now

02:54

but don't you think you have to sort of find that Sherpa, I'm gonna go to you. I haven't even I haven't even thought about the technology. I just know that the internet says I have to look into it. So I need to find somebody like you. Yeah, you go, and then I'll go and you give me the trusted technology? And if you don't know, you know, I don't know. Okay. All right. Background, give us a background for the listeners.

03:21

Okay, background, long background within digital my whole career. And always in that sort of Crosspoint between business and technology. I've always been driven by Okay, how are we helping the business really, even though I've been in super deep tech stuff, sometimes. Right now at a small company called Stream Analyze, we're out of Sweden, we provide an analytics platform that you can deploy on and install on any device out there, and allows you to bridge bring AI directly onto your device, rather than bringing all your IoT data into the cloud and doing AI there we run through,

03:57

run through that again, so I know this is this is the correct me if I'm wrong. I'm collecting data off of this asset, the asset is then you know, shipped to the cloud. The cloud then takes that data and then does the the analytics on it, whatever it deployed some AI, you know, whatever. That's where it resides. You're saying, hey, why don't you keep it sort of closer to you?

04:23

Well, why don't you do the analytics up front? You know, so one, when the data is generated, you do it right there on the asset. If it's on a car, if it's on a lawn mower, or chainsaw, I don't care. But you do that first analytics up front. And then you know, whatever results you need to ship up to the cloud and what you really need to save in terms of data that you ship up. It can be an alarm, or it can be just a heartbeat, or whatever it might be, but you don't need to take all your IoT data and store it. Do you know that's for sure.

04:55

Okay, so, the process is now I've got this asset I'm going to perform my analytics close to that asset, or on the ASR on the asset itself, let's say a motor. Yeah. So I'm pulling motor data. And I like, and in that I'm, I'm choosing this data, this piece of data in this piece of data. Yeah, and whatever, and I'm doing the analytics on that. And then all the other data, I can either make the decision of saying, that's future, or I can make decisions on that. But the real key is performing those analytics on site.

05:29

Yeah, exactly, exactly. And you can do that for multiple reasons it can be, you know, you need a quick decision, you know, something has to happen really fast, you don't have the time to step it up there. Or you just, you know, you want to you don't have good connectivity can also be another one. So if you have poor connectivity, that can be another issue. But there's also cost. I mean, this is a real real driver. By doing this, we're, you know, we're pushing down the amount of data traffic, both for your data plans and for storing it and processing on the cloud by at least 70%. We're seeing this across the board, and you know, that will save you money.

06:05

That makes sense that it sounds reasonable. So let's take that same analogy. I've got this motor, I'm doing my analytics on that asset. And I'm sitting

06:17

outside of the fact that be on site, I can look at the results of that, because it's right there. Right on the can I say edge? Yeah, or on edge on the edge on the edge, whatever. So it's right there. And I'm looking at it, I can make that those decisions, and then all the other data, I really don't care, and I'm not sending it up to the cloud. I'm just setting that up. Now, if you want it, that's your decision. It's your decision. But decision. And that, you know, that's, that's probably one of the hardest things I think people are having today in terms of data is making that decision of what data is important. What data do I want to store and what data do I want to do analytics on, somebody's got to make that decision. But what's happened historically, is now when they get access to all these great cheap sensors, and access with, you know, all the data transfer, let's store everything, just in case, just in case you might need it. This is where you come into black dark data and all this kind of stuff. Yeah. Dark data. That's the data you don't know you have or you don't need. I don't know the concept entirely. But no, no, no,

07:18

but that's it. Because you're right. But then again, the value of the data, right, let's say, this point, this point, this point, no brainer, right? There. Those are and everybody can shake their head and say, yeah, yeah, yeah. That's a no brainer. And you achieve, let's say, 85% value from just those things. Yep. And then all of a sudden, you just have what 15% To try to, you know, debate. And you're you want to send all of that up because of that. 15%, because you've already hit the the ones go ahead.

07:56

Yeah, but not only that, I think one thing, you know, even those 15% You might not need them historically, either. You just need them to do that calculation, right? Do the calculation upfront, and then ship up, whatever, you know, what do we need to really store long term? Really? I mean, some industries like health, tech and so forth, they have regulatory issues. Yeah, that's different. But but but in terms of taking, let's say, you want to know when something's about to break down, I don't need to know all the data when it was doing well, you just need to know what's happening when it's breaking down. This is what our clients are doing. So when we're working with them is they'll say, Okay, now we saw an anomaly, Okay, grab the data that happened a minute before that, and then that you want to ship up along with an alarm saying something went wrong? Because that's important. And that's a good decision.

08:44

How do you have that conversation? You must have that conversation, you must go into a client and saying, Yep, I got it. Yes. Our solution is here. It makes sense. Yes. Can't poke holes into it. That's where it needs to reside. But you're going to eventually have that conversation with the client and say, okay, just want to let you know, we're going to have this data conversation. How do you sort of work through what is of value? And what is just let it go? Well,

09:14

first of all, I mean, we have examples where we're done with other customers to give inspiration, right? That's one, many times our clients have an idea of what they want to do, but they don't know how to realize it. So then we can start talking about about how we do that. But I think the more important thing is they will say, let's start off real simple. Let's just send, you know, do a heartbeat or gets get some kind of information to understand what the data is to understand. Because they don't always know what they want to do with it. So they want to start off real simple and just monitor the machines or assets to see what's happening. Maybe let's just send up something when we have something going wrong. That's fine. So that will start off with a simple then they'll start evaluating this and saying, okay, Wait, now we have a tool that are dependent disposition that allows us to make decisions about what's important with our data. Now we can govern it. But if you don't have that, you start to get in the you want to take all the advanced stuff and govern everything from scratch, it will never happen because you haven't taken your baby steps yet. So this is often how we go about it. Some clients have good ideas of what they want, they have this is all the use case, we want to realize, can you do it? Well, okay, well, it's probably do these, you know, or this and maybe all of them. Some say, tell us what we want to do. So what can you help us with? And then we have to give them inspiration?

10:35

Because if, you know, I would, I would probably be the company on the latter part. I say, Hey, I, I know, I don't know where to start.

10:44

Yeah. And then we'll say, let's, let's start simple. Let's get to know what you have, and how that will provide value for you. And then we'll start unbundling that value over time, because you need to get to understand what it's doing.

10:58

So from a physical perspective, for Stream Analyze, right? Yep. You, you have a device, what does that know we

11:07

have a software, we're software, just some stuff. Where does that software reside

11:12

on that we assigned, we install ourselves directly onto the device or acid or machine into a piece of yarn with a memory and CPU that has access to that sensor data. So it can be on a telematic unit on a car, or it can be a little digital board that they've slapped on to a chainsaw or something like that, that has into an API or accent direct access to that stream data from the from the sensors, that's where we install.

11:45

So you could go out to that that asset. Whenever that motor.

11:49

Yeah, we install on any device, that's our whole tagline to bringing AI everywhere for everyone, right?

11:58

I didn't know that. I do know, I didn't know the fact that you could do that with it. No, I look at a motor, I see the motor, I know that there's stuff on that motor.

12:07

Older motors, of course, maybe you don't have that, you know, intelligent chip or whatever, then you need some kind of hardware device connected to it a gateway or router or something like that. But you know, we do not provide that hardware, we have that software components.

12:21

That's what I'm getting at. So it doesn't really matter. I got it. So you got the device. You use your software, into the device, wherever it is. And then all of a sudden, you're you're able to do what your your, your solution provides. Yeah.

12:35

And we we like what we like to say it's like, you know, real time Excel out on the edge. And why is that? Well, because, you know, you've worked in Excel?

12:45

Sadly, yes, I do. And I've created a lot of financial models.

12:51

Yeah, exactly. But who've made them all as you did. Yeah, yeah, exactly. We don't build models for our customers. We provide the platform for our customers to build their own models. That's the other kicker.

13:04

What does that look like? When he What does that? Well, what it

13:06

looks like is we install on that edge. Yeah, we have an installation and their their cloud or their server, and then when they have a front end. So now once they have that installed, you know that whole ecosystem up and running an infrastructure, they can see all the data they want on any device at any time. As long as there's connection, they'll see that data and then they'll say, okay, great, well use our tool, or conventional tools, because we can do that as well, to build models around that once they're built using our infrastructure, they can slap that down onto the device instantaneously. From a technical perspective, you don't need to do a firmware update, which for me, usually in you know, requires you to reboot the whole damn systems who's my friend? But here things are just running the whole time. Right? So you do you speed up your development time pretty quickly. We have had three customers tell us, you're like 20 times faster for our edge development than what we've had traditionally.

13:59

No kidding. Nope.

14:01

And we didn't ask for it. They holiday told us.

14:05

So I'm just I'm just making sure that I, so I'm able to contact you, your team? Yep. I've got assets you'll come in. Here might be a legacy or an older asset. You look at, you look at a situation I I'll say, I want to know, this is a critical process here. Yep. I want that process so that I can have greater insights into it. And I want to be able to pull it, analyze it, use some sort of algorithm, you know, AI, right there. And then then with you, you can go and let's say yep, yep. Need a device there. Yep. Anyway, be able to pull that together and then

14:50

and then we'll port our software to that device, and then install on their servers or cloud or whatever they have. Have those connect and then They have their infrastructure.

15:04

Where do you see it going? I mean, what? What's that evolution look like? I mean, I get it.

15:11

Well, if you look historically, yeah, any kind of compute, if you take it from a non Stream Analyzed perspective, but if you see any kind of compute, first computers were big mainframes, right? Yep. Then you went into client server, yeah. And then you went, you know, and then you went back into cloud. And now we're going out to edge. So but mainframes are still there. So this, this sort of tail will hang around, it won't change, it won't disappear. But this is what the evolution is. So you're going to be pushing a lot more of your compute power locally. Because people see AI as becoming a commodity. They're getting used to seeing it through all the chat bots out to all the chat, you know, the AI art. So now they get it. But they they're going to expect that my things are going to be really smart using AI. So you need to have that smartness directly on the device. And our vision is we're going to we're going to be the de facto standard for the industry for doing this. This is our vision. Call the be hag if you want. But this is what we say

16:12

no, that it sounds. It sounds accurate. And it's interesting, because there was this big, big push to get everything in the cloud. Got it, of course, but the cost the the is just like, wow, it used to be cheap, because we didn't have that much. But now it's it's killing us and we're dependent were super dependent on it. It's a big part, that you're just you're just being able to say, keep the important stuff here. But the important stuff where it needs to be exactly. Don't worry about all this other stuff. And we can work with you and help you through that. Yep.

16:52

And they're saving money that way.

16:54

Saving a lot for your reticle. Yeah, and

16:57

the other thing that's kind of interesting is the environmental aspect, right? Yeah, we don't talk too much about it. But it's always interesting to mention, though, about 4% of the world's energy consumption today is spent on the cloud storage processing. That's substantial. So if you can lower that, you know, and through this, you actually get the invite environmental perspective as well.

17:18

See, that's a whole nother conversation, because I don't my little brain is just struggled with that because you go around a event like this. And people are snapping shots and sending things into a cloud. And they're all just it's video that I just think about that. And it's going up in the cloud, it's going up in some sort of cloud baby. And that has to be cool. And it has to have some sort of place. I pay for my little, little portion of my cloud right there. God, you're awesome. Thanks. glad. Glad we were able to pull this off. It's always it's always I always give too much energy. Vice versa. Did she ask you this? Have you seen a different? I mean, there's a lot of activity, more activity at this conference

18:12

than last year than last year? Oh, yeah, I would say so. I think so.

18:16

That's how did they get a hold of you? And you know, sorry? How do they get a hold of who they want to contact you? Oh, they they

18:25

want to get right in here. Hey, Stream Analyze.com is a good one. LinkedIn, you talk to Daniel Spahr, I'll get you in touch with our company. That's too good. Wait,

18:35

there it is. We're gonna have all the contact information for Daniel out on Industrial Talk was his LinkedIn. Because you're on LinkedIn, I am gonna be your your great as always. Thanks, God. All right, thank you very much for joining Industrial Talk. We're gonna wrap it up on the other side. Stay tuned, we will be right back.

18:58

You're listening to the Industrial Talk Podcast Network.

19:03