Rethinking Predictive Maintenance - Episode One: Closing the Knowledge Gap - with Rachel Green
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.
This episode launches Rethinking Predictive Maintenance, a new Trend Detection podcast series exploring how manufacturers can retain expertise, capture new forms of machine information, evaluate predictive maintenance and expand successful approaches across multiple sites.
In this first episode, host Niall Sullivan is joined by Rachel Green to explore what happens when experienced manufacturing and maintenance professionals leave the workforce, taking decades of knowledge with them.
Rachel explains why technical information alone cannot replace the experience, judgement and credibility built over a long career.
She discusses how manufacturers can reduce their dependence on a small number of experts by capturing maintenance history, adding context to machine data and using predictive maintenance and AI to help people access the right knowledge when they need it.
In this episode, you’ll learn:
Why critical maintenance knowledge is so difficult to replace
How reliance on a small number of experts creates operational risk
Why standard procedures cannot capture every real-world maintenance scenario
How predictive maintenance and AI can make experience more accessible
Why technology should support experts and develop the next generation, rather than replace people
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
Being able to, you know, on the fly, work with a tool to support you in that is really quite a bit of a game changer and it's basically being a substitute for having that expert who may not be available in the moment. Welcome to the Trend detection podcast powered by Siemens. Hello and welcome to the the latest episode of the Trend detection podcast. I'm really pleased to be joined by a new, I say a new colleague, you are relatively new, Rachel, but also new to the podcast is what I meant as well, so I'd love to welcome you to the podcast first of all. Thank you for joining us today. Thank you, Niall. Happy to be here. Fantastic, fantastic. We've got, I was about to say we've got a really cool topic. We always have cool topics on this podcast, but we're talking about something that's very, you know, you know, incredibly topical, and it's not AI, although AI is going to play a part in this conversation, no doubt. But we're talking about actual, you know, people, and more specifically, the, the experts in the field or on the plant floor. Actually, what happens is Experts who have been working year after year for many years, lots of knowledge, lots of experience, no machines at the back of their hand when they leave. And we know that's a common sort of problem that, you know, gonna have to face very, very soon and into the future. So, we're going to cover quite a lot of topics within that, but I guess before we get started, Rachel, would be great to just introduce yourself, your background and your role at Siemens as well. It'd be great. Absolutely. Thanks, Niall. So, my name is Rachel Green. And I've joined Siemens with the goal of supporting predictive maintenance powered by Sensai. And in this role, I'm talking with a lot of our customers, a lot of manufacturing and industrial folks who are facing a lot of challenges on their shop floors. And oftentimes those challenges extend well beyond just, you know, how are we getting this PM done? How are we avoiding unplanned downtime. They're really dealing with who's showing up for this shift. Do they have the expertise in-house to deal with the problems they're facing? If they're outsourcing it to a vendor, you know, how are they overseeing those vendors services? And I'm coming towards this having spent the last decade in supporting different manufacturing IOT technologies. And it's been an amazing time to be working with manufacturers, having gone through the challenges around COVID, where we saw huge leaps in technology adoption because we had to, not necessarily because people wanted to. And then some of the supply chain challenges that have come after that. It's been a fascinating time to be in this industry. And one of the things that I like to highlight is all of the time that I've been working with customers, bringing them technology solutions, the recurring theme that they've always had is, Rachel, what we really need are the people to embrace these technology solutions and help us drive the change management that we need to make them daily solutions. And so, that's kind of been the recurring challenge and theme throughout this part of my career, so much so that I actually started teaching a manufacturing IOT technology course at one of our local colleges here in southeastern Michigan. Even if I could only turn out 1012, 20 students per semester, figuring that that would be part of, excuse me, part of helping that change management. Fantastic. And I guess that, you know, if we're, we're sort of maybe coming to the end of, like the, the, you know, full circle of what we're going to talk about today, but I guess that's part of the solution, right? You're sort of part of the solution in terms of what you're doing at Siemens plus what you're doing there, which is, is really cool. Well, well, I guess, so let's start with, if we think about sort of the scale of the challenge, because it is something that's talked about a lot, you know, lost knowledge leaving the manufacturing sector. So, what is the scale of this challenge from your conversations, Rachel? Yeah. I think for every leader, for every manager that's out there, they see the statistics, you know, I think some of the latest ones are, you know, that we have around 1/3 of the manufacturing workforce that is actually eligible for retirement. I'm sure we can fact check that. It might be a little bit on either side, and it probably varies by vertical, but either way, that's a huge number, and those are some of the most seasoned people in these facilities, the most institutional knowledge, the most respect. So, it can be a huge challenge, and over the next 5 years, this is definitely something that a lot of companies are going to navigate how they handle the retirement of key employees. And the biggest challenge is that there's not really a backlog of workers, a pool for them to hire from to replace that experience. And, you know, it's one thing to read the statistics in You know, the National Association of Manufacturing always puts out a study to help us understand what trends are on people's minds and what the statistics are. But the reality of it is, is, you know, I go to maintenance stand-up meetings with teams and look around and you can see, hey, if, if we're 1 year out, that person's probably not going to be here. They're saying this person is 6 months out from retiring, it looks like several other people might also be close to that, and just imagining. How that meeting would be different without those contributors is a huge impact for companies to navigate. Yeah, no, and absolutely, and you can imagine that as well, even with, because I was thinking even if you've got time to prepare, but I guess some of that much knowledge, that much experience, even 6 months, even a year is probably not, not long enough to, to really plan for that, because. Yeah. And they, you know, a lot of them make jokes about it, and they'll be like, this person has been driving maintenance for the last 37 years in our industry, and then they're like, and we've got this person who's 3 years out of a, you know, university program, and trying to bridge the gap between those two. And some of it I think is technical knowledge, but some of it is, if you've had 37 years to build up your who you are, what you've done, how you've saved the day, and especially because maintenance is not only doing the. Individual actions every day, every week, every year, but there's also some element of the hero because these are the folks that come in if there is some type of unanticipated failure, if there is some type of breakdown, and get everything back online. And being the hero over that time builds credibility, you know, these are people that can hack the situation to get something back up and running, even if it isn't the permanent solution. And that heroism has a following. So if you're coming in new and trying to build that credibility, build that you can also be the hero, that you can be someone that's confident in. It takes more than just technical knowledge to be in that role. No, yeah, it's a really interesting point. Yeah, yeah, the, the credibility as well, because actually, it also feeds into the, you know, if you think about the technology and you mentioned it in your introduction as well, sort of the change and adoption, because it's also, we talk about the trust in the technology as well, and part of the reason why sometimes it's harder, or takes a bit of time, let's say, to trust. The technology is because you've got these amazing people with all of this knowledge inside, so it's like we've got these people here who know, know all this stuff, right? So, so is that, that's part of the conversation as well, I'm sure. Absolutely. And one of the most telling stories that I've had in the last couple of years is, you know, machines have a tendency to break down once dinner is being served, once it's dark outside, once people's kids are on the sports field, or if it's a holiday. So, of course, this was a 3rd shift story and I had somebody who basically was reporting in that they were called from the plant. They were just going to bed. And the pivot that they, like, experienced viscerally is instead of grabbing for their car keys to go to the plant and help, you know, grab a tool and get involved in repairing the equipment, they first grabbed their laptop and they were going to the data being collected by these tools, by remote sensors and wanting to see what the data is telling them. Before needing to see the actual equipment themselves. And that change in mindset between knowing that we can rely on data to help us remotely through some of these technological tools versus having to be on site, particularly in the middle of the night, 45 minute drive from the plant, disturbing your family, you know, all those types of things. I mean, these are real pains that people are having. As you know, manufacturing facilities are never in the most convenient locations, so most people are commuting and to get back there in the middle of the night in our part of the world, you're always dealing with deer and all that kind of stuff. And I mean, this is a real change in people's workload to have data to rely on first instead of having to be there in person first. And again, to your point, I mean, if you take that point even further with the topic we're talking about today, so if you rely on one person who. Or, or a small number of people, I should say you have to go to the plant. It becomes even more problem, you know, problematic because you can see, are they available? Are they, you know, can they get to the plant in a reasonable time and things like that. So, if you take the data point away and there's a few really known experts that they rely on, and that's. Yeah. This was so exemplified. I actually talked with a company and they, their maintenance strategy. Was jokingly referred to as what would David do? And they had a person named David, who was essentially that hero for them. And to the point where not just at the facility he was at, but they would fly this person to other facilities to consult. They would get his input, all that type of thing. He had become the expert, but also the hero in this to help drive confidence in maintenance changes, repairs, planning, those types of things. But to your point, What happens if David wins the lottery or decides to retire or anything else, you know, we try and design systems that are not reliant on a single person or their expertise because we want to have the business continuity be more flexible than that. And you can always have that person's input, but being able to involve others and upskill others and rely on others is really the kind of thing where having a tool in place to do that will give you a lot more confidence in, in the business long term and not have a scenario where you're just relying on one person. Obviously, you might still have standout leaders and amazing experts, but having underlying systems to support everyone in becoming the next generation standout expert is really important too. Yeah, and I, I guess part, part of the, part of the issue as well is that the knowledge, you know, if we compare it again to the technology side is the knowledge is sort of trapped in. These people's head, it's one thing relying on them personally to fly them out all over the place to do it, but even if some of that knowledge was captured and, you know, stored and easily accessible, that doesn't solve the issue, of course, because it needs to be interpreted and nothing beats experience and knowledge, but yeah, so I, I guess. Yeah, I mean, I guess you sort of explained it, but what, why this knowledge is sort of trapped in heads and not documented anywhere, because which would help, I guess, towards this issue. I mean, I think, I think the want for us to be able to. Approach it in a knowledge-based way, in a standard operating procedure, and SOP is fantastic and the world is definitely gravitating that way and we want to have checklists and SOPs for everything, and I absolutely think those are critical and important. But the need for maintenance and predictive maintenance is because of the chaos of the operating environment and the variables that are introduced by the process itself, where the process is being done, who's the operator running it. And so you've got these factions basically at war. You've got the, you know, the SOP which tells you this is exactly the way it's gonna go, and then you've got reality, which tells you, well, you've got different operators, you've got different processes, you've got different products which might have different variables, and maintenance is kind of a, a chaotic space and especially if you've had a breakdown. You're going to that machine and just knowing they had to shut it down. Well, did they have to shut it down because quality of the product disintegrated, or did they have to shut it down because it's starting to rip itself apart and throw parts and pieces all over the floor? Those are two very different scenarios, but both reasons to shut a machine down, and it dictates a very different. Places to start in your troubleshooting technique. And so, one of the challenges for how do you create change management, how do you create a system that supports this, is creating something that's agile enough to say, here's where you start, but your troubleshooting might go in very, very different ways. And that's where some of the conversational IA is really powerful. We're seeing that ability to be agile enough to maintain the conversation during the repair. OK, start with the least invasive, go to the most invasive in terms of troubleshooting technique, but also, if you're finding this as you're doing your oil analysis, then pivot to this type of troubleshooting and being able to, you know, on the fly, work with a tool to support you in that. is really quite a bit of a game changer and it's basically being a substitute for having that expert who may not be available in the moment, may not be available anymore if they've departed the workforce and trying to stand up that institutional knowledge. Yeah, and I, and I see that as well, it's what, yeah, what, what I was thinking is that. You know, back to my, my point about the having some, you know, even if you document everything right, but you still have to sort of interpret that, that knowledge in some way. So, the tools you're referring to, you know, Germanative AI, I knew AI would come up. So, that isn't a big take, to be fair, but, but we know that's interesting. how you position that as sort of like, you know, not, you know, careful to say replacement, but when, you know, if the expert or the small subset of experts are unavailable, then you've, if you've captured the knowledge and documented it, then you could, you know, you, you could use a tool, AI to help. Yeah, and work with it, ask the questions you need, troubleshoot certain, certain options, look back at previous cases, if we're talking about Sensa, particularly compared to previous cases, things like that, and that's where you get, you know, not a complete replacement, but rather than, you know, you could ask, you know, what happens to go when it went down or whatever, you, you've got it documented, you can ask, you can interrogate the system and that kind of thing. Yeah, and, and I think, you know, there should be quite a bit of sensitivity around what are we replacing and what are we. Making more efficient because if we can get that person to spend more time on the floor, acting on the maintenance activity instead of going through old work orders, trying to find the maintenance guide, trying to call the person that did the last repair to find out what happened then, less time on the investigation or creating the plan and more time executing. I think that's really the, the variable there, and being able to execute with a high degree of confidence, even if you have less expert input is really where we want to get people. So, this might be, you know, your second shift maintenance person and instead of Having a 45 minute conversation with an expert, can they use a tool to refer to previous work orders, refer to previous outcomes, make sure that they've consulted the diagrams, the manuals, you know, all the spare part inventory, and then be ready to go, and maybe it's a 5-minute conversation with an expert in. Instead of a 45. Maybe it's just text exchange while they're getting themselves out to the equipment and getting the tools and the spare parts kitted and ready to go. Those are things that make a real difference when your downtime is being calculated by the seconds or the minutes. And that's, that's a real scenario. And I think the idea as well, I mean, yeah, again, easy to refer back to, to Sensai, but what I think fits well is, you know, in terms of prioritization as well, you know, the, the, the technology can, you know, it's not raising cases on everything for every small little deviation, but it's sort of pointing attention to where it needs it most. But then, and then subsequently, if you do need an, you know, a human expert to come into that, you can then focus their attention on what's. Needed most and not sort of running around trying to, I know, putting out fires is the typical phrase, isn't it, but, yeah, but it's, it, it's a lot more focused, it's focusing that expert where, you know, where they can be most, where they can be most useful, basically, rather than spreading themselves too thin across lots of different things. Yeah, I mean, you mentioned putting out fires and sometimes that's literal and sometimes it's figurative, uh, but, you know, the, the maintenance team, I have yet to go into a facility and have had them admit that they've got too many maintenance experts, that they've got too much, too many resources in those types of departments. I just am not seeing it in the US. So, we're coming in assuming that this is a department that is strapped. There's not enough time for the activities they likely need to do. And so, every day, every hour, they're basically balancing how are we spending our time and how are we getting the most output for that time. So, there's the, you know, what fires are we fighting today, but as you talk about tools like Sensai and some of the others that are out there, one of the things I love to do is take it into more of a strategic look. If you have a facility that has limited planned downtime, so maybe you do an annual shutdown, and that shutdown's 4 days, and during that day, you're going to bring in tens or hundreds of vendors to get many, many. The activities done because it's the only time you take the plant down. Using Sensii or a tool like that to prioritize those activities, using Sensii to tell you, hey, we had this on the list, but that equipment is still in a green status, maybe it doesn't need to be done on this shutdown, maybe we push that off. That can be just as helpful and valuable of an activity in terms of prioritizing what you need to do. It also pushing off what you can push off because you have strapped resources, and those types of use cases are really what our team spends time, you know, working with our customers on to say, OK, you've got this data, you've got this activity, how do you work through this and what's the best outcome for your facility based on the data that you're seeing and the resources you have. And this is what our customer success team does day in and day out is to help our teams with that, and it's a real challenge. Yeah, and, and yeah, it's interesting because you mentioned, because I know, of course, for, for a number of reasons, mainly, you know, military is, you know, downtime is usually or unplanned downtime is usually the big. …