Agile Predictive Monitoring - Part One - With Burak Polat

Welcome to the Trend Detection podcast, powered by Senseye, an industry leader in using AI to drive scalable and sustainable asset performance and reliability. This is a new publication designed to help you go away with ideas on how to achieve maintenance efficiencies.


For this 3-part series, I’m joined by Burak Polat, CEO at Skysens, a company that is focused on creating user centric IoT infrastructures for every enterprise.

In the first episode of this series, we discuss the challenges manufacturers are facing rolling out enterprise technologies, the benefits of a joined up approach and the differences between a traditional and agile approach to predictive monitoring. I hope you enjoy it.

Please subscribe via your favorite podcast provider (Apple, Google and Spotify) if you'd like to be notified about future episodes and let us know your feedback by leaving us a review.


Key topics covered (click to jump to the section)

  1. What is agile predictive monitoring?
  2. What are the key challenges manufacturers are facing?
  3. The differences between agile and traditional predictive monitoring
  4. Subscribe to our podcast

Niall Sullivan, Senseye: Okay. So I'd like to welcome you to the latest episode of the Trend Detection podcast. I'm delighted to welcome Burak Polat from Skysens today. We've got a really interesting conversation coming up. We're going to be talking about agile predictive monitoring and how Skysens approaches that, which is going to be a really interesting conversation, I think. But first of all, as usual, to get started, I think it'd be great for Burak to introduce himself and Skysens and what you do.

Burak Polat, Skysens: Yeah, of course. Thank you Niall for having us. I'm Burak, I'm a co-founder and CEO of Skysens. Skysens is a US-based IoT company.

We started in Turkey six, seven years ago with my partners. We all have industrial backgrounds. We work for different industrial manufacturers companies for years. We all graduated from engineering companies.

I studied outer space engineering and then we work for manufacturing companies. And then we started Skysens actually, because we saw a dilemma of this digitalization for the industrial enterprise. They have to do it, they want to do it, but there is a gap between the real world and the requirements. So we start Skysens from that idea. We moved the company to US.

At the same time, we got a couple of investments from European investors and currently we are mostly working for manufacturing companies. We also work for logistic companies like airports too, or airport seaports. Sometimes we also work for city councils or municipalities, depending on the region, but for IoT applications.

Niall Sullivan, Senseye: Fantastic. And really good intro there as well. So what I wanted to start with actually, you mentioned the dilemma between the real world and the actual requirements. So maybe you could dig into that. Start with what that actually means in practice.

Burak Polat, Skysens: Of course. So there's a famous photo, which we also use in our presentation to our customers, that when you think about the industrial requirements or digitalization requirements, there is an ongoing operations on the field. Standard industrial enterprises, they usually are hurdling with hundreds or at least 10s of operations simultaneously at the same area with the different assets, which all coming from different vendors from the different times.

But when you design a digitalization project and when you try to implement some analytics on top of it, you need to get data from all of these assets in order to make it some meaningful results. So, when they decide to do something like that, for example, when they decide to implement some predictive quality applications, they often realize that they don't have data yet, or they don't have any means to capture that data or data is not clear, is not sufficient.

And also teams, usually cannot easily use that infrastructure, that capabilities in the end, even if after they deploy everything, even after they do everything right, teams usually cannot really easily use that capabilities because of the law of training and the leaving the company kind of situations.

They take that knowledge and they put it to some other companies. They usually do that. I mean not in a bad way, but carrying that knowledge with them while leaving the company. So in essence, according to multiple research, 75% of every industrial digitalization efforts face some kind of failure. Either they're not sufficient and they're not inefficient as they promise to, or either the project is complete failure. So that's coming from that gap between the real world and the world that we think there is.

Niall Sullivan, Senseye: Absolutely. And I guess a good place to start as I said at the beginning that we're talking about agile predictive monitoring. So first of all, maybe let's dive into actually what that is, what your definition of that is. And how it plays a role in what you were just explaining there in filling that gap.


Burak Polat, Skysens: Of course. So it's a new approach to industrial digitalization efforts. It means a system actually, which is easily deployable. And it also includes end-to-end application, end-to-end integration, which can be even used by the front liners very easily with no training for the analytics for AI applications.

System basically includes the platform, data integration, capabilities, data adapters to existing industrial data and wireless network technology, and also portfolio of H devices, which, 90% of these devices are Plug and Play with long battery life with no cable and requirement. We develop this because, as I mentioned, we see industry 4.0 initiatives are very hard for the enterprises because most of the industrial enterprise, they don't have resources to hire high cost consultants or large companies like industrial foreigners to do everything for them.

So they struggle in some sense, in some level of the applications, while they're trying to implement this solution to their enterprise. That's why we developed this agile predictive monitoring capabilities, which covers every level of the requirement, but in a agile way, in more basic sense.

Niall Sullivan, Senseye: And what other challenges are they facing, which you can help to solve?

Burak Polat, Skysens: You mean in, as a application wise or you mean... Can you clarify the question?

Niall Sullivan, Senseye: Yeah, application wise. That'd be a good place to start. So what sort of challenges are you seeing on the front? I know you say about relying on external consultants is one challenge, let's say. Just other examples in that area.


Burak Polat, Skysens: When you think about the industrial enterprise, they don't only have the maintenance problem, but they do have energy efficiency problem or sustainability, which is a very prominent factor in today's world. And also they have a problem about the quality. They have a problem about the production. Maybe some health and safety application or sometimes logistics.

So, they all have these problems and they all affect the final outcome, which is the product and the quality of the product and the profitability of the company. So what we see in the industry today is, they decide that usually the management at the sea level doesn't interfere with ongoing operation. They don't go to, sometimes they do, but usually what we see in the market, is they don't go to the shop floor and they don't usually solve the problem for the front liner themselves.

So I mean, the front liners needs to solve these problems by themselves. And it's, of course, with the help of the engineering management or with the initiative of the engineering management, but not the CLO.

So we need to give these tools to them in order to solve these independent problems, but at the same time, because all the data is required to solve these problems, are coming from the physical world, either by manual filled forms or manual measurement, or by either IoT sources, such as IoT Edge devices or the connected machines or things like that. So when we go to a factory, when we deploy this system, we help them to solve multiple things at the same time. Usually those applications are beside the predictive maintenance applications.

We also saw the energy efficiency question, how much energy they are consuming. At what amount are they consuming the energy when you compare? Are they consuming the right amount of energy when you compare to last week's for example, or the last batch?

So we also solve these type of questions. We also help them to monitor their sustainability scores in real time with integrate to electrical energy or the profile of their electrical consumption, either coming from green sources or not green sources and also monitoring other energy sources, such as gas, such as other sources.

We also help them to employ the quality applications, especially for the process when manufacturing, for example, chemical factories or pharmaceuticals. It is very ambient, very environmentally affected product. You can imagine. So we also help them to monitor their quality levels when conjunction to IoT data sources.

We also do health and safety applications, for example, air quality monitoring for the office spaces or the areas where their workers get their lunch or something, or in working areas and also monitoring toxic gases or the noise or the vibration, which can affect the worker health.

And also we help them to develop some logistical applications, for example, indoor or outdoor location tracking for the trolleys or for the forklifts, things like that. So we do multiple applications at the same time. The concept that we developed doesn't provide everything end-to-end. We don't always do very high detail in every application, but we provide them a starting point for that application. And then if they want to improve that application in some way, in more deeper understanding, we bring someone to the table. We bring our partner to the table with a perfect integration.

Since we already solved the data question and visualization and the data analytical question. For example, we work with you for more in-detail predictive the maintenance requirements. We start the application, we deploy vibration sensors.

We analyze that, we do visualization. We do some small predictive applications maybe, but if they require, for example, a more embedded, more integrated application, which they can scale up to their entire facilities, to entire maintenance teams, we bring Senseye to the table and we help our customer together with the help of Senseye.

So it is basically a starting point for the enterprise. If they didn't start this journey yet, it is very easy for them just to hire us and start with us. If they already start, if they already gathered some kind of data, we can improve their capabilities. We can improve their data acquisition or some analytics capabilities, some data capture capabilities with a very easy and very CapEx friendly way. So we either act like a starting point or force multipliers, in that sense.

Niall Sullivan, Senseye: Yeah. And what I wanted to pick up on that. It's interesting. So is your approach... A lot of the time it's difficult. I think what we find as well is, that a lot of the approach is on a plant-by-plant basis and it's not a joined up approach across the enterprises. Is that something you see quite often as well from your experience as well?

Burak Polat, Skysens: Yes, correct. For example, when you compare to Fortune 5000 companies and the rest of the manufacturing enterprises, you see that if the enterprise already started, they usually struggle to let every worker or every operator inside this application.

They're having struggles with that because once they try to expand the system to additional section of the factor or additional team to be part of this solution, they either struggle to get the data, they either struggle to integrate the system to their existing workflow, or they struggle to customize the solution that they already have to the conjunction of that particular team's workflow. So, they are either struggling with the flexibility of the solution that they already have, or they don't have any solution at all.

So, for Fortune 5000 companies, it is easier because they might invest more to get some consultant and get some additional solution or something like that. But when you go below the 5000, things get little bit complicated very easily.

Niall Sullivan, Senseye: Yeah. And why is it more... Let's dive into it a bit more. Why is it more complicated for those, say smaller companies?

Burak Polat, Skysens: Because when you think about the competitive edge, what it is the manufacturing plant is manufacturing, it's a product. They are manufacturing product and they're selling it to some profit addition. And then they're creating a profit out of it.

But that product, in order to make that product as profitable as possible, they have multiple things that they have to consider. They have to consider working environment. They have to consider predictive maintenance, the interruptions in the production. They have to consider about the current status of the asset and performance of the asset.

Either they are using it efficiently or non efficient, they have to decide. They have to think about the energy consumption in real time. They have to think about the quality at the same time. And they don't have the skill set to do that.

Even if they find someone who can do everything, who can understand the concepts and they can put all the pieces together, they usually lose that worker to some more higher level of positioning their career.

So it is very hard for them to keep skilled labor, who has the capability to understand all these pieces. But they also need a starting point. They have to start somewhere and they can understand that competitiveness only can be achieved through efficiency in their case.

And with all these global seismic shifts of manufacturing and supply chains in the world, they are very excited to provide more premium services, more quality products, increase their efficiency, but they just don't have the means to do that. And that's where our strategy comes in actually.

Niall Sullivan, Senseye: And as part of that, do they require a lot of education, say, from your side to educate them on the benefits and how to do things? And the secondary question to that is, do they do a lot of education internally? So say you've got a champion say, who maybe is on board, do you help them to educate and maybe let's say change the culture internally, for want of a better phrase?

Burak Polat, Skysens: Of course. The training and education is, I think one of the biggest factor in these type of applications for the successful initiative. We always support them, 24/7. We support them, we try to educate them, we have a training program for the product. But this agile predictive monitoring approach includes no-code applications. I mean, a no-code platform and Plug and Play Edge devices.

It is impossible to do every device, every application just Plug and Play, but 85%, we can be fair to say, is Plug and Play devices and also no-code applications. So we are try to lower the barrier of entrance to use this application as low as possible with the lowest technical training requirements.

So, for example, when you compare to very high-end data handling platforms, we only take couple of days in order to, how to get the data, how to manage the data, how to create a flows out of it and do some energy efficiency applications, which sensor to be selected, et cetera.

It only takes a couple of days actually, but when you compare to other high level platforms or more prominent players in the market, they need to give training for a month or something. And once that guy leaves the company, they need to train somebody else and it goes and goes and goes. So the end-user becomes a education company in some sense. We are trying to destroy that, trying to clear that problem actually for the users.

Niall Sullivan, Senseye: Yeah. We experienced a similar thing on our side as well. Is a culture such a key part of it? Well, I wanted to go back to agile predictive monitoring again, just for a sec.

So, I just wondered whether you could compare the approaches between a traditional approach to predictive monitoring and an agile one, what the main differences are and what makes agile so much better?


Burak Polat, Skysens: Of course. I think the first thing is the deployment speed and the cost to implement the system. The traditional systems usually comes with a very robust main platform engines, but they don't offer something about getting the data off the ground or integrating the results to existing legacy systems to work side by side.

Their usual game plan is to send some integrators, to send some partner to the field who requires to do everything to make ready the data for the platform and then integrate it to this platform.

So in our approach, we provide hardware and integrations directly from our main team, from our company, we move to on-site and do everything for them, with the continuous support manner. It's a very effective approach, especially for maintenance and energy efficiency applications.

The other thing I can say is, the traditional practices, they usually require a lot of training in order to start it working.

This creates enormous risk for the national enterprises, because, as I mentioned, when the person leaves the company, they have to start over and over. So, at some point it becomes meaningless for the enterprise itself.

And another thing worth to talk, is we always work with higher, more focused platforms. I mean, when we bring the solution to our user, we always tell them, "Okay, this solution does this to that point but if you require more detailed analysis, we have already integrated to this, this, this partners platforms." So they can easily integrate their system to higher level of solutions or in depth solution in that particular application.

For example, for predictive maintenance, we start with the data acquisition. For our customers, we enable vibration sensors, temperature sensors and do the visualization, set up the network, get the data and do some alarming or some small anomaly detection applications.

But then if they require to expand it to all of their facilities and all of their assets, we bring Senseye to the table and we tell our customers, "Okay, we can integrate our system with the Senseye in no time. We already integrated that. So you don't have to do any additional investment in order to have a more focused predictive maintenance applications." So, we always provide that option to our customers. So it's a part of our strategy when we say agile predictive monitoring capabilities. But if you go to other traditional systems, you don't see that approach.

Niall Sullivan, Senseye: Oh, absolutely. You mentioned about integrations and I'm just wondering how, agile predictive monitoring, how does it integrate and how does it fit with other IoT platforms? How do they complement each other in the process? And maybe you could also touch on the issues with legacy systems as well, like connecting those together as well..

Burak Polat, Skysens: Yeah. Predictive monitoring is a starting point. We always tell our customers, it's a starting point or it's a force multiplier for the existing applications. You can think it either way. So, if they have a platform already in their usage, we can expand their capability with our predictive monitoring platform.

But if they don't have anything, it's a very good starting point. Because we are closer to physical space, we can integrate directly to legacy systems, push or pull data to, for example, systems or to any mode base or any other protocol based PLCs with conjunction to existing IoT platforms.

Agile predictive monitoring capability has a smaller analytics capabilities, but the users can also do visualization and data flow management at the same time. So for the higher analytics capabilities, they can still use their IoT platforms if they already have. So, it's a good starting point and it can work as a multiplier for the existing IoT platforms.


Subscribe to our Trend Detection podcast

We post weekly episodes on a wide range of different subjects including predictive maintenance, industry 4.0, digital transformation and much more.

Here are all the places you can access/discover the latest episodes of the Trend Detection podcast: