AI quality control powers productivity.

good well excuse me let me make sure we are broadcasting live here we are I just wanted to check my dials and we're not connected to Facebook but it will find its way through those troubleshooting efforts but we are broadcasting live to LinkedIn and to other podcast channels X and YouTube and many others so thank you for joining today's in-service podcast my name is John Carroll I'm the CEO and founder of the service council welcome to today's segment of the in-service podcast series a topic and a title that I think is really resonant with our audience AI quality controls powers excuse me AI quality controls powers quality excuse me so really excited about this this topic really excited about the discussion today and I'm really pleased to be joined by one of our newest board members his name is Thomas I'm gonna try it to the a we were pronouncing that and practicing and planning got it all right wonderful a warm welcome to you Thomas thank you so much for joining us from Deepomatic thank you very much thank you good morning very excited to be here outstanding so before we jump into our Q&A and an introduction to Thomas just some grounds keeping here we want to keep today's session interactive we want to allow for submission of Q&A and for you that are listening for those of you that are listening on LinkedIn we encourage you to comment we encourage you to comment and we encourage you to comment and we'll try and build the the comment the reactions the questions into today's discussion Thomas has agreed to think on his toes and respond to any questions that are submitted so without further ado let's go ahead and jump into today's discussion so really interesting one the concept of quality and automation and and how do these two things come together right and and so I want to first and foremost welcome back Thomas to the discussion and Thomas you know I know you very well you are a recent appointee to our technology advisory board thank you for all that you're doing in that capacity but for our listenership why don't we start with an introduction to you personally and then we'll get into the company yeah awesome with pleasure and good job on pronunciation by the way I know it's not it's not easy but very happy to be you know participating to this conversation I'm listening to them consistently so always a pleasure to have you know those being produced you know those being produced you know as my last name can attest born and raised in France my background did not necessarily destined me to a field service or even working at a tech startup I studied political science got a master's degree in corporate finance and market finance but I got a great opportunity during my study years to come to the US study university at the university of Georgia go dogs yes and there I met a very fantastic teacher his name is Bob Pinckney he's been a great entrepreneur all his life introduced me to entrepreneurship and it's also thanks to him that I got a first foray into field service I actually worked for a couple years at a coffee distribution company which was fully integrated including the distribution of kegerators of coffee coffee to offices and that meant having a delivery team installation team maintenance team and when we expanded from the tri-state area of Washington DC to New York metro I joined that team and helped support the logistics of it and you know that's really where you start understanding the importance of well the good quality of the installation of those services and a lot of what we're going to talk about today is going to relate to this and as much as much as our operation was not necessarily as large as some of our customers at diplomatic we really got to see firsthand how this impacts our day-to-day operation revenue costs etc when you don't deliver proper field service quality but after that got the opportunity to join diplomatic that's been now over five years now five years and a half almost got a couple of different roles but now I'm in charge of the North American market including Canada Canada where you know I travel around Canada where you know I travel around the country travel to Canada meeting different customers so it's very exciting and I'm very happy and grateful to be in this country. Well that's a great journey and in college kegerators meant a different thing it didn't mean cold brew of coffee. It meant something different. It did not and if you can this to go more precise it was nitro cold brew which means it looks very very much like a Guinness. For yourself a glass of nitro cold brew you're drinking Guinness in the office during the day. Outstanding. I'm Irish and we just celebrated St. Patrick's Day so you just brought back fond memories. Well wonderful. Let's go ahead and so thank you for the introduction to yourself. You know our listenership some of which are familiar with diplomatic some of which aren't. Could you introduce the organization? Talk a little bit about the journey that diplomatic has been on. Of course. So diplomatic if really we want to have an accurate picture I have to say it's it's a startup now that has been around for over 10 years so we've been working on the same technology for quite a while and I like to think of us as really expert in our field which is computer vision. If you're joining this session you might be more or less familiar to what AI is and computer vision but what I'll say now is is that in short computer vision is really just a subset of AI. It means we can make and deploy systems that can automatically recognize visual patterns in images and video. And from the very start what diplomatic was geared to do was enabling large companies to conceive and deploy but even more importantly maintain those systems in operation. For processes that are very much more often for processes that are very much linked to what we'll call the real world. So it's everything that's really outside of let's say a factory setting maybe everything where you're working in a messy environment a lot of unpredictability a lot of changes over time which you may heard is called the data drift right. So that's what really D-Punix has been doing for 10 years. We've worked with a whole range of different industries still have customers today from the pharma industry the pharma industry the pharma industry the pharma industry the automotive industry food service waste management those are happy customers of DeepPamatic but in the past three four years we've really honed on to one very specific aspect of what we do which is the ability to improve field service operation and also very specifically the deployment of telecommunication networks which is something that DeepPamatic has invested a lot of focus. And so today I'll conclude on this and so today I'll conclude on this what you'll see us talk about the most and what most of our customers do is use DeepPamatic to deploy systems that will live directly in the hands of any field techs out there doing work and for whom you might from whom you might you know require some reporting to make sure that the job they did was done properly to make sure that you know it's been done and you know where it's been done and so where DeepPamatic is involved is when the techs document their job through photos DeepPamatic will enable our customers to automatically validate the quality of the reporting automatically validate the quality of the job being done and interacting with that technician. Outstanding. It is a growing your brand is growing your tribe is growing I see a lot of new customers being added DNS comment and, as you know it's growing your prospects on a number of new customers being added, you know and then similarly expect to be a april 200-2008开始 soon he is the volt leb and so next week we go photos Petersburg and 35 we are still earning the feedback and saying you know listening. We're going to share some of the data today as we talk about some of the questions we had planned. But just flag your calendar next month, April. We'll be recasting that survey and potentially expanding it to look at not just the frontline agent that is in field, but also the contact center agent and doing a voice of the contact center agent. So there's an expansion of that series. So stay tuned for that. So we're going to break down today's topic into its subcomponents, right? In terms of the titling of today's session, AI quality control powers productivity. I fumbled that title as we kick things off today. So hopefully we can get through this a little bit better. But I want to break down the title because I think it's really important that we start with some definition that we actually hear from you, your opinion on the term quality. You know, what does quality mean to you? What are the core tenants of quality? Especially with respect to a service business. Yeah. So what I'll say is like a lot of things, it depends. And I think what we understand as quality will vary from one business to the other. But when we specifically talk about services, I think there are a couple of ideas and concepts that can be seen across the board. And I think that number one is probably when you do a service in the field, what you want is it to be achieved probably in the first visit, right? And that's a concept that we really like to talk about at Deep Emetic is that first time write rate. We talk about first time write automation a lot. And so when you send someone to perform a task, you don't want to have to come back to do it again. And so that's the most obvious one, because obviously, if you have to go again, it means you failed in that installation. It might not be because of the technician. It might not just be because of the environment. But obviously, that is that is going to be a lack of quality. But then you go into more specific aspect of quality, which might be the sustainability of what you've done through that job over time, right? You might do that, provide that service, it might be acceptable on the on for at first glance, right? Where you delivered your service, but maybe some aspect of it was not done. because we do be specialized Shiny It's a network. What happens is if you're building assets that are dispersed a little bit everywhere and those assets have small mistakes, small errors that might not prevent the service from being delivered. delivered, we see what we would call the creation of a technical debt, which over time adds up. And what I mean by that is if you do have an issue in the future and you need to work on certain piece of equipment you've deployed, certain pieces of the asset, but if you've not documented that asset properly, if you've not labeled that asset properly, if it's not being installed as per the specification and what the person would expect when fixing that problem, they might not be able to do it. That will be a problem. Or maybe the cut of service, maybe the problems will arise from this technical debt. So that is another aspect of the quality. And finally, what I'll say is it might also just impact the quality of the service being delivered. The service might be delivered, but maybe not as well as it should be. And that will cause problems related to well customer experience, your retention of those customers. And obviously this is a major focus. Absolutely. And I think, you know, one of the things that we hear a lot about is, especially in manufacturing industries, the expansion of the service network. Who is impacting the product experience and how do you support and create quality consistency across that network, especially when you think about expanding networks that include franchisees and dealers and independents and maybe even your competitors in a multi-vendor services offering supporting your products that are out in the field, right? So the complexity starts to unravel quality, right? So there's a lot of focus and emphasis on establishing and maintaining of quality across that extended service network, something that we're continuously hearing about from our community. Let's build on this discussion if we could, Thomas. Let's go and frame for our listenership, you know, anything less than quality. What is the impact of the organization? What is the cost of that? Right. Well, to answer this, I think we can go from the very specific, which I'll say is what is maybe more easily measurable to maybe a little bit more of a general view. And the most obvious one, and it's a little bit similar to my previous answer, it's just going to be the necessity to send someone back to a job that's not been done properly. So a repeat visit is going to be a very glaring cost, obvious impact on your productivity that we want to avoid. So that's going to be number one. The number two is going to depend for me on what industry you're in. But what I'm seeing a lot working in the telecommunication industry, utilities industries are going to be a loss of revenue when the consumption of your service is just not possible. I mentioned I work for a coffee distribution company. If someone installed the kegerator not properly, we're charging them for the amount of coffee they consume. Or if they don't consume coffee because they can't, it's missed revenue. So that's going to be another obvious impact in your operation. Yeah. Sorry. No, please. I was reacting. Please continue. But if you're not necessarily at a failure in terms of the service that was provided, but maybe the quality is not right there, that will impact cost in terms of customer service calls. For example, if you deliver an internet connection, that's not very good. And that might be just simply because the connection was not done properly. And that will slow down your connection. You'll get a lot of customer service calls that will be costly. It will have other impacts down the line in terms of churn. And more importantly, if that intervention was not done properly, where the connection doesn't deliver the right results, it might also be that you don't really know what causes it to, which means it's going to be very difficult for you to solve that problem. And now we can move to the things that are less and less easy to measure. But once you have an operation that relies on a pool of work, a pool of workers that's being less and less easy to tap into, and I think we can talk about this subject a little bit more down the line, but that means that anytime you have to use resources like a field crew or just one technician to go and repair something that was not done properly, that's a technician that might not be delivering a service to a customer who wants it, right? So if you're in an industry where there is a high demand growth in terms of what people expect from you, which is the case almost across old field service industries today, it means you really delay your time to revenue. You're using revenue, you're using resources, sorry, once again, to repair things that should have been done properly the first time, instead of connecting new subscribers, serving new customer, providing additional service, which will increase your top line. So those are going to be very important. And finally, the last thing I'll say is I talked a little bit about the concept of the technical debt. That's something that's extremely difficult to measure. But companies that deal especially with distributed assets understand that concept very well, because especially when they're physical assets, there is degradation, there are going to be problems, there are going to be issues that are outside of your control. But if there's a lot of little things, that means your asset has weak points in certain areas, a lack of knowledge, we talked about what happens in the back office as well, a lack of knowledge of what is where, how was it installed, how was it installed, how was it installed, that will have a very big impact in your capacity to address those problems, and those will be costly. And finally, if you do want to, you know, price your asset as simply as that, be it your customers, be it your physical infrastructure, if you're not able to accurately describe its quality, that will also impact you. Absolutely. Absolutely. And, you know, my team just alerted me that, Thomas, we're competing against the opening game in the NCAA college basketball tournament. I just alerted to this. So hopefully, we've got some folks that are, you know, bifurcating their focus on both the first opening game and this important podcast. Yeah, I am myself not a huge basketball fan. I didn't grow up with basketball. But a very good friend of mine has his brother playing one of the team that are playing today. Wonderful. I went to my first basketball game recently where they won. So I also made the sacrifice of coming here and not watch the game. Outstanding. Outstanding. Well, here we are. Let's continue. This is an important discussion. So let's continue. And one of the things that I want to look at is, you know, we conduct a service counsel conducts an annual survey called the KPIs and metrics benchmark survey. And this happens every year, typically in Q4. And in the survey, we learn the gap between best in class and average organizations. There are a couple of common characteristics, one of which was best in class had executive management oversight to setting and managing KPIs and metrics. The second was that best in class measured more metrics, 27 versus 18. So you know, that's a 50% 150% greater in terms of the metrics that they look at. The third was that best in class have achieved dynamic management of KPIs, meaning when we looked at and isolated the average performance in terms of key metrics, they were had a quarterly cadence and a monthly cadence of reviewing KPIs and metrics to understand how could we improve? How could we change against the things that we're the things that we're learning from that measurement? Whereas best in class, we're dynamically real time measuring it and reacting based on the ebbs and flows of their performance? And, you know, if we isolate, you know, this overarching trend that shot out from the data, there's this shift in mindset from achievement or outcome based metrics to supplement those with the metrics that are causal in nature, the ones that are causing the erosion. How do you see approaches to metrics? How do you see approaches to metrics evolving? What are you hearing from your customers and partners out there in the world? Yeah, I think this is a very, it's a great study and a good question for our topic because when we're talking about KPIs at the crossroad between quality and AI, it really is a question of, well, how do you make sure something is good? And, and, and, and really we're right there. And, and, and I just want to, to maybe tell a little story here in relation to this, because we have a customer, one of our first customer in the field service space, which is a large internet service provider in, in France. And they were very much evolving the way they're thinking about their network, how they build it, how they operate it in a way where they wanted to understand, more upstream, what results in a successful connection, a retained customer. And so what they have started to do was requiring a lot more reporting for, from their field technicians, from their vendors, from their subcontractors, et cetera. Collecting a very large amount of data to be able to conceive of better quality metrics in order to impact the end goal, which is, well, like we said, a successful connection and a successful connection and a subscriber that stays over the long, the long, a long period of time. And so this evolution meant that they needed a way to actually achieve this. And that's where AI and computer vision really is very powerful because an ISP in a company like France, like our customer means they were connecting several millions of customers every year. Now, if you want to make sure a connection is done properly, that you might start requiring five, 10 photos, you're suddenly taking tens of millions of photos every year. So you're starting to, you know what you want to do. You want to have almost an infinite pair of eyes with every technician that's out there to be able to look at what they're doing and start to understand who's good, who's bad. So they're starting to do this. And what's happening is they are seeing certain patterns emerge. And they're starting to think, all right, out of my vendors, four or five vendors across the country, I'm seeing one that looks like they're performing not as good as the other, right? And that's when they thought, well, let's see if we can have a better system to look at this data and actually have an accurate overview of what's happening, something we can digest and actually present to higher level executives. And that's where DeepMatic comes in. And suddenly, instead of looking at 8%, 10% of those tens of millions of photos that come in every year with a system that's just scalable, like DeepMatic, you can look at everything. You look at everything. And suddenly, it's not a question of sampling and just a fuzzy vision of what's actually happening, KPIs that are not precise. It's certainly, I'm looking at 100%. And I have confidence into the quality of the assessment. And the results, just to finish on that little example, is that the vendor they thought was not performing very well turned out to be actually totally on par with the others. But one vendor was actually flying under the radar with pretty poor level of performance. And that is something that goes at the heart of, well, if I'm looking ever earlier into the process that I'm putting in place before I get the results with accurate system, completely completely transformed my field operation. And that's what we're seeing when we talk about deploying AI for quality and improving your productivity. Outstanding. Outstanding. Yeah. Any additional thoughts? No, I hope that that makes sense for you in this regard. Of course it does. Yeah, no, I think it's a great response. And let's build on that topic of data, because I keep hearing data, data, data, data. And the concept of acquiring and organizing data is not new. There's been technology categories that have been building energy for the last decade plus, BI and analytics, knowledge management, well-defined and even growing categories. But the concept of generative AI kind of came on pretty quickly, first onto the scene. And we see investment priorities in AI continue to be prioritized. We did a early, early, early, and end of year plug-in with our community on technology investments. And AI jumped up from number four all the way to number one, from 27% to 62% most recently. So the investment in AI continues to be an increasing focus and priority. Can we begin in terms of data, just framing around a definition of what artificial intelligence is? And are there different options of AI to service organizations? Maybe your viewpoints on that topic? Yeah. I mean, look, when we talk about AI, artificial intelligence, it's at the end of the day, an umbrella, a large family of different technologies that in short, in short, is just trying to emulate human intelligence in certain forms, right? And ultimately, what's interesting in the way you're formulating this question is, what we're seeing is that as we advance in this technology, what we used to call artificial intelligence just a couple years ago, turns into just automation. And I think, you know, if we go down this path, it might lead us to some philosophical questions and very technical questions. And I'm not a philosopher or a doctor of computer science. But I think it's important to keep that in mind. And ultimately, when we talk about AI, one of the important things to look at when we look at the different terms that you hear and the different concepts is usually referring to how the system will digest different forms of data and how they'll learn it, right? So if you learn, you've heard about machine learning, which is basically the next subset, which is how you use algorithms that are not necessarily specifically explicitly programmed to do certain things, but look at a large data set to understand patterns, make some predictive analysis, etc. And now what the more advanced forms are going to be deep learning. That's another term you're probably hearing a lot. And deep learning is just trying to emulate a little bit closer the way the brain is, organized, right? So those neural networks is organized, right? So those neural networks is a term that you see a lot. But essentially, once you've said all this, what AI will allow you to do is just think about the different solutions that are available to you. And I think that is what matters the most when we hear about generative AI, large language models, those LLMs, it sounds like suddenly we have one type of technology that will allow us to do a lot of different things. And in some ways, and in some ways, it's true. But more importantly, it is what do you want to achieve and what tools are accessible to you. And if we break this down into, you know, what we're seeing and what might seem very instinctive, we'll see different, altogether, very different approach, right? So right now, a lot of AI is used for those chatbot-like systems. So interactive with a person that might be for customer service. So outward facing towards your customers. It can be internal, using natural language processing systems that can understand language. And on the other end of the spectrum, maybe would be something like computer evasion. At the end of the day, those technologies are similar, you're going to hear the umbrella term AI, but really, it will depend on what you're trying to deploy, right? And so here, it will go as to something I really want to focus on too, what AI will allow you to do is basically improve, make more precise, make faster and more scalable processes that in a sense, more or less already exist. And if they don't exist, they probably didn't exist just because it didn't make sense to do them. If you didn't have the type of scale that those automated systems will allow you to do. And I think that's a good transition from the example I gave you earlier. A lot of companies will not go to the path of that franchise B of saying, oh, I will put a complex process in place to gather tens of millions of data points that I cannot use because I cannot use hundreds of persons to look at these photos. But at the end, it is a process that you can conceive of that is very much doable with the current technology, but that AI very simply automates and make more precise. And so this is the lens through which you want to think about AI. is what are the processes internally I either wish I had or I could make much better. Outstanding. Outstanding. Thank you for your viewpoints on that. And I agree with the sentiment in terms of the application of AI. There's multiple settings in which it can be applied. And, you know, we see a lot of organizations attempting to build more of a foundational data strategy of which AI plays a role in. And then understanding the application of contextualized data across all these different application scenarios. So we're starting to see a little bit more of an organization to the approach of utilizing data and where does Gen AI fit and all these other good things. So we're going to be producing an AI handbook, actually, in the coming period. We've done a lot of research gathering some case studies and things like that. So stay tuned for that research, which is coming about. Let me ask you a question in terms of where we are with AI adoption and the market saturation. What percentage of organizations do you believe are using AI? Where do you think we are? Are we early stage still? Are we mid stage? Are we starting to achieve scale? What are your viewpoints there? Well, so that's a difficult question. And I think, honestly, I do not have the answer. And I think it's because it boils down to, how do you know, how do you use that technology? And I think today something that we'll see emerge as an idea and really our evolution of how we use this technology is, if you think about it, when you ask someone, well, are you using AI? It's a little bit like, you know, it may be in the future today asking someone, do you use computers? Do you use IT systems, right? Like I said, it will really depend on the way you leverage that technology for different applications. And so I don't think today. And so I don't think today there is any form of saturation. I think AI as a general tool for people to experiment with, to just jump into it with a lot of data point to just make wild research that might lead some interesting findings. That is something that we're saying bring more and more concepts into light in terms of how we should be deploying. It's a little bit like asking the AI, well, how should we use this technology, which obviously is not the ideal path to follow for a lot of companies. So most companies is not going this way. Most companies are going to be looking for solution they can leverage pretty much out of the box. And those are the type of companies that we're targeting and that today are really in need of understanding what the solutions are. And the concept of AI, like I said, might just be a little confusing because it's so broad. And that's why we'll see a lot of companies emerge with specific concepts. We talked about chatbots. We talked about, you know, there is such things as robotic process automation, for example, that might leverage AI. And those things bring more clarity in terms of, well, what do I want to use it for? And what is the impact in my organization? And so basically my point is, as we'll advance with this technology, we'll basically hear less about AI and more about the specific application from them. And that's what we want to be looking at. I think you're right. I love that viewpoint that you bring here. Because I think that AI has the potential to just become so pervasive in everything that we do. Like at some stage, I think that AI will augment, automate, automize, whatever you want to call it, human behavior. That it's not going to replace, it's just going to help. And I think that we will achieve a status of saturation where it's just so pervasive that it's part of everyday life. And I think it will achieve that status. It sounds like you have a viewpoint there. Yeah. And, you know, if you continue looking through this lens, when really we think about those systems as just being trying to imitate what we can do as humans, I think that's really where we see that that's just the reality is today, a lot of what organizations do are simply human processes, right? And they're more or less efficient. And it's not necessarily about replacing them, but that's where we want to start, right? It's looking inwards, looking at what we do every day. And see how that technology can just help make it better and easier. It certainly falls into a class of disruptive innovation. Let's leave it at that. How about we move into some of the serviceability challenges? We've been talking about the frontline and circling back to another research project that we always get really excited about. It happens every April, the Voice of the Field Service Engineer Survey. Okay. 88% of responding technicians and engineers last year during last year's data collection period indicated that repair, break-fix was a primary aspect of their work. And an equivalent, 88% believe the knowledge required to service products is increasing, whereas 86% indicate that the work requires greater technology, knowledge, and 3 out of 4, 75%, that the products that are more complex. So the job of the frontline is more difficult. It requires more inputs and outputs, more empowerment capabilities. Can you help us define some of the serviceability issues that you hear the frontline's facing day in and day out? And how can AI help solve that? Yeah. I mean, I think the elements that you've pointed out are what we're seeing across the board. And I think today, I feel like really the key part of what we're trying to say is just think about the processes that a person could do, but that's just not scalable enough, right? And so today, if we're looking at, well, I'm doing something very repetitive. I need to fill the same forms over and over again. Obviously, this technology is something that can become an assistant to these technicians to make that process less heavy on their day-to-day and focus on the quality of the work that they do. And one example, I mean, it's not really a specific example, but what I see when we interact a lot with those field service resources is instead of being required to do a lot of dreaded reporting on their day-to-day, which at the end of the day, ultimately is going to be looked at in a way that the technician feels they might be pullist in some ways. You know, there's someone looking over their shoulder to make sure that they did the job properly or that simply they just did it. And maybe they get a lot of negative responses when they didn't do everything perfectly, right? Something that AI will bring as an ally to these people is very simply that it's an assistant that they control as a tool, right? And so what we've done with computer vision with DeepMatic is when a technician is taking a photo for which they get instant feedback in terms of, well, you forgot something or here is how to do it better because like you said, the complexity of the equipment is slightly higher. And because of the repetibility of the job, it just human errors is something that happens. When that technician gets that robot feedback at the end of the day, it's something that they're very comfortable actually interacting with and actually correcting so that they are sure themselves that the work that they've done is done properly. And the biggest part, the most interesting part of it is where it becomes really impactful for them is once there's a system that validates that what they did is good, it's some form of insurance policy for them as well. And they have almost a guarantee that, well, nobody is going to actually look at what they did, except in obviously additional circumstances, but they know that this has been validated by themselves with the confidence that they didn't forget anything. And also that if any other unforeseen events happen that will impact the quality of the job that has been performed, they know they can also prove that what they did was done properly thanks to that system. And that goes into a lot of the transformation of how this field service job is perceived because now they can leverage technology in a way that supports them and not police them, right? And what our belief is that those field resources have today not been very served by those technology, by AI, even by IT technology overall. Overall, it has just been additional burdens because thanks to this technology, we can require more from them. We can track them better. We can track their work better. But because of the scale of all this data points before AI, it was really hard to make all of this work for them and not work just for the people in the back office. And I think that's the biggest improvement that we'll see as we improve those technologies. You said something that really struck a chord. And that was the concept of enabling, not policing. And I love that because a lot of the challenge of these digital transformations that are geared towards field service automation, they don't fail, but they struggle because of user adoption. And there's a real trickery around a mindset shift of the front line. Like we've seen categories, tech categories that have risen and fallen and have met and encountered a little bit of adoption challenge. And, you know, they kind of rise and fall in popularity because of the challenge encountered there with adoption. And one of the things that I think is really critical is the involvement of the front line in innovation. According to the same survey, 41% of field service engineers didn't believe they had a role in innovation. And 60% didn't believe that management listened and enacted their feedback into change management or improvement of the business. Right. So I think there's a real critical opportunity for leadership to listen to the front line, involve the front line, in these digital transformation. In these digital transformation continuums that everybody's on. So I just want to kind of double click on that comment because I think it's so critical and important. Well, yeah. And especially when we are focused on the concept of quality today, it means being very close to what the people who actually perform the work and the service are actually doing and what the reality of their environment really is. And so when I'm talking about, look, if you want to use AI in your organization, that means having a existing process or a wishful process to actually automate. A lot of those ideas actually come from the people who interact with your customer, with your asset in the field, because they will know what is important out there. You might have some ideas of what looks like quality, but that might not correspond to what the reality of it is. And I think it also comes back to what you were talking about earlier in terms of just having adaptive KPIs, adaptive understanding of what it is that will bring you to the result you're seeking. And that knowledge will come from the field. And again, it just goes back to saying, well, those are very granular areas where there's just so much to deal with. It's just too difficult to actually have those ideas come up and be conceptualized into systems that can be put across the board. Well, again, that is where AI comes as an answer and its capacity to, well, just answer to that need for really large scale processes, things that you want to be doing consistently. And, you know, just to kind of circle back on this topic of adoption and the whole, you know, what is causing erosion to not only throughput and utilization and automation and quality, but also what's creating friction in the career, right? Because one of the alarming trends and data that came out of the survey, the voice of the field service engineer was this continued trend of disengagement amongst field service technicians and engineers to the career. 80% of Gen Y and 66% of Gen X are non-committal to being a field service engineer for the duration of their career. Right. And some of the friction points that they encounter in their day, paperwork and admin, the feeling that they're being tracked. And that's kind of the policing comment that you made, which is why it was important and kind of spoke to me. But also the time they spend finding information and the pressure they have to work faster, not only by management for greater utilization, but also from the customer, right? They walk into scenarios that are not friendly and not so fun. And one of the things that we found in the data was that 81% of frontline agents are still picking up a phone and phoning a friend when they get stuck rather than using these really cool tools that we've made investments in. So adoption still is a challenge. What are you hearing from your customers in terms of improvement of adoption or approaches to improvement of adoption? Yeah. Yeah. I mean, I think that comes down to just the idea of making things simple and not just breaking it up into a lot of different pieces that you need to be able to handle in different ways at different times. And so, you know, when earlier I mentioned maybe to implement the right type of quality control process, you need an infinite pairs of eyes that can just look at everything that's being done. In some regards to have the right resources engaged, you want to be able to have an infinite set of mentors and teachers. And what is efficient with this concept of a friend on the other end of the phone of a mentor is that it means it's a central source of information. It's a central source of support of overall teaching, right? And so in this regards, it means that the more AI systems will deploy, the more sources of information are out there, the more important is to integrate those systems. And I think today that's something we're a big advocate of at Diplomatic because our technology is extremely focused on one aspect of the field service world, but it needs to be delivered through a consistent platform, let's say, where the technician doesn't have to juggle between different systems, different ways of interacting with the technology. And so the way I see it is, and that's why we're hearing from the rest of the community too, is we all kind of bring different bricks to that puzzle, different pieces to that puzzle, and it needs to all work together. And today for a technician to be able to actually leverage all this good technology on the field, it needs to all work together. And that goes back to the reason why we really think that the field service was not really served by this technology in the past as much as it was in the back office. It's because in the back office, you have your computer, which is a much easier machine to interact with, to pull through different sources of different tools, different data. On the field, you have your mobile device, right? It's a lot more awkward to interact with. You have different time constraints. And so all of this will only work if we have integrated systems that work hand in hand. And so when Deepomatic wants to deliver that feedback on quality control that was assessed on a photo, it needs to be done in the system that they use for other things as well, in the way that they can consume right then and there, right? It doesn't, it shouldn't be complicated to access. I love that. Interoperability, usability, critical to adoption. I absolutely love that. And that's the end of the day. If you're touching a process that makes sense, it already exists, that they're asking for, once you make it a part of their day to day, without adding any friction, the adoption is only natural. Right. We need to, we need to think about the workflow. We've thought about ad nauseum. Right. The, the customer workflow. How can we make that easy? How can we make that frictionless? How can we integrate? How can we make it easy to jump from one channel to the next? We need to parallel the approach that we take to our internal employees. It's that simple. There's a strong parallel between how we've enabled and created pathways for engagement of customers to how we do the same thing with our employees. So I, I love the viewpoints there. I think that's great. And, and, and just to conclude on this, it, what it means ultimately is that AI is not a silver bullet, right? And it's not, and, and just in the same way, every single technology that came before was also not a silver bullet. And, and AI seems like it's, it's different because it kind of seems like we can use AI to again, tell us how to use AI, but it's not going to work this way. And it's just going to require us to do our homework. Think about the processes, think about interoperability. Ask them what they need, what they don't want to do, et cetera. And because of the power of this technology today, we can probably answer almost 100% of all these concerns and make it actually worthwhile. But that just means putting our head together and understanding what type of process will be acceptable. Outstanding. Hey, let's, in the spirit of time, let's move into a different area. Let's move into what you've seen in terms, like we all appreciate that these transformations are a continuum. There's no start and there's no finish. However, your customers, what are they seeing in terms of a timeframe for implementation value received or witnessed? And can you quantify that value both initially and then over time at scale? Can you maybe frame that for our listenership? Yeah. So we, let's say we have two different types of, of, of customers. You'll have the customer who is already thinking about those processes and are just looking to have the technology embedded in that process. And they have the tool set around that enables it. In this regards, diplomatics deployment, talking from our perspective, is going to be a question of a couple months, right? It's extremely quick. That's what our technology actually enables. And I think that's what most today AI powered services are able to do. The deployment is, is, is pretty quick. And the value can be achieved really quickly. The second bucket is going to be the customers who don't necessarily have the processes in place. And that means, well, like I said, a lot of legwork to be able to make it viable for an AI process. You asked me about the value. So I'll really talk about our experience. There are very, basically two, two areas where diplomatic will provide a return investment that's measurable almost instantly. And that's going to be the elimination of certain repeat visit that any truck rolls to redo something that wasn't done properly. And also the reduction in the amount of resources that are dedicated to manual quality control. And those things can have a lot of resources that are available to people. And those things can have a large impact. What we're seeing is, in average, a 10 to 15 times return on investment in terms of just visit. Today, everything is more costly. And so sending a resource again somewhere, it's going to be extremely important to avoid that. So, you know, it might not be a satisfying answer to everybody, but really the conclusion is, AI is going to be only working for you if you're working for you if you're working for you if you already have a process that you can automate. And that's really the something that I want to leave to everybody who's listening to us is think about your processes from a human perspective. And this is where we can deliver the most value with automation and AI. We continue to hear a lot about inaccurate or inefficient dispatch and triage and diagnosis. We continue to hear about wrong skillset, wrong skillset, no fault found, inaccessible spare part to resolve the issue because we didn't diagnose it close enough. And you talk about this repeat visit and first call resolution, you know, achievement. There's a real big cost there. Our research, we've done some analysis of the different costs and what does it cost to dispatch. It's anywhere from $250 to $2,500 plus when it's multi-day and air travel, multi-day and all the things associated with the complexity of the issue, you know, depending on the asset. So that is a huge area and the cost analysis, the equation that you can create as you start to extract that against a workforce of 100, 200, 300, 400, 500 engineers. And, you know, they're not achieving first call resolution rates, one out of six times a day, six times a day, six times a day, six times a day, six times a day, one out of five times a day. Hopefully it's not that much. Depends, depends on the industry. But when I'm looking at the, the, the, simply the connection of subscriber to a telecommunication network, we have had discussion with companies who are expanding their services in new geographies or moving from one type of network connectivity to another go as high as 30%, 30% failure rates, which is extremely, difficult to handle. Difficult to handle over the long time and which you need to resolve as soon as possible. And this effective line of interaction from the field to the back office, enabling the workers to actually do a better job. The first time they go on the field is something that can have a, just such a big and quick impact in your operation that the return investment becomes, you know, extremely important in this regards. We've, we've, we've done some studies with, we've done some studies with Forrester, for example, who's look at a couple of our customer across the board to evaluate the impact in terms of those reduction in, of repeats, higher resolution rate. And in average, we're cutting that down about 50% in average. Obviously the larger the amount of, of repeat, the bigger the reduction. We've had customer who have very small amount of repeat visits who have quality built in their process already. And that might sound like not the best customer for diplomatic because we'll have a small impact, but that's actually most of the time the opposite because it means those customers have thought about their process really well. So that also means you can do a lot without the AI, but then once you put the AI in place, it makes everything so much easier and less costly. So someone who has a very small amount of repeat visit, it means they're spending a lot of money on the quality validation. So we will have that big impact. So we will have that big impact. Maybe we'll bring a 5% repeat to a 2% repeat, which already has big, big impact, but we might be able to reduce a team of maybe a hundred people going on the field to inspect your customer situation, your asset, et cetera, to practically none. And that will be a massive reduction. So it's, those are communicating vessels. And those are the things that you can measure right out the bat. But then like we talked earlier in this conversation, there's all the other little aspect that AI will enable that a manual process just will not deliver. Absolutely. A really, really interesting discussion. And, you know, I, I want to land on a sort of future proofing conversation because, you know, one of the books that I commonly go back to is this book called The Experience Economy. It was written in 1999. The co-authors, James Gilmore and Joe Pine framed it out back way back when. It's an old school book that is still pertinent to today's economy. And really, they were so ahead of their time. And they equate commoditization to the coffee industry in the book. And they talk about how, you know, at the raw bean level, it was 10 cents per cup. At the ground level, it was 25 cents per cup. At the pour a cup of coffee and you're able to charge a premium. So as you start to introduce services to the goods, you start to be able to charge premium. However, that gets commoditized because it's not differentiated. So then you see organizations like Starbucks introducing an experience wrapped around the coffee, charging a premium three, four or five dollars per cup of coffee. Then you see transformation achieved where there's home brewing, there's single pour, there's all sorts of transformations happening. There's outcome based methodologies being taken cost of coffee, the European version, an equivalent to Starbucks. They don't sell their machinery, they sell the poor and the outcome. So they've achieved outcome based services, we actually held a webinar with them. So I'm curious, what's next? What is that? And by the way, Joe Pine is authoring a new book called The Transformation Economy. So I can't wait until it's published. What's next? What is the next transformation for service? What are you tracking against in the future? Well, in some way, I have not read that book. So I will, I might have to go to go to it and grab a copy. But what I'm seeing is something that's already happening. And it follows kind of the same metaphor, let's say the same story as for coffee. I think what's going on with the service industry is we're moving thanks to this technology towards more and more self-service. Where the customer is able to either repair, troubleshoot on their own with a quality experience where it's not, well, let me Google it and it's a nightmare. But more thanks to the service of those AIs that can become eyes everywhere, that can become teachers everywhere. It empowers not only the technicians that work for you, but also your customer. And I think that is going to be a fundamental book. Book a meeting, book a meeting, wait for someone to show up. And that's something that is going to be transformed. So I think he's probably onto something here. Outstanding. Outstanding. I appreciate your viewpoints. Let's land on getting to know you a little bit better for our listenership. So two things. What value do you get from being a board member? And tell us something about yourself outside of work, something you're passionate about or something you're looking forward to. Well, I don't know if that came across, but I really, really believe that we will need to have a strong community centered around, well, what is necessary for the industry, but mostly how do we use this technology in a coordinated way? And I think that's what the service consult really brings is this community where everybody can hear from each other, discover the new systems that are going to be helpful to each other customers. And so, like I said, I systematically listen to this conversation that you organize. They always bring a lot. And so that's number one. For me, my current challenge is learning about wine. And so I'm going to be looking for a master certification in wine soon. I don't know if it's because I'm French, but maybe it's just in my blood and I hope it helps me succeed in this regards. Outstanding. That's outstanding. We'll have to share a couple, a glass of wine when we're together next. That's great. Next time I'm in Boston for sure. Outstanding. Well, listen, I want to thank you for joining me today. For our listening audience, he is Thomas Tullier. He's from Deepomatic. Thank you for being a board member for all that you do, being a board member. And thank you for the really interesting discussion. Allows our listenership to get to know Deepomatic and yourself a little bit better. For our listening audience, today's episode is available for download, consumption, sharing. So please visit our website or access it on whatever podcast channel you subscribe to. Thanks for listening. We'll see you on the next one. Thanks, John.