The AI Flywheel: Building Connected Environments to Deliver Real ROI for Your Organization

Okay. Got people coming in here. Awesome. Alright, we are going to go ahead and get started. Good afternoon, good morning, good evening everyone who is joining us today from wherever you are joining us from. We really appreciate you being here for today's presentation on, AI flywheel, how you can build connected environments to deliver real ROI for your organization. It's gonna be a super fascinating interesting call presentation today. I'm really excited for it. Hope you are too. My name is Josh Sparrow. I am a marketing manager here at Brightly Software. I will be your host for today. Let's see. Where is everyone able to see the screen, by the way? Anyone can just put in the chat. We're able to see. I just wanna double check that. Okay. All right. I'm seeing some yeses. Great. Sorry. I just wanted to make sure before I go too far. So like I said, my name is Josh Barrow. We're going to dive in. But before we do, I just want to go over a couple of housekeeping items. First being, yes, you will obviously be muted for the presentation. But as you are doing now, please feel free to to interact, provide any questions, comments you have in that Q and A function. Let us know if you're having any technical difficulties and we will do our best to resolve those for you. We will certainly try to get to questions at the end of the presentation. If we are unable to, we will, follow-up with a response afterwards. Along with that in your inbox, you will receive a recording of today's presentation as usual. If you want to share with colleagues or if you just want to hear Nigel speak, and listen to it a few dozen times, totally understand that as well. Lastly, at the end of the presentation, we will be sending out really quick survey. You'll get a quick survey. Any feedback you can provide us is greatly appreciated. Just helps us make sure we keep providing interesting content for you and making these as helpful and beneficial as possible. So now, yeah, without further ado, I'm gonna pass it over to our speaker today. He's our SVP, and head of product here at Brightly Software, Mr. Nigel Hughes. Nigel, please take it away. Awesome. Thank you. I hope you let me know in the chat if you, if you can't hear me. So, hi. As Josh has just said, my name is Nigel Hughes, I'm responsible for product at Brightly. So probably just to get us kicked off, you know, as we talk about AI, I we've all seen the the headlines. One of the things, I think, you know, we don't all talk about is we might not all be ready for what's coming. I I'm a I think AI itself, obviously, is a a really amazing tool, and I'm a big believer in humans. You know? What differentiates us? Can we run faster than a cheetah? Nope. But we can build a car that does, you know, and we're we're really good with tools. So I I think we're always gonna be, ahead of AI itself and that it's a tool for us, and it is gonna transform the way we work. But I was lucky enough to go to Stanford University, I don't know, three or four months ago, and we were reviewing the early results of the effectiveness of AI from the very early adopters, which were all in the in the banking industry. And one of the most interesting things I saw there was most of the projects, and and it's up on the screen now, ninety five percent is definitely most, failed to deliver any return at all. That meant, in many cases, no effect on productivity, you know, cost cost neutral. There was no giant saving. There was no giant boost in productivity. Some lost money and time. It it it got worse, not not better. So, you know, the natural question to ask is, well, what's what's getting in the way? And and primarily, there were three core factors that differentiated those that didn't work and the kind of five percent that actually blew the doors off ten, fifty times productivity improvements, dramatic jumps in quality, in innovation. So what was going wrong? So the first thing is just a super, you know, kind of natural and human thing. This is a new topic. It takes time. The businesses don't have twenty years of AI experience hanging around. So a lot of projects, they they ran forwards not knowing even, perhaps what was gonna happen if they were successful, and that led to much, much larger, than expected, increase of of cost. The biggest one, I think, for me that really stood out was data. It could be insufficient data. We just we just don't have enough. It doesn't go far enough back in history. Or the datasets were too narrow. You know, it it has a little bit of data, but they you would need more context to do it. And then finally, something that plagued and kind of amazed me that this was common in banking, but poor quality data. So you've got the data. It's pretty wide set. Goes back a long way. Not always accurate. And so kind of the takeaway from that is, if any of these if they don't kill it dead, will drive up costs. And if you've got all of these, you might you might have a a real challenge in the in the organization. And I don't think it's really, really different in our, in our world. And that's really what I'd like to explore today and then kind of look at what can happen when we get it right. But while I've got you all as a group, I'd I'd actually like to see what experiences you've all had. I know this is a real kind of group of innovators. If you could have a look, you should have a a quiz right now. Take a look. If you've got an AI initiative and there's an option if you haven't, are they hitting? Are you are you getting something like what you hoped for? So if you could kind of start submitting your question your answer now, that would be awesome. Just let us know how it's going completely anonymous. I'm not sure I want my manager knowing how all of mine are going. So, yeah, I can see a couple of people have have picked now. Oh, really building up. Because, you know, I just just wanna get a sense of your experience. What I love is this tool builds up dramatic tension. I can't see what you're submitting before, so when I bring up the results, we'll be looking at it together. Getting close to halfway now. That's really good. Cool. Let's just give it a few more seconds. Yay. Still cool. Still a few of you pitching in. Thank you so much. Okay. Okay. I'm gonna look at results. So what have we got here? We've got two point seven percent better than we could have hoped. I kind of you're doing better overall, but it's kind of the size. Kinda great. Kinda twenty percent they're delivering at some level and two two point seven, like, really blowing it away. That that's kinda that's really great. No one's disappointed. We've got thirty four percent who aren't getting a return yet. And I think given the expertise thing, that's kinda smart as an answer, you know, keep keep going. And then forty six percent that haven't got any running yet, and I think I think there are many, many people in that boat. That that's awesome. Thank you so much for for that bit of feedback. I I think this is why I I think that the that the the groundwork on on data is is absolutely key. And my vision as we we look at products, I kind of always think of this as there's lots of buzzwords here, but a kind of infinite digital thread that that extends. You know, if you think we design a building, we do the architecture for a building, eventually, we we build it, and then we start commissioning it. That's kinda happens once where we do obviously go back and recommission at points, but but, you know, that happens once. Asset life cycle management for me is actually an an infinite ongoing loop, and data is flowing around that absolutely all the time. And that's what, you know, enables us to understand how preventative maintenance work extends the life and drives down cost of ownership. I think it's what, enables us to take real world performance, interpret that data and feed that back into our plans and our budgets and and really kind of get these into an integrated portfolio, which means those tools are sharing data. So my kind of plan is already dependent on on all of those data data points. And actually, I'm not sure how many of you know. A few years ago, Siemens Siemens bought Bright we're Brightly a Siemens company. And this concept is actually really important in Siemens as well. So this lines up really nicely. Siemens have this idea of a digital twin, where you design, so that would be the architecture for for our buildings, realize that's building the building, get it commissioned, and then, you're into utilization. And so, again, Siemens, which is this, you know, kind of a tech company that's if you've if you're using anything, in fact, I was thinking the room right now probably was designed with Siemens software. The silicon chip in your phone, the car you drove in probably had its body designed in in Siemens three d design. And what's so exciting, I think, coming in from Brightly is, well, we could be the same for the utilization piece, the maintenance piece of our facilities, buildings, road networks, everything. So I'm kind of invested in data anyway, but how do we get all of that kind of great data? I I think right now when I I spend time out out there talking to clients, I see struggling with data that's outdated or incomplete just don't have time to capture it. And I think outdated data is sometimes worse than missing because it kind of looks right but actually there's a problem. I also think you you often don't have, the full and complete history, and that's like a we think of it like a snapshot of an athlete running down a track and jumping a hurdle, but you can only see the the athlete. Everything else behind is is blurry, which means we don't know how many hurdles you've already jumped. We don't know how many hurdles you've already still got to jump. You actually don't know anything about the context, the people around them. And so you don't know if you're that athlete is winning the race, is actually the last athlete in a row, and whether they've got one more jump. So they're in front and they've won. You actually think you know, but you don't. And that completely breaks that whole digital continuity and AI eats data. So you've got to have the data to drive that AI process. So that all so it makes sense. Know, hopefully, you you know, you you like the vision, definitely wanna get to AI. I know it looks like over half of you are already doing things with AI. So how do you get that data? And I think the the the first step is actually it's not cool. It's not a buzzword. It's have great usability in the in the software. So whether it's a a technician accurately capturing the work order completion data in the the mobile app, it could be perhaps you're working in public infrastructure, something to do with the road, you're you're managing fixed assets around around a city and you're you're trying to allocate people to manage it or or or even, you know, you're running a a hospital or a care facility and, you know, the data you're capturing is is needed to maintain your certification, it it's actually it's not easier than just doing it with a piece of paper because you don't in a piece of paper, you can do absolutely anything. But once you've got a piece of software that that constrains what you can do, it it demands people have got devices in their hands and things. It's an investment. But I think, and I see kind of our responsibility from the from the the kind of product side is to give you a return on that investment, many, many returns on that investment as you go. So I think usability is actually a critical part of this story. Great usability is captured and is correct. Sorry, I didn't do the builds, there we go. And one of the first areas where we're investing in AI and this isn't the kind of cool check GPT headline grabbing AI but I always love the thinking of AI is actually usability itself as well. So I call this boring AI but that's because it's supposed to serve us, not the other way around. So one of the important features we've recently released is is just nameplate scanning. Right? So all of the data about an asset, if you don't have make model serial number right, kind of you're stuck at the very beginning. So all you do, grab your phone, take a picture of the nameplate, it grabs it and it grabs all of that data. Fantastic. That's much better than any user interface for for typing, etcetera. Know, you've you've got all of that data captured. And we kind of thought, well, that's great. But a bit like I said about commissioning a building, you know, that that's not happening every day. We are working with, we are working with the assets every day, and and most of our assets are already in place. We're not necessarily doing huge amounts of new ones. So same same capability in the product for existing assets that are in the system. So kind of first investment from us in getting you and our clients into a great place is, well, let's heal the data that's that's already in the in the software, go and correct anything that's wrong, maybe even download the manuals and the service service bulletins and ensure that the data is not only complete, but it's of high quality, that is accurate from a historical point of view as well. And the outcome of that, that means just forget AI for a second. Just that investment delivers a return. Your assets are gonna be maintained in the right way, and you're gonna get more out of them. And depending on what industry you're in, that could mean some really, really important things. It can mean your school remains secure. Your production line remains running, and people can get from home to work or home to a hospital, and when you get there, that the ventilator is ready and able to to go ahead. So our dynamic data really enables you to do that. So we're making it easier to capture the data and keep it right. Where possible, we're actually trying to remove the need for you to even capture the data or worry about whether it's right. We're gonna try and fix it if we can. I also think it's critical that as a software person, it's kind of really tempting to think, well, know, I could really fix this if I owned all the data. Absolutely every piece of data of that. I'm gonna be the everything solution. The truth is even with all the AI vibe coding we can do now, actually that's not possible. In fact, it's it's probably not even desirable. Just as we have specialists in our working world, we kind of need them in our SaaS products too. The likelihood is I don't even know what I don't know about an ERP or what it really takes to manage a real time manufacturing execution system. So my philosophy and I think it becomes all the more important in an AI enriched world is that we recognize we're just part of an ecosystem. And I think as we think about our AI flywheel, this thing that as we speed it up it feeds on itself and keeps itself going, this is all the more critical. If all of our systems know, out out in our our facilities and our and our cities are running in silos with disconnected data and processes, well, then we already know what's gonna happen. Our AI projects are gonna fail at the first hurdle, and there's a a real risk we don't get our money back. So instead of trying to rule the world, we need to be a great partner or your digital ecosystem is is gonna be correct, but not necessarily consistent. There'll be different data in different places, and that that sounds terrible. So the way I think about it is if oh, sorry. We're onto the poll. Sorry. So one of the things I wanted to just check-in with you before I I went on, just take a sample from this audience. Have any of you integrated your CMMS with other systems? And I, you know, start submitting your answers whenever you're ready. But just really wanted to get a sense, and we always get a great great range of clients from from small to big. So, you know, if you don't have a system to integrate with that makes sense, that's awesome. But then if you have, it would be great to understand, what you've done and and how it worked out for you. You know, one of the things I often hear is it it's it's too hard to do or when you do it, it's too hard to keep going or it, you know, just never bubbled up to be a priority, or just, yep. Did it. Awesome. Thanks very much. So I'll give you a second to, to to pick your option. And I apologize if there's not one that's perfect, but it would just be great to get a sense. Oh, thank you. Coming in nice and quickly. Just give you a few more seconds. Okay. Let's take a look at the results. Thanks very much for all of your engagement, folks. That's awesome. So we've got a very small number. Just does doesn't make sense for us, and that's absolutely fine. Always makes me feel sad, but there it is. Twenty percent. That's a that's a fifth of all of everyone. It's either too hard or it's too hard to keep working. Forty five point six percent are not a priority. I I think that's really, really common, and and I think the good news is AI is perhaps gonna help with that priority and and great. Fifth of people, they've done it, and it's working brilliantly. Thanks thanks so much for that. It's really, really interesting. So if you're gonna if you'll allow me to stretch the digital thread, digital metaphor just a little bit too far, the way I kind of think, why I think it is important and particularly in a modern in the modern AI age, integration is really, really key. Is really what we should be thinking about is weaving all of our data together into digital fabric and really getting great products, each one doing what it does best, keeping them aligned but not, you know, kind of weighed down with each other. So what integration studio that's the the part of the Bright ePortfolio that does that is is helps us be that good partner. So, you know, there's a couple of categories here where something is used by a lot of people in the industry. We might have an off the shelf common integration, you know, something like NetSuite, d three sixty five, BuildingEck. It's all of those kinds of tools. You buy it. You set up a little bit of configuration to to to wire things together, but it's largely copy paste, and you're done. And now part purchase order status is right and kind of sounds boring to you, you know, kind of think about, well, but why do you care? We don't wanna walk to the stockroom and find the part isn't there. We just look on the system and not go and waste that time, maybe get something else done. And it could be even worse than that. You might have multiple buildings on a on a on a campus with with multiple, buildings, and you maybe even have to drive between them if you're, you know, a school district or something. And so if you could connect to your real time system, Desigo is a is another one, you could just go and have a look about complaint that the room's too cold. Maybe go and look at the window sensor and say, yeah, it's too cold because the window's open and just make a phone call and say, close the window, you'll hit your set point. No problem. So, you know, it's just a couple of examples where correct connected data deliver another return. And, you know, just like using a piece of software, put some constraints on. There's work involved in in, you know, integrating multiple systems together. But this is just even haven't even used AI yet, and we're already getting a return from this investment in data. Of course, everybody works a little differently. You might have, built something special, that we don't know anything about, and that's that's that's absolutely fine. So we also provide, APIs. So particularly, you know, especially common for larger clients that we need to help them tie together multiple systems. Some of the systems only exist at that client, but, you know, just sometimes it's just not common in the market, doesn't mean it's not critical for you and that we we we shouldn't be helping. And so with APIs, it it's you're able to do this for for data we don't even know about, and and I think that's really critical. And then we can look a little bit further out. So AI agentic systems, and I don't know how well that all the buzzwords are known, but I always think of an agent like a conductor at an orchestra. And all of the instruments are sitting there. They're all sitting there quietly waiting for the orca the the conductor to wave his baton in their direction, and then they're gonna go ahead and, you know, bang the drums, play the keys on the piano. That's agential AI. So you've got a a conductor there asking multiple different little AA LLMs, large language models, the things that are sitting behind chatbots to go away and do a bunch of tasks. And just like we're going to fit into a digital fabric, we have to fit into a fabric of AI as well and support us driving other tools and other tools driving us responding to, other people's conductors. So really turning this into something much stronger than just a thread, but a digital fabric that you're able to use to to get more and more benefits, to optimize what you're doing and not do not do the wrong things. And I also think it's time for you to kind of get paid here. We've got a product called Data Share. And the payback here is how can we go beyond just our data before we even get to AI, enable you to just get something back from that digital factory? So going back to that picture of the athlete jumping the hurdle, how do we go from just the Brightly data, the athlete jumping the hurdle to show you the full context, perhaps the full full history, perhaps you to be able to see visualize what the impact on data rates were from work orders over the last year and tie that into potential salary change coming in from the HR systems. Perhaps it's you know which work orders, which parts hold up the most work orders are causing the most extended lead time. Or perhaps you know you were given budget to do something or you want to do something and you want to justify it to your to your boss, but it's going to involve getting data from the ERP system for part cost and lead time, the HR system for what labor costs are, and fusing it all together. And that's exactly what data share is for. And it kind of sits in the middle, call it like the center of the bow tie, means actually it doesn't really matter which of our tools you're using. The data is there in a common format and it uses industry standards to help you tie that together. And so it's really about you fusing and being able to visualize that digital fabric to get what you want out of it to either understand what's going on or perhaps position some investment you'd like to make. We're getting pretty close though now to the AI piece. This is you looking at the data, but there's just kind of one more piece or the beginning piece of the AI flywheel. And so think about the investments we've been making so far. So we've made sure our users are capturing the right data. We've integrated it together with a broader network, making sure that's all consistent, and we've been looking at the data. Well, here's the first place we can now get start to get AI to pay a benefit with with analytics. Okay. So here, you know, in this case, we've got a a forecasting dashboard, But here, we're able to leverage the fact we talk to thousands and thousands of clients every year, understand what's driving things, and we're able to present sets of dashboards and data which highlight interesting things for for you that, you you know, you might want to act on or, you know, understand work order forecast. You're gonna have a spike, you know, from for many people, there are spikes over the summer at Christmas time when there concerts are coming on at your venues. And so being able to see those forecasts, it's possible for us to design just like we design a good user experience, a great set of analytics for you to fit to to to guide you and draw your attentions to things. And AI is right behind looking for where is the discontinuity in the trend and and what do we need to show. So that's really the the the kind of solving of the data problem and getting our first first items. So, you know, dynamic data is making it really easy to get the data in when when, you know, it's it's just data that has come from a human. We've got the users, and we can integrate it all together and provide analytics either done by you or by our tools to to help you get that. So already a lot of benefits coming out of this. The next thing, though, that we really want to, build into is, like, where where are we gonna go go next? It's kind of all cool. It's also a a little bit of kind of where we are. We're at a we are at this technical inflection point that I I introduced at the beginning. So I wanna talk a little bit now, about some of the investments we've made and are making. And there is a a real benefit in being part of Siemens here that I think is is kind of really exciting. I don't know if you know who these two fellows are, but on the right is Roland Bush. He's my boss's boss's boss. He's our CEO of Siemens. And at CES, which is the consumer electronics show that always occurs in Vegas, feels like biggest boondoggle in the world. But at CES, Roland was on main stage with Jensen there on the left, who's the CEO of NVIDIA. And if any of you know, NVIDIA is making this the silicon chips that are driving this AI revolution. And actually, Roland and Jensen have been partnering for a long, long, long time. We've been lucky enough to be able to listen to Jensen and Roland talk actually for for years, long before Siemens started decided to to to acquire Brightly. So what they were announcing at at CES is just a further deepening partnership to build out the industrial AI operating system for manufacturing. That's a little bit beyond the scope for today's webinar. But what I wanted to kind of share with you all is, you know, this project is fantastic, but what we've been able to do is reach out to our partners in Siemens, take advantage. Remember what causes those AI projects to fail? You know, Lack of expertise. Well, luckily, we've got from from the the fruits of this of this partnership, we've been able to reach in and understand what does and doesn't work ourselves, taking advantage of the investment that Big Siemens made for all of us. So when you think about leveraging the digital twin and Brightly's kind of model digital twin of asset lifecycle management, the different stages and the different places AI can contribute and it's giving us some stepping stones. So first of all, we capture the data, right? Then as we just saw, we can have some analytics that could be actually enabling you to handcraft it easily and in a really maintainable way. The next step though is going beyond just, hey, here's a thing to here's something that should have something done about it, and this is the advice on what you should do. What and then finally, we could have the AI take action. It could do something. I I think there's gonna be a a time for that. But, you know, right now, I think it's critical we keep the human in the loop and and retain control, and we we can loosen that over time if we need to. So little mouse pointer on there. I'm gonna ask somebody to click on something to to make sure it it happens rather than just handing the keys to the to the the castle. So I really want to focus on advice. And I think going back to just just AI to make the user experience better, but instead start to turn AI into a I'm gonna steal a quote from Steve Jobs now, you know, bicycle for the mind. Something that enables us to take the same body and go three times as fast. So so you can get where you're going really quickly. So I'd like to know what you think about that last one, keeping that human in the loop. So my question for all of you is, would you like to hand the keys to the AI and have it act without human intervention? So first option, no, and I never will. You've seen the Terminator. Right? Unfortunately, Arnold couldn't be here today, but if you can pick that option, if that's how you feel. No. We need to give it a little bit more time, to mature. I think that's a very prudent response. Yes. But it needs to be very transparent. I think that's that's also another very valid approach. And then finally, yep, no concerns at all. And, you know, I'll let you all start to get your answers in, but I've absolutely heard that one Often paired with, you know, you've met people. Right? You know, they make mistakes too, and we'll just need the same processes to catch the the problems from AI. I've also had some of the best conversations in my career in product with people when they if their answer is no, and actually it's never as glib as a bad Terminator joke. Actually, normally, there's all kinds of super interesting feedback that comes that I'll share with you as we get a little deeper into the slide deck. So, got a great bunch of answers. This is one everyone can answer. I'm sure you know how you feel about this. This wall just beginning to slow up now. Come on, folks. Let's get to halfway. There we go. Okay. So let's see what you told us. Thirty percent. Yep. We're not giving the keys to Arnold. Fifty percent, you know, measured. Let's give it a bit more time to mature. Twenty percent on transparency and very much where where I'm only one point six percent on on zero zero concerns. Definitely think this is something the the the the tailor, tail measure twice, cut once here. So thanks so much for your feedback there. Really great to really great to see. And I mentioned that I've had some of the the great great conversations driven by by this question when I when I chat to clients. And one one of the great examples, we've got this we've invested in this duplicate work order detection capability. And actually it turns out AI is amazingly good at spotting duplicates using symptom descriptions, parts, any number of other factors, and it it reads everything. So it can spot with a really high degree of accuracy that some work request has already been made. It's already gonna be done. So we could act, right? We could just automatically, delete it. Turns out, you know, there's a couple of things here. So first of all, a high degree of accuracy is not the same as perfectly accurate. So we wanna be responsible. We don't wanna delete something that was critical and needed to be done. But as we designed it, as I said talk to clients, we learned actually, it goes way way beyond that. So here was the the use case that that I was I was taught. So let's say you're running a senior living home and one of your residents submits a request to fix the lamp in their room. Okay, that's a request. If they do that again and again for the same lamp, that's not duplicate work. That's a diagnostic step. You know, the maintenance manager said, well, as soon as we get a resident doing that, I go straight to the on duty doctor because perhaps these are early signs of the onset of dementia and that we need to up our level of care for that individual. It's a diagnostic step. So really great opportunity. So we can make it a bicycle. We can spot the duplicates and then focus on an amazing user experience for the user to see all of the key data, make the design decision, and and make the merge. But here's the flywheel. So all of that happens and the human takes the action. But here's the flywheel. By catching them and removing the duplicates, you've just improved the quality of the data, which means future iterations of the AI are going to be better. It won't learn that actually, no, that's not duplicate, that it's normal to have like like that. That through those three words, significant difference. Or you're trying to understand what most of your work orders are. Well, now that report is completely wrong because there's so many duplicates distorting it. This is the flywheel. Better data enables AI. AI enables better data. So, you know, you don't end up ordering more light bulbs for a lamp. You instead get care to where it's really, really needed to help that person. Another one, that, we invested in is scheduling. This is a really complicated problem. Sometimes it's way beyond just doing it in your your own team. You need to do it across teams, consultants, even people outside your organization. And so smart scheduling enables you to rapidly create plans for your organization. But again, wanna keep human in the loop. So design a great user experience so that you get the plan. It's good enough. You can understand what it's done, but you can tweak it around. I think the other advantage here is you get to decide how you spend the benefits. So maybe doing the planning saves you couple of hours every morning as you're trying to get the work done. How do you use that time? You could take the time saving and get something else done off your task list. You could also line it up against your production schedules or other work, you know, data share anybody, and make sure it's a good fit. It's not adding risk to something else that's going on in your organization. Or you could tweak the goals you gave it and actually try out four or five different versions. Well, what if I prioritize this over this? Can I squeeze a little bit more in? And again, that's that's the goal is not to make you feel like Mickey Mouse in Fantasia while, you know, all the brooms and buckets are marching around flooding flooding the room with water. But to give you a return on investment on that great data. And and by the way, because the schedule was generated and it's all data and it dotted every I and every t, all of that data is good too. So we're feeding into that flywheel yet again. Of course, an AI webinar wouldn't be an AR webinar without Copilot. And don't get me wrong, it's great. It can read all of the manuals, provide step by step instructions to fix the rattling noise you reported, and there'll be the right instructions because dynamic data made sure it had the latest version, of the manual, and it grabbed the latest service bulletin where they tweaked the procedure with a new recommendation that increases performance. But remember that athlete jumping the the hurdle. Right? We we wanna see that full thing. Because you don't have a bunch of duplicate work orders because you had a great experience and all your data is right, it can read the history. It can look at the PM schedule and what's coming up. It can summarize it in multiple languages. So it doesn't matter if the manual's in English, but you you'd like to read it in Spanish. You can do that and ensure that you've got the full context of the entire race and not just a snapshot. And so, you know, again, just like duplicate work orders, here's the flywheel. When we were looking at this, we looked at the millions of work orders that are created in our software each and every year. Most of the descriptions were fixed, replaced, filtered, didn't say whether it was oil or air. It was bad data, not because technicians are lazy. They're not. I'm stressed just thinking about all the work they do because it's really hard to type on a mobile. And talk to text doesn't work in a noisy playground or in a noisy factory floor, or it just plain feels weird to be talking to your phone in the middle of middle of your working day or because they're in a hurry because the working list is long. So you type as little as possible and you move on. So one of the things our Copilot does is capture all of that for you. You followed its advice, said you were gonna do that. It can capture that data for you, and we feed the flywheel again, turning it turning it around and ensuring that ensuring that the next time you come around, the data's even better than it was before. Your investment is giving you a bigger return on your investment. So oh, that's not working. So the circle is complete. We've now gone all the way around. We've improved our data and now we're using our improved data to feed an AI which is helping us continue to improve our data and make sure it's right and give folks like me ten new ways of finding out how we can give you another return for that investment you've made. So we invest in helping you capture it. We invest in just using the data to help you understand what's going on and what you might like to do or perhaps what you should pay attention to, or we can then take it to the next level and give you some advice and guidance and hopefully even make your job data and make it easier and make important tasks easier and easier to get right. And then it can feed itself again to go around on that AI flywheel. So that concludes what I prepared for you today. And with that, I think we've got some time for questions. Yeah, thank you Nigel. So yeah I've kind of aggregated some of the questions here. I will say the sort of the the biggest question I am seeing, and I'll give you a chance to answer it or however you want to go about it. But a lot of people asking if when these capabilities are going to be featured in, you know, products, AE, Energy Manager, Predictor. Yeah. What are your thoughts there? So so let me give you the give you the answers. So for AE, that's been our front runner in in building out a bunch of these tools. Although I do wanna come back to predictor if I can. So for AE, some of them are available right now. Dynamic Data with all of the AI product onboarding is available right now as is DataShare. The others, any any second any second now. So I think Copilot and Duplicate, they have been in the hands of some clients. Please do reach out to us. We'd love to hear that you'd like to get your hands on them, so please reach out to us. That's something we can work together on today. Predictor is really interesting. So first of all, I'm going to give an answer but it's a little bit of a nerdy answer. So absolutely we're trying to build out the way we've built it is we're building these things as components. So we'll build it on one product, perhaps a product that that where clients tend to have the absolutely best data. That proves it out. It works. Clients actually getting a return on their investment in our into our tools, and then we can easily use it for others. So tools, whether it's Predictor, Workshub, Confirm, everything we make does now have access to these components, and this is why I'd love you to reach out for us and be noisy and demanding because that will really help me just organize which order we we do that in. The really cool thing is Predictor is AI at its heart, by the way. It uses genetic algorithms to produce its, its plans and its investment strategies. So actually, Brightly has been investing in AI for a little while and Predictors is a great example of that and benefits so many people. So it will come more directly, but I think it's kind of cool that Predictors has been AI since before it was cool. Awesome. Another question here. Can building information systems feed live monitoring data into Brightly Solutions? Yeah. Absolutely. We've got a couple of ways, of doing that. The one of the easy ones easiest ones is BuildingX. That integrates directly with Brightly. You can even, configure it to automatically create work orders for you. We can also integrate with third parties. So, you know, just other systems of BuildingX doesn't already support it. Actually, another benefit, Siemens. Siemens Digital Industries make a little box you can buy and just plug in, and BuildingX knows how to talk to the box. And x three is called the x three hundred, really great name. But but it will talk to lots lots of things. And then if that doesn't support it, as I promised, we've got APIs and we can stitch just about anything together. So yeah. Absolutely. And we we have actually some clients, you know, doing that right now today. That stuff is in production. Awesome. Let's see, looking through. Will AI be able to create a task list for preventive maintenance and inject it into the work order? We are not there yet, but yes. So Copilot actually will tell you what you should do. What it's not doing is that last step at the moment of actually I think that's great and if you'd like a job in product, do keep an eye on our jobs board. Actually getting it to turn that into a formal list and put it into the data flywheel, absolutely that's coming. First version of the Copilot will report it to you and we'll take that next step. Awesome. Let's see. Can the planned embedded AI clean data, for and the example they gave was find these work categories and change them to x y z. Yeah. That's I mean, that's really interesting. I think that's something we have tools we use AI to do that internally. So for example, if you're moving from one of our older solutions onto something like a a one of the more modern ones, you know, we have tools that actually do that. But at the moment, just as I was saying, they actually flag all the records where the data just looks wrong. And, you know, often the AI can can tell you why it looks wrong. I'd love to build that into the software and give you a a a button to do it. Not there yet, but that is something we're doing internally. Internally. Okay. So we'll get that. Yeah. Let's see. It's a little bit of a long one here. This was from Russell. He said his team doesn't have good condition size and material data now for their assets. They're adding spaces on the work orders for their field crews to start entering this data when they work on an asset. Could there be a way for AI to read the work orders and add that new asset condition data size material into the AE GIS data in real time rather than waiting to download the work order data at the end of the year. Yeah, absolutely. In fact, we're we're working on that. That's a use case we're actively exploring as well as using the GIS data in the other way around to help with scheduling and things like that. So the scheduler, you know, could take that into account. A hundred percent not in the first version of the of the software, but that's that's exactly that's possible, and it's it's a really great use case. And also I really appreciate your kind of candor in the challenges with asset data. I think you're in lot of people are in that boat. So thank you. Let's see. Just clarification. By Copilot, do you mean Microsoft Copilot? No. No. I don't. I I mean Brightly Copilot. We haven't named it. We've just just calling it the maintenance copilot. Maybe we should give it a a fancy name. I would love any suggestions in the chat. Nothing rude, please. But, no, we're just calling it the maintenance copilot for now. But, no, it's not Microsoft copilot, and it's entirely Siemens technology. That's where that NVIDIA partnership really, really plays. We've got a we've got our own versions of, of everything. And for me, that's really important because I'm a big believer in ethical AI, and I want to know how our client's data is being used. I'm not handing it off to a third party. Awesome. Okay. Got one more here for you. Will, this functionality allow, users to adjust staff and assets? For example, Jim switches positions with Jane, things like that. Is that is that will that be part of the functionality? I'm not sure I get the question. Is that for example in scheduling when you when you tweak who's doing what or responding to a change in the HR system? I'm not sure if we can I'm allowed to ask follow-up questions. That's fair. Let's see if there is any additional insight there. Yeah. Not totally sure the additional kind of context for that question but that's that's from Daniel. Daniel, if you'll, you know, wanna wanna send us in little more information, can get back to you on that. A lot of the other questions, there were questions around things like pricing and future updates that maybe we will can kind of maybe we won't answer on this call, but we can we can, you know, come up with it, figure out the answers and and respond afterwards. We have a few more minutes. I want to see if any more come through. Just seeing some options for AI names that are very interesting. I know. I I was I actually because my temptation is to stay on and just chat with all of with all of you. So, love it. Yeah. Okay. We do get one more from Stanley here. Can AI learn how to emulate PMs that have several levels of maintenance with weekly, monthly and annual, PMs so they can do away with the redundant work? Yeah. So the the answer is actually and it's really, really, really good question. The answer is yes. But it is one of the areas where we get more consistent results on the overall correctness. And and just I I'd like to bake that a little bit more before we we let it out because it's a really great use case. And even in computer science, kind of we call that coalescing when you kind of figure out how all of these schedules interact and it's a really it's a thing humans are good at, but computers haven't been. So just superb idea, but we're just getting it to to do that super efficiently and not generate a plan that actually is quite complex because it typically has to be over a time period like a year. It's quite hard for a human to easily review it. So we've got a few UX problems and and things to solve first. Great. Cool. I think that's kinda nope. Anything to add, Nigel? Any last thoughts? No. I said cool. I just said cool. No. No. That's it for me. I I really do wanna audience, though. Thank you so much for your participation, your great questions. Please do reach out to us. You know, we'd love to to hear from you to to talk about all the, you know, kind of everything, various options, things, ideas, etcetera. So please do reach out to us through your client success. We'd love to talk to you. Yeah, absolutely. And everyone, like I said, if you can take a moment, at the end of the call to provide feedback to the survey, that would be wonderful. But otherwise, yeah, thank you everyone. I hope you enjoy the rest of your day. Thanks, Nigel. Thanks, everybody.