The Truth About Predictive Maintenance: Copilot special - with Nat Ford

Trend Detection Podcast ·

Show notes

Welcome to the Trend Detection podcast, brought to you by Senseye Predictive Maintenance – which gives you visibility and insights into all your assets, from single machines to full plants to help you reduce downtime, increase knowledge sharing and accelerate digital transformation across your organization.

How everyday frustrations—like buying insurance—mirror the complexity faced in industrial environments, and why even small simplifications can have a big impact.

What copilots actually are, and how they help guide users through complex processes using natural, conversational language.

Why tools like Adobe Copilot are game-changers, and how their capabilities are being mirrored in industrial copilots like Senseye’s.

How interacting with AI doesn’t require perfect prompts—copilots are built to iterate with you and improve results over time.

How engineers and operators can use copilots to access insights from data instantly—without building dashboards or learning query languages.

You can find out more about how Senseye Predictive Maintenance can reduce unplanned downtime and contribute towards improved sustainability within your manufacturing plants, by visiting: www.siemens.com/senseye-predictive-maintenance

Transcript

So again, so defining what it is that we're, we're really discussing in this context, I think of, I think of, and I think it is correct, I was a co-pilot as um, in the context of software or computer-led processes, um, a piece of functionality that gives you another window into the thing that you are looking at. So in the context of SEDite, for example, a predictive maintenance platform, the co-pilot is an additional way. Arguably will become the main way, but an additional way of looking into the data that augments it with a greater, more worldly understanding of, of maintenance, etc. but the fundamental and important point of it is that it's conversational and you can have a chat with it about the thing in front of you. So it, it kind of brings the platform to life, um, but it doesn't do your job. Welcome to the trend detection podcast powered by Siemens, and now your host, Nel Sullivan. Hello and welcome to the latest episode of the trend detection podcast. Um, it's been a little while, but I'm really pleased to welcome back Nat Ford, um, a regular on the podcast, so welcome back, Nat. I hope you're well. Thanks for having me. No, fantastic to have you back. Um, I guess for, for regular listeners, we, we've done a, a pretty regular thing over the past few years. We've been running the podcast, The Truth about predicting Maintenance. Um, so I think we might be under that banner today, but. Um, we're actually extending our, well, we're, we're always looking to extend our topics out, um, from that. And today we're talking about co-pilots and what co-pilots mean both from in an industrial settings, which we know Siemens supports lots of different, um, co-pilot use cases, um, but also how that relates to what we're doing in our personal life, which I know Natt is a passionate advocate for. The use of it. I'm starting tentatively to, to get into it as well and just how, where the crossovers are and how they, they sort of link together, I guess. Um, but what I'm gonna do is, um, again, something we do quite regularly is I'll give that the platform to sort of set the scene of where we're gonna go and what we're gonna talk about. Um, so we'll do that first, and then we'll dive into the topic. So over to you, Natt. Thanks. No, thanks. And again, thanks for having me again. Hi, everybody. Um. Yeah, let's do a big fat disclaimer at the beginning of this session. So we're gonna be talking about co-pilots. Co-pilots are relatively new on the scene, um, and undoubtedly in a year from now they will have changed. So cliche, cliche, alert, the only constant is change. Um, but that said, let's have a chat today about what these new emerging co-pilots are good for, what they do, what they help with, what they're not, um. Picked out a few examples actually of other industries of where they've used so we can talk about the, the impact they have on different jobs and and roles, um, and I think it's an interesting conversation to have because I'm like all new technologies. Um, there's a lot of expectation in the market, a lot of hype, sometimes a bit of disappointment, so it's good to be clear in our minds on what it is that a co-pilot should be, so that when we're talking to our client base about our co-pilots, um, we're fair and reasonable, measured, um, and we know what we're talking about. So interesting times to be in sales where there's new emerging technologies that we're actually embedding in and making use of with our product set cos you're, you're part voyeur, part consumer, and also then going out and um evangelizing and talking about things. So it's interesting times, but just to loop that back to my opening statement, this is today my views, your views. Um, representative nothing but whatever's going on in our head at the moment, but should be, if anyone's interested in listening, um, an interesting conversation about where we're at in a minute. Yeah, yeah, it's our sort of interpretation based on all the conversations we're having with various different people and you, and yourself, in your case, um, with customers, which is actually where I wanted to start. So, when we're, when you're talking to. Uh, clients and customers and potential customers, what are they actually saying when we're bringing up the topic of generative AI or co-pilots? What, you know, Well, it's in their minds at the moment, but it's obviously most important. Yeah, you put two things into one bag there, and they do belong in the same bag, but actually they're perceived very differently. Um, I think, I, my perception is they're perceived very differently in, in the market. So, um, that, that for the sake of this, let's agree that G AI is a, a blanket term, um, you know, when you, you go to check GPT and get some fancy picture made for a slide deck. Um, you could argue that that's a co-pilot, but the co-pilots that we're talking about today are the ones which are, are self-proclaimed pro, um, co-pilots, or maybe that's a bit too sci-fi. The makers of them have proclaimed that they are co-pilots. Um, and they're there to assist with the process. So let, let's differentiate because there's all, you know, I mean I do stuff with check GPT all day every day, um. But this is very much about um an extra layer of something which assists with an existing product, I'd say. But anyway, to your, to your question. I think we've had a couple of things have happened. There's a re-emergence in the market of the great fantasists, people who think that finally computers can do everything and anything, who have somewhat unreasonable demands um and expectations from vendors like us. Um, and at the other end of the scale, there's people who don't really know. This is more than norm. People don't know what value something can bring. So, you know, as is always the case, it's our job to go and help them understand what, what extra value there is by using a copilot. So we're, so, yeah, so to, to summarize that. There's an extra layer to our role now when it comes to things like co-pilots, which is a tempering, tempering fantasy and helping people to understand the art of the gospel. And so I guess it's this with any sort of, if you talk about technology and softwareware in general terms, cos there's always sort of early adopters, right? The people are, I mean, you. You maybe described yourself as that, as someone who uses chat GPT every day. So there's that difference you're saying between the, the people who are like, get it straight away and want to implement it everywhere, personally, professionally, and whatever, and the others who are still either skeptical or don't just, uh, what is it? You know, I, I don't really know what you're talking about kind of thing. So there's a big difference between those two groups, let's say. There is, um, I, I mean, let's talk about it. What do I use, what do I use co-pilots for? So, on a day to day basis, I use all the, the, um, The probably the best known, the Microsoft Copilot. And I use it for fact checking. I use it for gathering information that's gonna help, um, augment or make better communication. I use it for ideas for slides. I use it for note summary. And I hasten to add, for anybody who's worried about me, um, summarizing notes in the conversation you've had with me, uh, we, we have a, an internal system which is, um, You know, you will not take information that we put in and make it generally available, so it's ring-fenced, um, so it's a, a professional version of the Microsoft platform. But I find it increasingly useful. I, you know, I. I think there's, there's a, there's a place and a role for that type of tool sitting there augmenting everyday activities. Making us all slightly better at the things that we do and faster at the things that we do. Um, I also do all the other stuff like getting it to write poems and songs and ridiculous pictures, but I suspect that will wear off in time. That's a kind of novelty factor, isn't it? What about you? Do you, do you have a, a daily? Usage of your co-pilot. Yeah, so, um, so I'll talk, I'll talk from both sides of the coin. So professionally, it sort of transformed, actually, cos we're doing a podcast right now, it's, it's the perfect use case where before I'd sit around and think about the perfect questions to ask you, Nat, or other people I'm interviewing. Now I just ask chat to UPT. I've got this subject, I've got this person. Think of 10 questions. Um, and the good thing is, the questions, as you know, for me, they're, they're not meant to be a script, but they're meant to be like where the conversation dips and dives and comes out of, and we might skip some. So it just gives me a general flow. I can send someone and say, these are the questions we're, we're probably gonna ask to get you thinking. That takes 10 seconds versus maybe half an hour, you know, it sounds like a lot, but sitting down and thinking about it, so it saves all that time. Um, we then summarize it afterwards. So we take the transcripts and summarize it, um, using internal tools, um, that, that you described there to, you know, to, to summarize the conversation, bullet point it, take out key quotes and things like that. So it's just say, again, the time saving is, is off, you know, for the whole team, not just me, you know, included in on the editing side too. Um, that's a bit concerning though, because, you know, given that I then used the co-pilot to give me all the answers I'm going to say, then we could really cut out the middle, Mandy, me and you. But actually, actually that brings me to a point, and we, we had an internal session across the, the business talking about different usages. Of different platforms. And I had a fantastic, a fantastic use case. That somebody put forward, which was prior to giving a presentation. He'd taken the topic and asked it to be discussed by, you know, with those, it's not co-pilot per se, but the platforms that will do a mock up. Um, webinar like this with a couple of, a couple of a fake, a fake gnats. And whilst he's not expecting any quality, um, the point is, it prompts his thinking. So by sitting and listening to that for 20 minutes, by the time he's listening to that, it's given him all of that sort of, all of those different topics and footholds to be able to then go and have an interesting conversation. I thought it was a great use. Um, and there are no rules, right? No, no, exactly, and that, and that sort of brings an interesting point in my, in my eyes, that, that, the fact that you can create a podcast, which, and I remember you sharing that is again quite scary cause I love doing podcasts, but it did sound scarily, um. You know, if, if you weren't, if you weren't in the know, you probably would think it was, um, you probably would think it was two real life people chatting about a subject. But, um, and I guess that's part of it as well, uh, in a couple of ways about do you believe the, what the AI is telling you? And we can talk about this from a, um, a personal perspective. I was actually speaking to someone from a G Gartner analyst earlier this morning. We, we, we talked about this, how. You know, you can ask questions and unless you know the answer, you just believe it, right? You just believe whatever it tells you cos it, you know, that it's, it's a great tool, right? So, uh, uh, I'm thinking of the, the angle to take this, but. In, if we talk about, you know, industrial co-pilots, you know, including sen sized co-pilot, how, how do we sort of differentiate between, you know, what's believable, you know, so it's not just hallucinating, I think it's the term, right? And how do we sort of navigate that with customers. It, it's a globally used term, I've never been convinced by the use of the word hallucinating because it suggests that, uh, one thing is a hallucination whilst another thing is clear thought, and I don't think either are true about any form of GI. They're just simply putting together words in response to a, a, a cue, um. But within Siemens, all of the co-pilots that we use that sit on top of some kind of general AI are um ring-fenced, and they, they will only talk about the things that they are supposed to talk about. Um, they are based on the information that has been made available to them for the purpose of the function they fulfill. Um, so I'm sure there's probably some, this, you know, it's emerging technology, right, but we, we, as an organization, we've put in place the things you need to, so you, it's not gonna go off and start, um, giving you spurious statistics or spurious facts. Uh, we probably need all of our data scientists on the call to discuss that and give a, a, a full answer. Um, but I think in the, in the everyday use, absolutely. I was, uh, trying to have some interesting conversation with a member of my family recently and I asked, I asked, um, A given general platform, if it could tell me interesting things that happen in a second at a kind of macro universal level, um, and a micro level. So it gave me a bunch of facts about things that happen and like numbers of stars born, numbers of stars that die, um, the amount of rain that falls on planet Earth and, you know, things like cell division within your body. And when I fact checked it, it was a beautiful response, but almost everything it was was incorrect at some level. But then, hey, if you ever, you know, you read a, read a newspaper, if you've ever been involved in a story and you know the insight, it's almost always factually incorrect. So maybe. Much better in those contexts than we can put in um value for, but I think, um, let's, let's steer back in from that conversation because that's a, that's a kind of a, a human-wide problem and within industry we have ways of resolving that, um. And let, let's think then about, you know, why, why do we want to have co-pilots in our life? We all, we all know how to do what we do, don't we? Um, so what is a co-pilot? So again, so defining what it is that we're, we're really discussing in this context, I think of, I think of, and I think it is correct, I was a co-pilot. As um in the context of software or computer-led processes, um, a piece of functionality that gives you another window into the thing or what you are looking at, so in the context of sensor for example, a predictive maintenance platform, the co-pilot is an additional way. Arguably will become the main way, but an additional way of looking into the data, that augments it with a greater, more worldly understanding of, of maintenance, etc. but the fundamental and important point of it is that it's conversational and you can have a chat with it about the thing in front of you. So it, it kind of brings the platform to life, um, but it doesn't do your job. Should we, should we jump in and talk about things that, that it can be and isn't, or did you have other things you wanted to discuss first? No, no, let's go into it, let's go straight into it, yeah, OK. So Let's think about what it is at first, or shouldn't be, right, so we can't determine any truths on this, this is our opinions. But a co-pilot, I think. You know, we talk about being conversational, but it's not a chatbot, right, just answers questions, it's something that should interact and help you do stuff. Um, it definitely shouldn't be a replacement for human judgment, and if you think of that in the context of Sensii, it's not there to make decisions about maintenance. It's there to enhance your decision making process. Um, it isn't, we, we've mixed this up a little bit this morning, we've talked about. Um Generic Gen AI products. Um, and co-pilots, but it's not generic gene. A co-pilot should be specifically created for and trained on specific use cases. Um, and become an expert in a very specific field and help. I mean, it could be other things, but that, you know, I think that's the sort of best delivery. It, it shouldn't be, you know, if we're talking about being augmenting by being an interface, it shouldn't just be another dashboard. It should be more than that, sit with inside the workflow. Uh, and be part of, you know, be on your side as opposed to being the dashboard. Um Everybody now is talking about agentic AI. It shouldn't be autonomous. It should be, it's still, you know, this is still, it's still a window into your data for you, right? Um I think that's probably a pretty good premise of things that shouldn't be. Oh let's add as well the thing that we always have to add about gen about um predictive maintenance. It's not magic, um, underpinned by. Um, logic, good data, good design, good human input, all those things. Yeah, and, and to, and to your point, you know, at the end of the day, if, if you don't take action from it. You know, but it's like, yeah, it's like, it's like, you know, in the sense like, if you don't take action from it, then it's, it's not much use, right? So it's actually, it's actually, yeah, I, I, I love, I've always loved that positioning of decision support, not, you know, this is gonna happen at this exact time. Kind of thing. It's always another tool in the toolbox. I think one of our customers once described to me, which I quite like. Right, exactly, exactly. And, uh, another way of getting more value out of the platforms that you're using and getting more value out of you, actually that's a good way of putting it, we're, you're getting more value out of the platform and more value out of yourself. As a user. Yeah, and that's where the, the personal, I don't, and you're right, we have sort of put generative AI and copilots in the same bucket. Maybe that's a good thing cos maybe people do put it in the same bucket and just interchangeably use them. So we're sort of breaking that apart a little bit and, and defining that a little bit, um, a little bit for people, cos. Yeah, I mean, maybe I'm using it interchangeably, so maybe I've got a little bit to learn there. But it's important to distinguish between the two, right? It's difficult, isn't it, because, uh, there, there's, there's things like chat GPT, it's just not easy to say, uh, but it's become a standard shorthand for all different platforms which are generative AI. Generative AI is not necessarily an understood term. Um, and then when you come to the co-pilot, because of the success of one organization, that's what everybody thinks it is. Um, but really it's a generic term that refers to. Anything that's a co-pilot sits within the system. So yeah, let's hope that over the next year these terms consolidate themselves and we get some general market agreement on, on what things are. We can try and drive that, but, uh, you know, but I think equal, yeah, it's interesting you mentioned that because you think something like Google is just synonymous with searching the internet, right? So it could be that, oh, let's just, well, it doesn't quite roll off the tongue like it's in the same way, but let's beat, which that it doesn't Google's, yeah, yeah, well done to their naming and branding because that just rolls off the tongue perfectly, doesn't it? But Yeah, but maybe, maybe it'll evolve into another name or maybe that they'll think about that a little bit more. But at the moment, you're right, when people think of that, it's Chachi BT which gives them a strong position in the market, I guess. Oh, absolutely, yeah, yeah. Um, yeah, I, so I was, you know, I was wondering that. So you're obviously, you know, both pers, you know, personally of a quite well described as like an advanced user. And again, with my conversation this morning with a Gartner analyst, he was talking about using it in a personal context as still as like a novelty. You know, I know you mentioned about creating poems and so, so you, you do that as well. But, um, and I know our colleague Alex uses it. I remember him saying like planning an itinerary for a holiday and all sorts of things. It's amazing stuff, but. …

Full transcript (WebVTT)