 From San Francisco, it's theCUBE, covering OSISoft, PyWorld 2018. Brought to you by OSISoft. Hey, welcome back, everybody. Jeff Frick here with theCUBE. We're at OSISoft in downtown San Francisco for PyWorld 2018, about 3,000 people talking about really OT operations as it slowly marries with IT, and industrial Internet of Things. They've been doing it here for a long time. We're excited to have our next guest and practitioners out in the fields who've got solutions, and they're actually doing good work. So, Sankeet Ampakar, he's the SVP of Falconry. Sorry about that. No problem. Welcome, and Michael Ricci, the VP at Seek, welcome. Thank you very much. So, before we get started, just a quick summary of what your companies do. Sure, so in the case of Falconry, we do what's called operation machine learning, and what it means is really applying machine learning to business operations and data analytics to really drive improvement and efficiencies in their operations. And what's unique about that is it's kind of like a data scientist in a box. So you don't require data scientists on your side, you can actually have your own practitioners and operations people use it. And just plug into your algorithm. And just plug it into what existing their infrastructure is. Okay, and what about Seek? So Seek is an analytics application for process manufacturing data, for example, OSISoft Py. And really, what our focus is, there's incredible innovation out there. The open source and the machine learning and the big data and so forth. And we're about closing the gap between what's possible and what's practical in terms of the applications that people use every day in process manufacturing. So it's just funny, because big data's all the rage and machine learning's all the rage and AI and industrial internet of things and IOT. And yet these guys have been doing it for like 40 years. Without IP based sensors. Without 5G, without Hadoop 40 years ago. So why have we not heard about this and what kind of opportunities now open up when the rest of the IT infrastructure space and we do get 5G and we do have IP connected devices and everybody's ready to get the sensor data. It's a whole new revolution. Exactly, because what we're seeing right now is people have data in their systems. They just haven't leveraged it to the full capability. So as you start getting more and more data and especially if you have a Pi system, you have access to all that data now. How can you fully leverage what you have and really drive new insights from that? And that's really what's driving all this stuff. And you brought up some good points like with Wi-Fi and 5G and other sources where information which initially was not connected can now be connected. You have now full visibility into your entire systems and you can actually be able to control things that before you had to send a person out there and kind of go and tweak and turn and get working. So it's really changing how you digitize your infrastructure. I think that's, it's become a bit of a buzzword unfortunately, but the digitization of your industrial operations is actually real and it's happening right now, right? It's funny you bring that up because you could argue the original big data was manufacturing data and they were just missing a branding team to call it something cool, right? So the original big data was manufacturing data, there's a lot of it, there's been a lot for a long time. Now they are ahead in the sense that they know how to store it and do a great job with the Pi infrastructure for example. And now, as you said, it's about that next step not only for the manufacturing environment but for those IOT environments that are just starting to collect and process data. So now if we can close the gap on modern analytics, right? And modern analytics capabilities with the data they've collected, what that means is businesses are going to get more benefits. It's not about sensors, it's not about data collection. It's about business benefits to the bottom line, the ability to see it and get insights from data. So it's really interesting because so many startups get started because they see some inefficiency, whether it's empty rooms that can be Airbnb or it's cars that sit 90% of the time that can turn into an Uber or a Lyft, you would think that in some of these old line manufacturing that a lot of that inefficiency would have already been rung out but as we keep hearing from stories here time and time again, whether it's getting better yield out of your gold ore or getting better yield out of your water system, there's still a ton of inefficiency and opportunity yet to be extracted and that's before we add machine learning. Well, that's the difference between I've got the data and I've got the science or I've got the calculations. It's too hard and takes too long to get the insight to impact the outcome, if that makes sense. If it takes me more time to do the analysis in the spreadsheet or pen and paper to impact the outcome of the batch, I'm not going to do it. But with these modern analytics, hey, I can get the insight quickly and I can make a change to what I'm doing or prevent something from happening and now it's worth doing. So I've got the data, got the insights. And if you think about it like today, for example, you have control systems in place that have been there for 20 years that basically can do what we call real-time control. So you're doing a batch process and you're monitoring that stuff, it can do that stuff perfectly well. Does it make sense to put something new just to get another 2%? Maybe not. But what about if you can now predict, not just real-time, but predict what's going to happen six hours, 12 hours, two days, a week ahead of time? That's entirely brand new. And the problem is looking at your data you have today there's just way too much data for you to humanly possibly do that. So therefore it never really got touched as much. Now as you have the tool sets that have come from the IT side have come from the big data analytics, now you apply them over here, suddenly you're uncovering basically net new benefits that you can get that just before were not easily accessible. Right, right. Yes, go ahead. I was going to say 30 years after all the data was created and collected, unplanned downtime, right? Is still a bugaboo of so many of these industries. Unplanned downtime means whoops, we didn't expect that to happen. Machine failure, something going down. Another set of analytics is going to be required to really stamp that out and no things in advance as you just pointed out. So what are the notions that gets kicked around a lot, right, is data's a new oil, right? And, you know, I'm not going to go there. But one thing that is clear is that data used to be a liability, used to be expensive to store, right, expensive to keep. And you hear time, I mean there's a really great movie sponsored by EMC, this big data movie that they did. And they talked to horrible story about, you know, these EKG machines that would be kicking out data all the time on a tape that would go to the floor with predictive data that could tell you when someone was going to have an issue, but the nurses only came in and checked it once an hour for whatever the protocol was. It's just horrible. So, you know, have the industrial companies now realized that beyond what's on their balance sheet and their capital expense and these huge infrastructure projects, they actually have a lot of value in their data. We see it in tech companies all the time. Why do these companies have this valuation? It's not a multiple of revenue, it's because they got the data. But we haven't really seen it morph into kind of modern, not modern, more old line asset based companies where there isn't a line item yet. Soon there's going to be interesting to see how the accounting principles change, where you get credit for this data. People getting it now, are they seeing the value? Absolutely, they're getting at the pressure that they have to now realize the benefits of the data possibility, right? Mean that they recognize, look, my next benefit out of my balance statement comes from my, McKinsey calls it competing on analytics, right? My ability to do analytics drives that balance sheet results, okay now, what are the right analytics and what am I looking for in terms of outcomes? So they absolutely get it, it's just been too hard. The gap between the innovation in our consumer and IT lives and what's been generally available in the OT space has been too high for too long. That's what we're working on closing. And there's two things, actually bring up a good point with the McKinsey article because McKinsey's predicting that 20% of the next, actually the next 20% increase in productivity rise has actually come from data analytics being applied to manufacturing and being applied to process operations, right? And it's interesting because it's not like this stuff did not exist before. If you look at it right now, there's about 15% adoption rate of advanced analytics in manufacturing. I'm not talking about your standard real-time stuff, I'm talking more of the advanced stuff. But if you look at the adoption and what's expected, by 2020 they're saying that's gonna go up to 53% of all manufacturing out there, all process automation out there. So what it means is right now, this year, 2018 and 2019 is we're gonna see a huge amount of adoption where people have been doing pilots until now maybe or doing a little bit of trials up to now are actually, they've actually stepped in and we're seeing real purchase orders for real production applications and it's happening in every industry. That's the interesting thing too, it's not just okay. Before it used to be semiconductors are leading or automotive is leading or maybe oil and gas. We're seeing it pretty much in every single one right now because everyone has the data, everyone knows it's not being utilized and they're saying where can I get my next advantage from because it is a competitive advantage now. If your competitor is doing a better job with their data than you are, then you want to make sure that you're able to leverage it itself. Goldman Sachs actually built on that, wrote an article on productivity in oil and gas and chill from brawn to bite to brains. And the whole point was the next chunk of innovation is going to come from the brains and the analytics that are possible and how to optimize those outcomes. So it's very clearly seen. So the other buzz that's happening right now is that all the machines are going to take our jobs and we're going to have universal basic income, lay on the beach or be laying out here on Market Street one or the three, I'm not sure which, but clearly the evidence is contrary and really we're seeing that here, especially with some of the stuff even without the analytics. It's the combination of the machine with the data and a little bit of an application on top of that to enable people to make their decisions. And some of the use cases that have been coming out of the show are fascinating to me, the scale of the impact. One of the water companies, it was like 50% of the water between the time it goes out of the processing plant to the spigot at the house, 50%. These are huge inefficiencies. So the opportunity just seems endless. I was just going to say, so do you have any of your kind of favorite stories where it's just like mind-blowingly, in hindsight maybe obvious, but it wasn't at the time until they actually dug into this data? Sure, so you bring up a really good point because it's not really about replacing any work or it's actually augmenting what the work can do. You're making them much more efficient with what they're able to do because they're the ones making the decisions at the end of the day. So there's a couple of interesting use cases that we've been seeing. And I'll give you one coming from the mining side where, for example, they've been having an issue where on the conveyor belt, depending on the quality of the ore, that ore was starting to get blocking into the part of the machine that does the crushing and does the grinding. And that when it goes down is about $30,000 per hour. Takes them somewhere between five hours to a full day. So that can be like $720,000 per day. And it happens twice a week. So you do the math. That's lost productivity. That's lost productivity right off the bat, right? That happens twice a week. And this is not, this is not like a massively large company. This is like a mining company out in Wyoming having an issue like this. So obviously there's a big problem over there to solve. And the beauty of it is you can take the data. The data can actually anticipate and say three steps before it reaches that grinding part of the cycle that this batch of ore, which is moving through right now has a problem. And therefore what they were able to do was they were able to go and slow the process down. So you're still having output and productivity. Have the ore removed and then basically continue the process on. They've got to the point where they're so confident now that the actual operator now is able to close that loop remotely and basically whenever that warning happens they can say yes, here's a bad batch, automatically get it taken off and keeps going on. So, but you have the operator in the loop. The operator is the one making the decision what to do about this. This is not being done for them. So it's not, while it helps in automating, it's not an automation. It's still a person in the loop. And that's always going to be the case. I just think one of the things that Falconry and C-CAP in common is that focus on the engineer or on the operator, the person, and then taking advantage of their expertise, their experience, their education, they know a lot about those plants and assets. It's just too hard to do the analytics by hand. So if they can use the Falconry or CEC to get the insight more quickly then they get the better production result. But tapping rather than replacing that expertise in that engineer, that front-line worker, absolutely critical. Because there's 20, 30 years of experience in some of these plants and some of these assets. You want to tap what they know. Because they've seen it, just help them do something more quickly. Soft institutional knowledge is really hard to replicate. Absolutely. And I still keep hearing about everything on Excel too. It's just a fascinating market penetration of Microsoft in Excel. 30 years later. I have my data on a CSV file. Can you do something with it? Yes, can you do something with it? And it's from three weeks ago and I finally figured out the export. So before I let you guys go, kind of the thoughts on the show, we're here at OSI Soft. Have you been here before? It's our first time. I see people walking around on the 15-year badge. This is just amazing. It's like the most successful company you've ever heard about that's right across the bay and been operating for 40 years. So kind of general impressions, some takeaways from some of the sessions. What are you guys here for? So OSI Soft does a really great value for essentially the industrial operations team. Because basically they're bringing them data that actually can really change what they do in their operations, can really make a big difference. And in terms of the users, they're very sophisticated. They actually, you don't have to convince them and say, hey, data's important. They know the data's important. They have been doing stuff with their data and they're able to actually show really good use cases. If you go into any one of these, I was sitting in the transmission distribution one and it's amazing, even in industry like transmission distribution, which you think is a regulated industry, have been doing tremendous amounts of stuff in terms of how they've been using the data of their pie system and improving operations and actually making things much more efficient for you and I to your point, that there's so much of loss between the energy generated to a finite reaching your light bulb at home and imagine that they're making significant improvements in that so that there's less loss of power when it comes to you. I mean, it's more benefits for all of us basically. Oh, for sure. I just think, it's funny you mentioned the OSI Soft. Is it known and I can see and understand that but this is the largest user conference they've had. They've doubled the partner space that they've got. 3,000 people here. So I think the recognition of, okay, before I can get the insights from the data, I've got to have it sort of well stored in the high infrastructure, is growing among organizations. So that's why you see the growth in the user conference. And once it's there, then we can kick in. Advanced analytics on top to go from the data collection stored and managed to the insights that drive better business outcomes. Yeah, it's in so much easier, right? To get those efficiencies versus rip and replays. Leave the data where it is, get your engineers involved. Just fix the leak. 50% of my water's coming out of the leak, it's crazy. All right, second Michael, we got to leave it there. Thanks for sharing a few minutes with us. Sure, thanks for having us. Very much appreciated. All right, I'm Jeff Frick, you're watching The Cube from OSI Soft 2018. Thanks for watching.