Tim Whelan with Datch

SUMMARY KEYWORDS

data, ai, asset management, asset, tim, enter, frontline worker, industrial, palo alto networks, maintenance, solution, enablement, work, prompt, piece, ultimately, system, worker, focused, cmms

00:00

Industrial Talk is brought to you by Palo Alto Networks. You've heard me talk about my friends at Palo Alto Networks zero trust OT security delivering a comprehensive security solution for all OT assets, networks and remote operations. But did you know that the Palo Alto Networks solution provides over 1100 app IDs for OT protocols, over 500 profiles for critical OT assets, and over 650 OT specific threat signatures, now that's best in class security delivered, learn more about Palo Alto Networks, zero trust IoT security solutions, and how you can achieve 351% ROI over five years. Learn more about the Palo Alto Networks solution by going out to Palo Alto Networks.com. That's Palo Alto Networks.com.

00:57

Welcome to the Industrial Talk podcast with Scott MacKenzie. Scott is a passionate industry professional dedicated to transferring cutting edge industry focused innovations and trends while highlighting the men and women who keep the world moving. So put on your hard hat, grab your work boots, and let's go

01:14

Alright, once again, welcome to Industrial Talk a platform that is dedicated to you industrial professionals and companies that get it done. Because you're bold, brave, you dare greatly you innovate to collaborate, you solve problems. And therefore you're making Yes, you are making the world a better place we are broadcasting on site SMRP 31 is the place in Orlando, Florida. And that is something that if you would have any passion, any desire, any need to get engaged, go to SMRP.org start there, that is your first step right off the bat. That's SMRP.org They have all the resources for you to be a successful asset management, reliability, maintenance professional, big time, as an rp.org. Tim is in the house. We're gonna be talking AI, so let's get cracking. All right, Tim, don't let us down. Bring the energy. Do you do whatever, you know, AI? AI is definitely the the, the word of the day for anything

02:20

very, very buzzy, for sure.

02:23

To understand, but

02:25

I think probably what we hear often is, I think people are getting a little bit more hip to the buzzword. And they're actually asking the question like, well, how is this AI? What is it doing? Yeah, do for us? Yeah. And, you know, I think it is a big challenge. Since so many other companies are touting AI for us to really demonstrate how

02:47

do I trust it? Right? How do I know? Yeah, I mean, yeah, it's cool. I can I can upload a quick statement and give me a summary in sort of, I take it for what it's worth, and I run with the summary. But but how do I know that? That one AI is better than the other? I don't know how you do that? Yeah.

03:08

I mean, it's definitely an issue that each individual will have to work through and make sure that they're doing their research and their due diligence. But I think, you know, like, it's an we're an important phase where we're moving away from like, just saying buzzy AI stuff. And then now trying to actually relay to companies and, you know, individuals like, this is a tangible piece of AI. And this is how it can enhance what you're doing. So trying to actually prove out and show the benefit of where AI can help someone, particularly frontline workers, and you know, people in maintenance and reliability and asset management. You know, there's some really tangible use cases where AI is very beneficial. Before

03:58

we get into those. I broke the cardinal rule of interviewing, I don't even know who Tim is so nor to, if I don't, the listeners don't know, give us background.

04:08

Yeah, absolutely. My name is Tim Whelan. I'm the head of sales for Datch. That's da t ch Datch.io. And, you know, we are a an AI powered enterprise mobility solution. So sorry. And then my background is I've been in SAS sales for over 10 years in industrial adjacent markets. So formerly selling like EHS solutions, as well as IoT platforms and data analytic platform, data analytic platform, excuse me, and now currently focusing on like data acquisition and frontline worker enablement. So yeah, Datura is an enterprise mobility solution. We're powered by AI. We leverage natural language processing. So frontline workers can actually speak to mobile devices as they're completing tasks. And we'll collect that data, we can enrich it, and then we can route it into any back office systems that it needs to end up in. And we can prompt any actions that need to happen from there.

05:10

All right, so now everybody knows that Tim has mad skills. And and I don't. So that's it's just the way it is. It's always an OK, let's talk a little bit about AI. We briefly mentioned it, but how is ai ai transforming the asset management world?

05:27

Well, I'd say that AI has the potential to transform the asset management world. Right now we're looking at this sort of adoption phase where, you know, large companies are, you know, looking into AI, looking into their kind of more of their stance on AI and how they can incorporate it, incorporate it into their stacks, I'd say that it has the potential to enhance what they're doing by, you know, in our case, we can collect cleaner, more contextualized data. And what that will allow is for asset management teams, and reliability teams to make better asset lifecycle decisions. Yeah, so that's the main piece, but ultimately, to we're trying to reduce unplanned downtime, right, get it down to the minimum. So if we can help them understand, you know, their assets a little bit better, and maybe predict when things are going to fail a little bit better than, you know, obviously, then we can, we can keep that uptime at the most positive.

06:34

Give us some use cases, in the world of asset management, where, where this will be. I mean, it's, we all have a similar conversation, and the conversation is I'm collecting data, I run that data through some sort of analytic and, you know, parameters of some sort. And then if it goes outside those parameters, and that's bad, and, you know, fires off something, how does? How does AI? How will AI differ from that?

07:05

Yeah, I mean, that's a really good use case. But I'd say that the, the core tenant of that is clean data. And that's what we're trying to solve. So right now. You know, think about industry for Dotto, also pretty, pretty buzzy, and it's very broad. And there's a lot of difference.

07:25

It's a miscellaneous file. Everything is exactly the industry.

07:30

But one of the key pieces, right, one of the building blocks to anything that you want to do within industry. cortado is data like that, you they need more data, and now more than ever. But, yeah, so in order to accomplish those goals of where you're feeding it into some algorithm, and then it spits something out, and it says, Hey, this is the most optimal settings or this asset is going to fail, then or these are the things that you need to look out for, you know, you need, you need good data. And right now, there's there's plenty of different systems and tools that can grab data from machines, right? So if you're thinking about IoT sensors, or grabbing data often PLCs that that data is traditionally very trustworthy. Now it's the human entered data. That is historically unreliable and finger, right. But you know, just when you think about it, let's say that you've got 500 field workers, right, right there, and you're relying on them to enter data back into the system. Right there. Before anything else, you have 500 different variables of how data is going to be entered, right? So there's no standardization when you're looking at who's entering data, how they're entering data. So you know, there's some some great use cases for AI would be worker enablement, plus data enrichment, right? So the worker worker enablement piece is like, how do I offer better technology and better experiences for the worker, so they can do their job a little bit more effectively and efficiently? So the way that we will do that, right is we offer mobility, right? So instead of, you know, having to go back to a laptop, or you know, a kiosk or something like that, to close out work orders, for example, a maintenance technician can use their phone application, and they can speak data back into their phone, right? So this is conversational AI, what they'll do is what we'll do is interpret what they actually mean, not just take exactly what they say and dump it into the field, but will actually interpret what they mean. So you know, there's like, oh, we

09:43

gotta we gotta step on that one for a minute. Are you saying that I can. I just performed maintenance on this motor, and then instead of the fat finger in that thing again, all I have to do is say, Yeah, I changed the bushings and this that and yeah, The thing and that that AI will step in and, and be able to summarize that in a more cogent way than me.

10:07

I mean, it will. So our language model will be able to take what you said clean it up, right? So if you are like me and you throw in a lot of bums and ahhs and you need a little bit of time to sort of bring your thoughts together, and you fill in the space with filler words, yeah, can eliminate that. But also, if you say things like, you know, completed work on an extruder, this morning, and I replaced, I'm not this morning, this afternoon at four o'clock, it can completely erase that, I mean, this morning and just put in, I completed this work at four o'clock, right. But I think the more powerful piece of of like the NLP and like, what it's understanding is, when we're integrating with another system, take like SAP plant maintenance, for example. You know, we can enter, we know what fields need to be entered, data needs to be entered into when they're closing out a work order. For example, if someone enters a block of unstructured text, and then just say something like, Hey, you know, went to extruder, number one, replace the screw took three hours use part ABC, we can take that data and structure it back into those fields just from that like paragraph of text. So as a frontline worker is, you know, leaving one job and moving over to its next ticket, they can close out that work order. But what that will allow them to do again, since they have mobility, and they can enter this in real time, it allows them to enter more of that contextualized data that they might forget otherwise, if the internet later, right, so if they say, Hey, I replaced the screw, but guess what I you know, I noticed that the flywheel was wearing right there was metal shavings in the box, maybe that's something they don't remember, if they're entering it in at the end of the day, and maybe three weeks from now, flywheel blows that assets down for three days. So giving them tools to do their job a little bit better will help will help the enterprise ultimately get that more contextualized data. You know, and then obviously, things like real time translation and transcript transcription, we have plenty of frontline workers that are not, you know, that speak a native language other than English, or, you know, it doesn't matter what country you're in, but, you know, we have the capability to translate and transcribe in real time. So again, these are sort of that worker, these worker enablement, tools that allow them to do their job more efficiently and better. And then we also have the data enrichment side, right. So this could be prompting questions based off of data that's entered. So if you know, they say, Hey, you know, again, I'm using the same terms over and over again, had to replace the screw. When I opened the box, it was black, right? And it's like, we can prompt questions from there, we can say things like, was there metal shavings in the box? Was it making a weird noise before it went down? Was it whining was it screeching, right. And these are things that our model is picking up and prompting questions based off of that. And the goal is to get better data back to the planners and the schedulers. So they can do a better job of preparing the maintenance worker for, you know, for whatever the task is going to be. We all know those stories where a scheduler or planner, you know, puts one work order out maintenance person goes to the asset, they open it up, and they're like, this is not the work that needs to be done. And they have to go back to the office, yeah, get a new plan, get new tools, and it's just extremely inefficient. So we can work on that. But again, ultimately, like we mentioned, we want to keep uptime, as optimized as possible. So, so yeah, if we can, if we can get better data back to them to hopefully preventing, stealing, predict and prevent, when assets might have issues or when parts need to be replaced, then ultimately, the entire enterprise is going to run a lot more efficiently.

13:49

So So with that said, it. So I'm taking the work order, I'm updating the work order I, I've analyzed this particular piece of equipment, and then I I verbally enter in the information, which I think is just brilliant, honestly. And I, I'll be the I'll be the poster child for not wanting to enter data. Sure, you know, and or pencil whip in it, or whatever it is, yeah, so I'm able to sort of verbally say, Hey, this is this, this is this and have a have an AI platform that will sort of clean it up and make it succinct so that planners and others can understand what's actually happening or what happened at that asset. Is is just, that's a great use case. Sure that I that I really, really liked. Do I have an opportunity to to spit out my my analysis and then correct it? Yeah, yeah, absolutely.

14:46

You know, it doesn't have to be it's not like an automatic fire back into sh SAP like you can you can check the work for you.

14:53

Eventually it's going to be because I know that I'm not going to want to spend time trying to correct my I don't know it just seems like it would be a great thing. Yeah.

15:02

I mean, so again, like the NLP, the natural language processing is very good. Like, I mean, it's, it's very, very good. I mean, we're talking about, you know, upwards of close to 100% accuracy in multiple different languages 60 different languages actually. So, yeah, the technology is very good, of course, you know, people like the opportunity to read their notes, check their work, and maybe add some more contextualization, right, and we allow that opportunity. So it doesn't publish automatically unless you hit publish, but you know, that's more four, I think there's going to be common adoption curve, where it's like, people need to get comfortable with really new technology, that it's actually going to work the way that they expect it to, once they get comfortable with that technology, because they've had experience with it working correctly, then it will become more of a just, I know, this is working, I don't even need to look at my screen, I can just speak to my phone after saying, Hey, Dad, and then keep walking, doing maybe I'm working on an asset, even while I'm doing this stuff, and I can rely that it's going to go back.

16:06

So how do you? How do you deploy it? How do you come in? I'm an I'm an existing organization, I have a, of course, I have my CMMS and has the historical record of I'm not asking about how clean the data is associated with that asset, but I have data is associated with the asset, how do you how do you incorporate this AI technology into my existing? You know,

16:33

yeah, so you know, I think there's baby steps to give a broad picture here, too, like what Dad wants to be, ultimately, is a single interface of front end piece of software that someone interacts with, right? For every back office system that an enterprise works with. So we're talking about plugging in your EIN, you know, or CMMS, or, you know, your could be your hrs software, right. So you can imagine being at a maintenance person is at a, an asset, and they're, they're entering data, or they're speaking data back into their phone, they're saying, hey, you know, just completed work order number 123. By the way, it's my wife's birthday tomorrow, and I want to request off, right, taking that unstructured data, routing into two different systems, one back into the am the other time off requests into the HRIS system for a supervisor to approve or deny time off, right? That's ultimately what we want to be is like that, that front end tool, like the only front end tool that they use, but sorry, so nowadays, for the deployment, we usually focus on something like, you know, in em, we have a turnkey integration into SAP plan payments, specifically with us for HANA. So since we have that turnkey integration, and we have this data for asset management solution that is ready out of the box. So it's, you know, anything that a maintenance person would need, they can, you know, doesn't need to be configured, excuse me customized in any sort of way. But we complete the implementation over, you know, a week to two, which will be plugged into that back office system. And then, you know, just simple as downloading an app on a on a mobile phone for, for all the users, so we can deploy it within a month. So option is, is very easy, because it's sort of like using Facebook, or Instagram or anything that they're used to using on phone.

18:30

So it really is a phone interface. Yeah, absolutely. So I can take it and I can say, get on the app, speak into it. And then magically, it goes into the areas that's where you want to go, I get it. That's the use case, that's where you want to go. But to do it in such a way that that it makes it so efficient. And you don't have to, you're trying to eliminate the typing.

18:54

Yeah, there's a ton of friction when it comes to data entry. Right. And that's why there's been so many different technologies that focused on eliminating humans from from data entry, right again. So if we're thinking about IoT, excuse me IoT initiatives, and, you know, just data pipelines that are pulling directly from machines, the goal was to eliminate humans to make sure that that data is accurate. The problem is, when we're dealing with asset management, it's impossible to eliminate the human right, like these are the people that are completing the work. And and so to get the data back in as to what work was done, or what work needs to be done, you have to have humans, I think, unfortunately, as an industry, we sort of accepted that there's no good way to make that aspect any better to get cleaner data. And we think that's lazy. And we know that there's a way to actually make it better. And it's by offering enablement tools to the frontline worker that gives them a better experience and allows them to enter data in real time. And then again, those enrichment tools where it's like, I'm going to take the data I'm going to interpret what you're saying and I'm going to prompt questions to enrich that data and make But the quality of it much higher when it's going back into the systems of record.

20:06

I think it's bad. It's bad technology. I can't, I can't complain about it. I really think that's cool stuff is still at the tip of the iceberg. When it comes for sure, how do they get ahold of you, Tim?

20:19

So you can find me my email address is Tim at dat da tch.com. Or you can find me on LinkedIn. It's Tim Whalen. That's w h e l a n and yet be able to connect with anyone I love having conversations like this and I love sort of the debate of how technology can can help within this industrial center. Amazing.

20:45

All right, listen, there's we're gonna have all the contact information for Tim out on Industrial Talk. So if you're not following his contact information will be there reach out that is your number one thing that you need to do, because he's got mad skills, we are broadcasting from SMRP 31 COMM to SMRP 32. When you're I mean, you just got to make that a priority. Go out to SM rp.org Become a maintenance professional reliability professional and asset management professional that is your first stop S M. r p.org. We're gonna have another great conversation shortly. So stay tuned, we will be right back.

21:18

You're listening to the Industrial Talk Podcast Network.

21:28

Incredible stack card for Tim. Tim I'm looking at it right now LinkedIn, go out to Tim Whelan. Contact information will be out on Industrial Talk. Datch is the company d a t c h.io. Right. And that is the Marion of AI solutions. It's happening with asset management, and how to improve that whole profession. It's an exciting time. Don't get me wrong. You need to reach out with Tim and you need to find that trusted individual that you can ask questions and be able to see how AI can help your business. It's happening. It just is find those trusted resources. Industrial Talk is here for you industrial professionals to amplify your voice. Be a part of this ecosystem, this expanded ecosystem and be a part of that community that really solves problems. We've got a lot going on out there. Go out to Industrial Talk Be bold, be brave dare greatly. Absolutely hanging out with him. Change the world.