How Generative AI Turns Maintenance Data Into Action with Jay Hack

Your CMMS probably knows more about your plant than any one person does, but most of that information is buried in reports, menus and half-completed work orders.

While we were at Accelerate 2026, we sat down with Jay Hack to look at four practical ways generative AI is being used to make maintenance and reliability work easier, faster and more accurate.

The first was the ability to simply talk to your data inside eMaint. Instead of needing to know how to build reports or search through different parts of the system, you can ask a normal question and get an answer back from the CMMS.

But as Jay explains, the AI itself is not necessarily the hardest part. The real challenge is making sure the right people have access to the right information, especially across different sites, departments and levels of the business.

We then looked at two areas that could make a real difference on the plant floor.

The first is automated SOP generation. The system can scan OEM manuals and technical PDFs and turn that information into practical procedures that can be added to work orders. There is still a human involved in checking and approving the content, but it could save a huge amount of time and help improve consistency.

The second is voice-based work requests. Technicians and operators can speak naturally into the system, even while they are out on the plant, and the CMMS can then populate the relevant fields.

Instead of receiving a work request that just says “pump broken”, you can capture what the operator saw, heard or experienced and create a much better maintenance history.

We also discussed how AI could support global teams by searching document libraries, translating manuals and helping standardise maintenance and reliability practices across different sites and countries.

Looking further ahead, there is also the potential for AI-powered competency mapping, using a person’s actual work order history and experience to better understand skills and identify development gaps.

This was a really practical conversation about where AI can genuinely support maintenance teams, rather than just adding more technology for the sake of it.

If you are interested in AI in maintenance, CMMS adoption, preventive maintenance and digital transformation that actually works, give this episode a listen.

What is the first maintenance workflow you would want AI to improve?

Transcript

Welcome From Accelerate 2026

SPEAKER_00 0:00

Hello, welcome back to another episode of the Reliability Gang podcast. I have a very special person with me today. I would like to introduce Jay Hack. How are you keeping, Jay?

SPEAKER_01 0:08

I'm keeping well. It's great to be here. It's great to have you at Accelerate 2026. And this is just a fantastic conference, and I loved sitting down with people like you who are out on the front lines and leading companies and leading people, moving maintenance and reliability forward.

SPEAKER_00 0:24

What an event it's been this year. Honestly, I've enjoyed it so much. Every time I come in, I get invited, meet so many people, so many people network, and so many engineering managers and people that obviously adopting the e-mailing software or other products and seeing everything like that. And I've got to say, like, Jay brings an incredible energy to this as well. So, like, really honored to have you on here to have a chat about some of the new features as well that we've seen. So, Jay's been kind of giving me a little bit of an insight to some of the cool features. And one thing we did do on Fort Leadership Day was trying to bridge generational gaps, trying to see how we can use technology to also make it easier for maybe different generations and what we've got. Jay was so excited the other day because he was like, Will, I wanted to pick your brains up, we've got some new features. What do you think? Because I think we have some good convos about kind of how we can use tech to influence what we're doing. I want you to talk about some of the new features that we've got with e-Main and what you've got. And how does that have an impact with general people that come in here?

SPEAKER_01 1:17

Absolutely. And I'll start off with a little bit

Why Generative AI Matters Now

SPEAKER_01 1:20

about what I shared in my keynote. One of the key insights we've looked over over the past few years is generative AI and a lot of these technology features. I mean, it's like a rocket ship. You know, it's a sci-fos, isn't it? You know, I shared my personal story about how I use Chad GPT for uh my health goals. You know, I'm like tracking uh my sleep on my Apple Watch. I got my strength training goals and my food and you know, lab reports, you know, you upload it into Chad GPT and it'll say, you're not sleeping enough, you need to eat better. And like you can really accomplish some stuff. But if you look at how we do this in industry and professional corporations uh with maintenance reliability, we're not there yet. And so this was this was one of our insights uh kind of coming out of Accelerate last year. So we said that we're gonna figure out and we set the goal is by Accelerate 2026, we will have three new features launched. We ended up with four, which is really cool. Well, you know, about the about the Tolegate, then we might said we're gonna find three solid use cases, and so how we thought about this, Will, is we didn't just want technology for technology's sake, we wanted three practical applications. And so here's what we came up

Talk To Your Maintenance Data

SPEAKER_01 2:25

with. First one we came up with was talking to your data. This is the most obvious one, and there are other companies doing this, but we thought we'll tell you about our little twist on it. But we want to be able to actually ask for information in the system of record with emaint and get back actual like results. Okay, wow. So think about this. Uh, and this is the classic example. When I first tested this maybe six or eight months ago, I asked, How many forklifts do we have in all of our facilities? And we had a mock email account, it'll tell you what kind of batteries are we using. How often do the batteries need to be changed? And it's just spitting back information where before you had to go through and like triangulate reports, and you know, you had to be a primary user of the system that's working in e-main daily in some cases, or at least familiar with the reporting structure now. Like you may be the casual user, so we've expanded the people that can access the email database for information.

SPEAKER_00 3:15

Super cool. That is so handy because I think one thing that I found in my experience with CMMS, and we are maintained to a lot of optimization, and we kind of look for the gaps and see how they can work across lots of different CMMS systems. And we've experienced lots across the past, to be honest. And I think one thing about industry, one challenge is that if they've been using it for a long period of time, there's a lot of fear of changing and oh, we've done all of this work, how are we going to lose anything? Or you've got embedded purchasing systems that are very difficult to move away from different things and stuff like that. But what I do see a lot of is a lot of people struggle to use a system for basic things, and I think that feature there is very valuable in the sense that anyone can find information by voice, voicing it. Is it a voice recognition?

SPEAKER_01 3:55

We that's a separate feature. We'll get to that. We go, I'm jumping ahead of me. You're jumping ahead of me. We're gonna jump ahead. So let's just think of Chat GPT, yeah, except you're querying your system of record within EMA, you're querying the database.

SPEAKER_00 4:07

So you basically ask it, I want to see this, it will generate everything from it and it'll give you what you need.

SPEAKER_01 4:11

And so here's the important thing: really easy to set that up. What's really, really hard, harder than actually programming the AI, is thinking through all the implications. Okay, I see. And so we've done this in a very, very safe manner and a professional manner for our customers. And so I'll give you a few examples. The first time we we did this in the mock environment, you know, facility A can see all facility B's information and it gets mixed together, and you're like, okay, that's probably not great. Or uh maybe there's some proprietary information that you don't want the average person, you need some role to use cool stuff. Or maybe, you know, we had a write feature, which uh, you know, it's writing something to the database and it maybe it's going to the other plant. So we went through and we really were thoughtful and we partitioned this in a way where we think we nailed it in terms of uh having someone be able to consume information where they're supposed to and access what they're supposed to, and actually really think through this. So everyone should feel really good about this. It'll be in beta, still uh access for all e-mail customers and X5 now for the next month or two, and then it'll be uh fully in the production environment. But the second order effects, I'm gonna make a prediction in the industrial space that will be an increasing issue as people kind of for probably the next year or two, as people kind of work to think through all the implications.

SPEAKER_00 5:26

It's massive. Wow, yeah. I can I can see that being huge. And I think with software in in general, usually sometimes the smallest updates have the biggest impact. But you've got to really think your way through them problem-solving things. And how do you say, well, if I can solve a small problem, how's that able to solve a bigger problem across the whole organization? Which I really love that feature for that. That definitely, for me, if I was using that as a maintenance manager, I'd find lots of value within that as well. So that's that's amazing. So, what's the other feature today?

SPEAKER_01 5:53

So

Auto Generate SOPs From Manuals

SPEAKER_01 5:54

let's talk about our build feature. Yeah, it's an SOP generation. So this is really cool. If you think about how many OEM manuals there are across the world, you see them on the shelf, don't you? You walk in, there's a whole massive bookshelf of manuals. Yeah, I get it. And what ends up happening is one of our observations is let's say you have a motor or pump, you know, you'll upload the procedures you need into your CMMS, but then there's a procedure that's in the manual that you didn't, and then five years later you need it, you have to go back. So this feature is really cool because it allows you to do this in an automated fashion. You take manual, upload the PDF, it'll scan, or any type of maybe it's a non-OEM manual and it's just a general maintenance thing. It'll scan that PDF document and generate the SOPs that you need for work orders. Wow.

SPEAKER_00 6:40

And it's automatically. So you you put their manual in, it will feed it'll say we found these 30 different things in the manual. This is what we need to do.

SPEAKER_01 6:49

And these are the four different motors it applies to, and it'll automatically generate steps in there. And then the maintenance we do want a human in the loop because it is AI, so let's say it's 80 or 90 percent correct, and maybe in your facility you want to have things done a little bit different way to a higher standard or maybe extend an interval. All that's okay. The human is in the loop that's approving these, of course, yeah, and maintain control over your maintenance procedures, but talk about time savings.

SPEAKER_00 7:19

Wow, yeah, that's that's huge. That's huge. I mean, I think to be able to, and again, this is another problem in the industry, right? This is why I quite like these solutions because they are kind of solving frontline issues that we've got. When we do PM optimization as well, and we're looking at these things, we are looking through manuals and all the rest of it, but where is it going? How is it actually being generated within the system? And I think information and data, there's a lot of things that could probably do that, but how is centralizing the same system is the important part, really, isn't it? If you can manage that workflow and gradually start adding all of these manuals in, and also look what you're doing with your database, you're actually now forcing yourself to put all the information in one place rather than being on the shelf.

SPEAKER_01 7:55

You know, it's brilliant. Yeah, but not only PMs, if you have something that's not a PM and it's in the maintenance book, that's now in your procedure. Yeah, it's not.

SPEAKER_00 8:03

So let's say it's SOPs and just the way that you know just the way it's done.

SPEAKER_01 8:06

So you can get you can get everything in there. That's really, really powerful and really time-saving.

SPEAKER_00 8:11

Yeah, I agree. And I think that's where the industry needs to improve. I do feel like we don't have enough SOPs. There's not a lot of an alignment with say engineer A and engineer B doing the same jobs. How do you know and how can you measure that they're doing the same thing and it's consistent? Because that's what maintenance is, right? It's consistency and it's making sure that you can continue to do the same process no matter who it is, and you still get the same results, which can be difficult to do depending on who's doing this stuff, right?

SPEAKER_01 8:35

So let's go to our third feature. This is the feature you seem most excited about, you know. So I'm like firing up here. This

Voice Created Work Orders

SPEAKER_01 8:42

is the the talk feature. So in natural language, you can use your voice to generate a work order. And so let's say I'm walking through the facility, I got my mobile phone out, I see something that uh needs to happen. I can pull up my phone and just talk into it and be like, hey, I'm walking through the facility at point ABC. I notice that this pump uh it's making a noise, and uh this is what's happening. Whatever my natural language is, it will parse out what that is and put it into the exact fields of the work order.

unknown 9:10

Wow.

SPEAKER_00 9:10

So it will okay. Just let me just get this uh this straight. Say I'm work order requests. Yeah, work order. So I'm walking around site, I see a pump leaking. Yeah, like, right, okay, I named the asset. This is where the asset is, here's the location, here's where it's leaking. It will transcribe that and it will put it in where it is with that asset within within the work order request. Within the work order request. Oh, okay, I see amazing.

SPEAKER_01 9:34

And so so what it is is it's a shortcut. People, you know, instead of pump broken or pump making noise, or instead of someone seeing it, saying I need to report that, getting back and completely forgetting.

SPEAKER_00 9:45

Because let's let's be honest, this I'll be honest with you, it within maintenance, this happens all the time. I don't really think that engineers want to not put this stuff in the system. That that's of course they don't, but what happens is complacency kicks in. And I think this is where technology can help the general engineer. If it's easier, you've got to make it so easy for the guys to do something. So when they do do it, it becomes more habitual. But like most CMS systems that we're dealing with nowadays, I know some have smart apps and you can even put it in, but to even be able to voice note in it's just so much easier. And I think that is quite revolutionary in the sense that the easier you make it, the more information you feel.

SPEAKER_01 10:26

And the more information you get, the better your maintenance is more, the more data you have, the more records you have. So what the desired intention is, is lower the barriers of entry for someone creating a work order request and then increasing the fidelity of uh what that work order request actually has. So you have more information.

SPEAKER_00 10:43

But that's what I think it is, because I think when you write down what you see in your brain, essentially to what you write is very different. Do you know what I mean? And how you explain things because I think some people, when they write things, struggle to explain what they think, if that makes sense. Because if we're saying it, I feel like you're gonna get more information there. And if you do that across all of your technicians, and say if you've got five or six technicians across the site, and or even operators, you know, operators are the guys uh next to the machine, right? They're the ones hearing it, feeding or whatever. Giving that feature to them guys is gonna significantly improve the whole process.

SPEAKER_01 11:15

So that's what we're hoping. So we've been tested uh ahead of launch with about or uh about 12 or so customers, and I think we're gonna see a lot is what was the feedback? Was the feedback good? Is that what you kind of feedback is good? I think what we need to see is how is it used in the wild? Probably had two dozen people, you know, play around with this, but we can extrapolate this from 20 text to 2,000 text. Yeah, how does that scale up? How does that scale? What are we finding? Is are we finding maybe the microphone isn't capturing in a noisy environment or we need to adjust things? But look, we can do that. And so that's one of the things that's really cool about software, right? Is you can iterate and you can make directions, improve it, yeah. And we can find out, and we don't have to like bring things back in, it just we can just push it out.

SPEAKER_00 11:56

And I think that's what you know this is about sometimes. These features don't seem like, oh my god, this is like rockets, but it's the implementation and the ability to be able to make sure that because like systems can be built to work, right, but they're only as good as the users and what we do to be able to implement them. Do you know what I mean? So making it easier to use, I think generally is going to improve overall output of it all, and it's gonna allow clients to be able to get on top where they keep their central information, really. Do you know what I mean? Because that is the brain box of their whole operation, really, is such an important part. And um, many sites we see don't really utilize it to the full potential. So you're missing out there on efficiencies and effectiveness through the maintenance practices. So, yeah, I I love them personally. Like, I think I see a lot of value in them, and I'm hoping that you guys see your customers, and hopefully you get the feedback, and hopefully you get the value coming near it. Yeah, and it'll take this again, it takes a while for these things to happen and you know and go through, but there's definitely some long-term value there.

SPEAKER_01 12:52

You ready for our uh feature number four?

SPEAKER_00 12:54

Is this oh easy? I don't think I've heard this one, so this is gonna be doo-doo.

SPEAKER_01 12:57

Uh

Translate Manuals Across Global Sites

SPEAKER_01 12:58

this is the learn feature. This is a derivative of of uh the AI assist, which is the uh talk to your data feature. So when we were testing this, this is one we hadn't planned, but it came up in testing, and this was like quite catalyzing, it's the easiest of the four to do, and people were most excited about it. And so you don't you can't predict these things. And I was just, you know, I'll tell you, this is real. I was on my way walking to this podcast, and literally one of my customers stopped me and they're like, they were talking about I have a plan in Japan, Germany, and the US, and the translations between these are really hard. And we designed a tool where when you have document data repositories, you can ask for information and it will scan through the manual, search for that, parse out what you're looking for, and then translate it into the language that you need. Wow, that's incredible. And so this is such a simple thing with the AI technology, but such the most powerful thing. And okay, like it's 90% correct or 95% correct, but for most applications, it's gonna work, yeah. It's gonna work, and it's gonna give you uh enough insight to know what you need. And so what we're seeing is, and and you probably realize this with your customers, is we're seeing more global rollouts to be made across countries and cultures. We're seeing a desire to standardize and have more digital transformation. And one of the key insights that we have is this is hard with language or cultural barriers.

SPEAKER_00 14:17

Yeah, yeah.

SPEAKER_01 14:17

And so this feature is designed to help address and alleviate some of those concerns.

SPEAKER_00 14:22

I can see that being really valuable even for some, maybe some of our customers as well, because we're doing a lot more work in Europe now. And in Europe, you know, it's a big place. I mean, often compared to the US, but you've got lots of countries within Europe, and you can travel across borders and there's different language barriers between them. Does that make sense? Yeah. But what we're trying to do as well as a function in some of these customers is start to standardise and align a lot of the SOPs, the standards, the practice. We need that aligned right across all of the plants. But what is difficult sometimes is the language barriers. So the principles are the same in terms of what we're trying to teach them in terms of reliability practice, but how do we get around the language barriers and stuff like that? Which is you know, this here seems like a great solution to some of that, and you know, some of them things of how do we make sure we can communicate in the same practice of reliability but across different countries, which could I could imagine we're starting to find it a bit challenging because we're now like, oh, okay, they do speak English, but they need these manuals in German. Do you know what I mean? So we're like our normal content doesn't like oh, we have to you know change a few rounds and all the rest of it. So, you know, I get it, I get it. I I see the value in that.

SPEAKER_01 15:25

And just for you, Will, I can give you a preview of some of the things that we're thinking about. These are not new features or uh anything on our roadmap, but there's things that we're thinking about that are really cool and exciting. You interested? Come on, give it to me.

AI For Skills And Competency Mapping

SPEAKER_01 15:40

Let's let's so so think about this with uh work order history. Let's take your company and the reliability gang. You have how many employees now? Ten, ten guys work less. So let's flip this over. Let's say it was in a facility and you knew every work order every one of your people were working on. And so you can make assumptions saying, Oh, well, well, Crane ain't available today, and I need to have a vibration job, and it's really hard. It's a it's a uh technical job, yeah, yeah. It's a technical job, and like I have uh one of my other employees available. You could ask the AI potentially, hey, this person has done all these types of work orders. You could see every work order that employee's done. Do you think he could do this job?

SPEAKER_00 16:18

Okay, that's very okay, yeah, yeah.

SPEAKER_01 16:21

And it'll be like, hey, look, he probably doesn't have the right experience. You know, he may need to be able to, he may need some more. Oh, or absolutely, he's done all these, and we think he can do it based off this fact pattern. We've seen these people be successful. So trendy. So there's some competency things that we're looking at. And maybe, I don't know, I'm gonna say five years in the future, maybe a technician has a portable record where they now have all record of every job they've ever done. So now you have a different credential. I mean, the possibilities here are quite endless. Yeah, I can imagine. And super exciting. And I want to be clear this is this is something that we're thinking about. Yeah, we haven't quite figured out what the future looks like, but this is coming where we will know so much more information about asset types, maintenance professionals, and how the intersections are than we ever have in the past. You can ask questions of I got a vibration tech, how long will it take him to become a certified person in ultrasound based off of this experience? And you'll have a prediction of how fast he can work.

SPEAKER_00 17:18

That's pretty cool. That's pretty cool. What I love about that is as well, I think, again, businesses like ours is quite small. But if you were looking at big organizations that we deal with, competencies is a problem. Like you've got guys there that may be on the same level, but you've got to look at some of the jobs that are done in a plant. They can be so vast, you know, they're not they're not all the same thing, you know. When it comes down to what we're doing, vibration analysis, okay, there's a lot of things we do, but say, for example, you've got guys in the plant, and it's, you know, they need to be able to fix an electrical problem, and you've got electricians and you've got your electrical team, there's still abilities within that electrical team that you might have an engineer's not quite capable enough to do some of them jobs because of their experience, not just based on their qualification. So I think competencies and stuff like that, that that would be huge.

SPEAKER_01 18:00

That would so there's a lot of interest here. You know, they might say, Yeah, you could learn how to do that job in a week. Jay, man, you're gonna you know, I might take me a year. This is real, you know.

SPEAKER_00 18:10

But um And the idea and the premise is very good. And this I think this is why I like speaking with you, Jay, because you're always thinking about how to solve the problems. Do you know what I mean? And and I think as more of a problem-solving kind of thought process, it's almost like a philosophy, isn't it? You know, we've got these systems, we know that they exist, but what really splits the margin, what really makes the difference is how do we approach the problems and how do we actually say on a ground level, how can I help you? You know, how can I help the technician there that's are fed up with old CMMS or fed up but doesn't use it or doesn't believe in it? How do we inspire them guys to get on board? And a lot of these things are are designed for that kind of front force workforce to work more effectively, and that's where the buyer needs to be. You know, these guys need to use it because if they don't, the whole system falls on itself.

SPEAKER_01 18:57

We

Bridging Generations Through Usable Tech

SPEAKER_01 18:58

talked a little bit about uh before the podcast started about the generational gap. I mean, like young people want to use technology, you know, they don't want to use like uh something that's with like complex transactional codes. They want to use something that that's easy. You know, I joke, you saw me tag you in a post about my 14-year-old daughter. You know, we're trying to get her get her excited to be a part of the reliability game. Oh, it's like brilliant great choice. Her usage of technology, it's like fantastic. She uses, she's learning in school the same program eMate uses uh to do their wireframes.

SPEAKER_00 19:30

It's crazy, right? And like you think of the this generation and and how they're using tech now, you think of what they can be doing in the next 10 years, 15 years.

SPEAKER_01 19:38

She will have an expectation if she's working in this industry that those tools are available.

SPEAKER_00 19:43

Yeah, it is just a norm for her.

SPEAKER_01 19:45

And so, like, if it doesn't have it, it's gonna be like, man, that stinks. So we gotta be really on the forefront of trying to do this and trying to make it uh I I you know, we're only as good as customer adoption. Yeah. And so really being thoughtful, and I think these features kind of help us along that adoption curve.

SPEAKER_00 20:00

Yeah, and it kind of bridges the gap as well between the older generations because, you know, they have all the knowledge, right? And they've got all this experience, and we've got this new generation coming in that will adopt tech quite easy. And if we can use some of the things that you're doing here to almost bridge that gap a little bit, you know, you've got collaboration then, you've got the knowledge that it's potentially going to be lost very soon. Going into, you know, younger generation as well that are able to use the technology, and if you can marry the two together, I think that is a recipe for success. That's that leads me to my last question, Jay. Okay. Speaker, great podcast. This is kind of a bit of a closing thought. We know technology is moving fast, it's it's rapid. You know, you spoke about your daughter using all of these codes and programs. You've got kids programming code now and doing all of this crazy stuff. Technology's moving, and I don't think we can stop it. It is kind of on its own train. But the culture part of what we're trying to do and align with is very slow, and trying to get people to use some of these systems and know the whys is can be a little bit elongated. How do we speed up the process of really getting people to understand and align that technology gap to the culture gap that we're finding? Really difficult with some customers, some customers better than some. But you know, you've got customers wanting to put AI and everything, they haven't even really sort of their CMMS outwards or and all these things.

Closing Thoughts On Culture Change

SPEAKER_01 21:10

So there's a couple things. This is a leadership question. Yeah. And there's a few things here I'm really hopeful on. One is information becomes more centralized, and IT systems like eMate become more deployed at senior levels in the organization. Like digital transformation officer, that title didn't really exist 15 years ago. Now that's a primary stakeholder here. So the conversations have become elevated, the awareness becomes elevated, the data becomes more consolidated and visible, the impact of not doing this becomes more visible. It gets more senior leader attention that drives a culture change. And if you look at this, if you look at the attendees just at our conference, you know, go back 15 years ago. Today we have directors and VPs of maintenance that attend, like that was not as common. That's why we started Thought Leadership Day is to have programming for both users of the system, but also directors and VPs that administer the system as well and think about what they're doing in their organization and decisions. Yeah, that's a fantastic trend that plays in this favor. I think also as organizations become more lean, you know, the leadership and the relationships become more important. Employees have a choice of where they work. Yeah, yeah, you're right. Yeah, and they're gonna want to choose somewhere and they're gonna want to choose, right? Yeah, they're gonna want to choose to be in a good culture, and so this will force companies to adapt. And so I think those two trends play very much in favor of some of those cultural shifts. No, that's a great idea. And having people like you in the industry also, that's a little bit exciting, starting new companies. I'm very hopeful for the future.

SPEAKER_00 22:44

Yeah, so am I. And you know what? I love that point that you've raised in terms of you know, is the company got a good culture? Because you need to attract the talent as well, right? You can't have a culture that doesn't look for the future, doesn't look for the visions, because people are not going to align with that now, and especially with the new generation, they want to see progress, so they want to see that that actually adding value to the progress. And if you're not adopting that culture, you have to be forced to. So, yeah, that's a really good point.

SPEAKER_01 23:06

But I just want to thank you for it's a pleasure as always.

SPEAKER_00 23:09

Honestly.

SPEAKER_01 23:09

Uh, it's been, I guess, I last time I saw you was at Main Tech, but I always like catch it up, but I always like seeing what's going on with the reliability gang.

SPEAKER_00 23:16

Well, again, always a pleasure, Jay. And you always bring a real good energy, a good vibe. And I just could interview you all day if we could. Like, I don't got enough time to have to keep going. But honestly, thank you so much for your time, and I really appreciate you coming on the Reliability Gang podcast. All right. Thank you. Take care, Jay.