Name: Building AI-Ready Capacity: Managing the Handoffs That Make or Break Data Centre Projects
Uploaded: 2026-09-02T16:30:07.293Z
Duration: 1 h 30 min
Description: Building AI-Ready Capacity: Managing the Handoffs That Make or Break Data Centre Projects

Transcript for "Building AI-Ready Capacity: Managing the Handoffs That Make or Break Data Centre Projects": Hello, everyone, and welcome to today's webinar. I'm Matt, and I will be your host for today's session brought to you by Data Center Magazine in partnership with SideTracker, Salesforce, and Accenture. Now AI and cloud demand have put data centers right at the heart of pretty much every infrastructure conversation happening right now. Everyone wants capacity, and they want it faster. Whether you're developing sites, running colocation, or managing delivery on the ground, delivery is the challenge as power availability, land readiness, permitting, long lead equipment, specialist labor, and commissioning set the pace across the industry. Today, we're going to get into where that risk really builds. It's usually what happens at the handoffs when a site becomes a funded program, when development hands off to construction, and when a finished building has to become a live service ready asset. Joining me to talk through all of this are three experts working across the full life cycle. Michael Roche, director of product management at SiteTracker, James Hebelfate, managing director of software and platforms at Accenture, and Brent Healy, senior director of solution engineering at Salesforce. Guys, welcome. Thanks all for being here. How are we doing today? Good. Thank you. How are you, Matt? Very. good. Very good. Thank you. Let's get into it then. We've got a lot to get through. I think just as an intro to get things moving a little bit, a bit of context for for everyone watching, I'm going to ask for a a brief introduction to each of you, thirty seconds to a minute. We'll go around the room. Michael, I will pick on you first, as I can see you staring right back at me. So, why don't you go for it? Take the floor. Cool. My name is Michael Roach. I lead our core product team at SideTracker. SiteTracker is a critical infrastructure asset management platform. So we help companies like data center operators, developers, fiber companies, energy companies to, ultimately plan, build, operate, and maintain their critical infrastructure assets. So, of course, we'll be talking about data centers today, and my job is really spending time with people delivering complex data center programs, turning their break points into, our product. So glad to be here today. Great. Thank you, Michael. And, James? Hi there. I'm James Hilthwaite, a managing director at Accenture. I work across our teams looking after AI compute, data center delivery, and operating model transformation with a specific focus on Neo clouds and colos. My focus is simple. It's how do we turn investment and announced capacity into productive service ready compute working hand in hand with our partners, across the spectrum. Thank you. And and last but certainly not least, Brent, we'll come to you. How are everybody? I'm Brent Healy with Salesforce. I'm based here in Atlanta, Georgia in The United States. I've been here at Salesforce for more than seven years now, and I lead our high-tech media and telecom solution engineers. Most people know us for CRM, and tracking sales and opportunities in cases, but the thread that runs through the data center program is the same thread we've run for twenty seven years now, keeping every stakeholder, every commitment, every status in sync on one trusted platform now with AI throughout. SiteTracker runs the project. We help operators run the customer and service relationship that the project, exists to serve. So that's that's our role. Wonderful. Thank you everybody. Look, let's get into it. We'll start at the beginning, I guess. Site selection funded project, it seems like a good place to start. And, Michael, I will come to you first again, just to get some context for everybody who who's listening and watching. Obviously, as we said in the intro there, everybody's talking about the the the capacity rates right now, and and rightly so, I think we'll all agree. But, I guess as a starter, why don't you tell us why this is actually an execution problem rather than a a demand problem, for example? Yeah. I I think when we when we speak to our customers, really, it's it's a known entity that the the the demand is that the the backlogs are are huge and and years long. And so really for for us, it's about being able to support our customers in converting that, those contract that contracted demand into something that's energized and, a service ready capacity. And, really, every item that's going to affect that timeline is on the delivery side. It's about power, interconnection queues, long lead equipment, being able to get the specialist labor that you need, and then running commissioning. So, yeah, it's not really about can we sell capacity. That's, that's a known, entity. But, it's it's really the question of of everything in my pipeline, what's going to be service ready on the date that was initially committed and what's blocking that right now. It's it's actually funny. When we when we speak to, our prospects, as they're coming into SiteTracker, we we'll usually ask how long have you been at the company. And, in data center, more so than any other sector, what we found is the the individual is saying, I'm I'm here as of yesterday, or I I've been here. for two or three weeks. And it's it's they've been brought in specifically to, to solve this, this challenge, which which is ultimately an execution problem. Mhmm. Okay. So so thank you for that. Setting the scene a bit. I guess maybe James, Brent, perhaps if if you'd like to come in, an open question, following up. What what's changed about what makes a good, in inverted commas, site today? Obviously, the assumption is that that it used to be about the actual piece of land maybe or the location, but I guess it sounds from from what Michael's saying that it's a lot more complicated than that now. Can can one of you or would one of you like to expand on that? Yeah. Sure. I can I can take that one? I think it's I totally agree with everything you said so far. The the good site is no longer a parcel of land. It's it's an ecosystem decision. Obviously, the proximity to power generation and the ability for power to be transferred to that site is critical, but increasingly permitting, community concerns, access to, to cooling water, all of these things are really important as well. And then you move into the the other bit that I think is a build on on what Michael was saying effectively around the the the complex supply chain that supplies these AI data centers. So they need to, integrate upstream with multiple OEMs with people and companies outside of the kind of typical core technology delivery supply chain is becoming increasingly relevant as well, incorporating power generation potentially behind the meter, as well as switchgear and and other things as well that that, with the demand that's happening at the moment, the the lead times of these are are going through the roof. As well as that, the interconnection requests are driving very much increased lead times for for the, for the ability to connect to the grid as well, and that that's driving, for example, extensions to more than four years in in The US and and significantly more elsewhere. Transformer lead times as well, increasing from sort of twenty four to thirty months to three to five years. So this is a massive supply chain problem as well as a planning and scheduling problem all underpinned by that's necessary to orchestrate that. Thanks, James. And and are there any of the open question, I guess, to to to the room. Following on, we we've talked about some of those constraints and and the challenges. Are are there any for people will inevitably listening, you know, working through their own challenges and so on. But are there any of those, whether it's it's power, you talk about utility timelines, permitting, that are particularly difficult or or the real bottleneck for teams that, I guess, are kind of boots on the ground right now at the moment? think power is the hardest one. Right? I mean, we're seeing, you know, Elon, for example, putting, you know, on-site generators, you know, on one side of a river to avoid restrictions. Right? The the creativity here in The US, to get around state and local regulations is is all around power because that that is the biggest constraint right now, that we're seeing. And so, you know, understanding the capacity visibility, it it's it's it's a massive retention issue, not just a construction issue. So, it's forcing the creativity. It's forcing, some of these developers into space, which, you know, maybe sounded like science fiction a year ago, but, these constraints really are looking having us look upward and not just outward, quite literally. We just heard Sundar say that yesterday, about Google's aspirations to this as well, which kind of is bringing more, more and more critic criticality to the power conversation. Thanks, Brent. Thanks. James, I'll come back to you if I can. You you mentioned community engagement. Obviously, we all know that that's an increasingly important part of this conversation, and and and we should cover it. I mean, you mentioned it. So so to coming to you on that one, how much does getting that aspect right early on affect an entire timeline of a project? And and can can you give any key pointers as as to how you approach getting that that challenge right? Yeah. I think it's it is it's it's absolutely critical. It's, it's almost binary now that, engagement of the local community and as as Brent says at the kind of state or federal or even even city level is is increasingly important. I think this was probably more of a concern in parts of Europe historically, with regards to land land availability, water and and actually, sustainability as well and concerns around that. That's increasingly becoming more, of a worry in The US as as as because partially because power is such a constraint, and therefore people are starting to build behind the meter power, and that has other implications as well. So I think a lot of that is really important. Now it's very clear that it's it historically, the industry has has treated this as as more of a late communications exercise and and potentially, can you know, anything that can be done to shift left on this and and and bring this in earlier, is is really important. And that local support affects planning, the the political situation, and then ultimately the social license to actually operate in in an area as well. So we're seeing increasing, increasing in, incentives being paid by developers to local communities, deals being done with local governments to make sure that this, this building is in is in the kind of, interest of the local area as well. And by I guess the key thing here is by finding out any concerns on this earlier by canvassing and bringing the community along for the, for the ride, that there is still time to make changes to the design to reflect local concerns. And if that stuff comes up late, that's more of an objection, which is much harder to deal with, once the design is locked and and building has started. I mean, just as a data point, we do estimate that $64,000,000,000 in US projects has been blocked or delayed since 2024. So a significant percentage of, of of programs are are are under threat as a result of that, failure to engage this, the community at an earlier stage. Wow. If I. could just Sorry. if I. could. just on that, what James has said and and what Brent said earlier. I I think if I look at this from a tooling perspective, thinking about community engagement and and commitments that are made, what we see is that those those commitments are are rightfully made, but then the the data becomes stale. So, typically, it lives as a as a PDF in a folder, or it just lives in someone's email account. When it comes to, some of the constraints that that Brent was talking about, particularly on power, We we know that the we know that these things are problems, but I I think from our perspective, it's actually the constraint that you can't see move is the one that kills you. So it's it's about having that single source of truth and and everyone really singing from the same hymn sheet when it comes to all of this information and and a lot of moving parts in a in complex programs like these. And I. think. to to build on what Michael's saying, I think that's so crucial. Right? We don't think about community and stakeholder alignment as something that might be tracked in a computer system. Right? But that is so crucial of making sure we're keeping everybody aligned, and there isn't some email thread that maybe a mayor was on, but the, you know, the board wasn't on or the community itself. Right? We need to make sure that that's actually part of the development before shovels go into the ground that we're actually taking into account, you know, bringing the community along for the ride, and not just expecting them to join us because, of course, this is a great idea. Right? It needs to we need to all communicate and have that in a single source of truth as Michael's saying. Completely agree. Okay. Thank you, Evan. So so so to to lead us into the next set section then, at what point does the site, I guess, stop being a plan? We talked about the plan a lot and and the commitment, and and at what stage does it become a funded, committed program to to move to that next stage? Yeah. I mean, I'll start. I mean, a site becomes. committed when the data is committed. Right? Like, once once we have the data elements of what we're going to be doing, the approvals, the interconnection milestones, all of that gets bound into one one record, instead of living in someone's inbox. That's this idea that someone's tweaking on on their own. And so one once we have the the very first signal that this is gonna be a project, that's when we start tracking all of this and and and keeping it keeping it in mind. Because as as you've heard, some of these things are getting delayed four or five, six years or more. And so they're they're gonna disappear if we don't actually keep track of all of these work streams that Salesforce and and SiteTracker can work together to maintain. That's or staying organized and keeping everybody in sync is the project. Right? There's something, oh, well, you know, we have to get the power. Yes. We have to do all those things, but keeping all those things aligned is actually the management of the project is the project. Thank you, Brent. Look. Let let's take it to the next stage. Let let's look at development and and construction. James, I'll come to you. You you talked about delays. Obviously, hopefully, I'm not getting you to repeat yourself too much, but we've talked about how a lot of those delays, I guess, start before construction as as we talked about there. Where do things typically go wrong if if you're involved in these projects, and and what what can you look out for? I think it's this concept of this long supply chain that that sits outside of the direct control. So and by supply chain, I'm I mean, the physical movement of goods, but I also mean the the fact that the permitting processes that need to be in place and the, the the ecosystem partners that need to be engaged in order to ensure that and and continue to engage to Brent's point. You know? There's no point getting a date and then saying that's gonna be fixed forever and never updating that. Having the real time insight into what's actually happening in those in those enabling activities is what's critical. So the delays begin before construction if you don't get that design and map that out in a really clear way and really understand where those data signals are coming from and and track and manage those. At that point, if if you if you've assumed those are fixed, then, you know, that's gonna be that's a very poor assumption. Enabling to make sure that those customer those are linked forward to the customer commitments of of when you're gonna energize, when you're going to have your first, profitable token, that's, that's critical as well. So basically mapping out that understanding that end to end process and the data that supports that through the whole program is is the critical situation. I think it's both James and Brent said here is really the the rubber hits the road when you you've got a committed program and you've got a committed, record in in the system ultimately. Now, typically, we we haven't seen a single system being able to to handle the entire life cycle. So you've got development running in spreadsheets, deal tools. Construction might run-in a product tool, and then you've got a handoff between them that's usually a meeting and a and a document dump if you're lucky. And so, the the information that's that's required, by the construction team hasn't actually made it across from the development team. So all of that data is in the organization somewhere, but it it's not in the right hands at the right time. Okay. So so so what what gets what gets lost during that that handover when it's not managed cleanly, when it's not managed correctly? We've talked a lot about, obviously, the importance of data and alignment and permits. All of those documents that you must have, what tends to get lost during that, I guess, badly managed process? Yes. It's context. Right? Alright. It's it's permits. It's procurement commitments. It's design assumptions. They're all decisions that have an owner and they have a date. And when. they live in six disconnected tools like Michael's talking about, the next team inherits the what, but they don't get the why. And so it's so important to have these, you know, connected systems of records that can see what each team is doing simultaneously so everyone has the context of who did what and why, what happened. You know? Was there, you know, an issue with a particular community or a particular, you know, nature rights group or human rights group or, you know, it you know, it there there's all kinds of of the, you know, things that are popping out of, these types of conversations and build outs because they're they're massive. It can span, you know, multiple counties or cities. We need to make sure that we're we all are working together and seeing that context to work together. And and that's what moves the data across the handoff. It's just not a document or a series of emails like we've talked about. It really is a unified record where all the work is being tracked in one place so that we can all work together like an orchestra does, right, with with a maestro, so to speak, seeing all the different sections playing together. That's that's a very important analogy to what it takes to make this work well. It's a great analogy, Brent. I I guess to follow-up with that, Michael, that that those tools you've talked about, I guess we understand that that they're not necessarily connected. Why aren't they enough anymore? Is it just that they are they are legacy? Is it is it the pace of development? Is it the pace of demand or the pace of growth? Why aren't those tools enough anymore? Because people will undoubtedly be watching and and using some of those tools. Right? Yeah. I I think I I wouldn't say it's particularly, a legacy problem. I I would I would say it's more of a a disciplinary problem. Okay. So it's that, you know, p p six is great for running a critical path schedule. Procore is great for running job sites. BIM tools are are great for coordinating the design, but they're they're really deep in in a single discipline. And, actually, they're they're talking about, an individual project, an individual data science campus. But what they can't do is they can't they can't carry the site across the phase. So none of those tools actually exist in the development phase, is is one thing we found. You know, site trackers had great success looking at the development and origination phase of of data centers. And then none of them handle the, the handoff to operations well either. They also don't connect across the portfolio. So if you're an executive looking at thirty thirty sites, you can't you can't go and ask p six quickly, okay, which one which one needs attention this week? That that's just not a a question that that, that tool is designed to answer. Okay. So so so look. You know, let let's kind of set the scene. Let let's give some examples, some context, and some grounding. You you we've all talked about connected the the the the importance of connection. Michael, I'll ask you to follow-up, but Brent, James, please come in for for people watching, listening, owning, running multiple sites at once. Maybe give us an example of what what that real kind of connected view across all of those looks like in practice and and where those benefits really come in for you. Yeah. So for us, it's running a standardized, stage gate model. So it's having, a a common, a common vernacular being that project milestones, cost forecast assumptions, risks that are all rolling up from live project data. So it doesn't become a monthly PowerPoint collection exercise. It doesn't become spreadsheets that take two weeks to collect to to give an executive an an answer to what you might think is a is a fairly simple question. And, really, the payoff that we see by having the connected tools is is making sure that you've got, decision surface quicker. So, you're seeing a slip the day that it surfaces rather than, in the monthly review when it when it's too late. And I think the way I would challenge the audience to think here, if you're thinking about connectivities let me make it a bit more concrete. If you're a CFO and you're a VP of construction and you get asked the same question, what's going to energize this quarter? Are you both going to come out with the same answer? And if you don't have the same answer, you don't have a connected view. You've just got some very powerful tools with different reporting chains. Yeah. I'll plus one on that. I think one of the things that that we hear a lot, and and I I like this term is helpful. Right? You could have these multiple systems, but it's the seams that show. Right? When the seams are showing, that's that's the area where you need that one executive view, where we can see every site's delivery status with portfolio dashboards and capacity forecasting. We can, you know, for a long time, have been able to do predictive, probabilistic AI, even before generative. We could we could do that and understand as these things are growing, how do we adjust the plan in real time to make sure that we're not building ourselves into a corner, or or too small and that we have the right capacity, and that the customer is in a healthy place. I'll I'll give a a real example. We have a a UK telecom operator customer that put our analytics platform on top of their site tracker data that prevented a £3,000,000 plus per year loss from missed site contract expiration tracking. And so, like, that that that's a scene that was showing that we spotted because these systems were interconnected. We have the analytics layer on top. So crucial to make sure that we have that visibility and and the predictability of what's going to happen before it happens, as Michael's saying. That's great. Thank you, Brent. James, anything to add on that, your perspective? I mean, I agree with everything. I think the challenge there is also that not all sites are developed the same. So having some sort of a way, some some abstraction layer that allows you to talk to, like, we did mention to the c suite at a certain level, it gives them a view on this sort of stage gate. But then the complexity that that hides is beneath in terms of different design solutions, different chipsets, different scale of delivery, different commercial models with, the the core infrastructure providers, for example. You know, that's that level of complexity needs to be kind of abstracted below. And without a kind of strong process and data model supporting that, which is obviously underpinned by these kind of best of breed solutions that we've discussed, it it's very hard to get that executive view of of, you know, the the like for like comparison. Great. Thank you. Thanks, everyone. You're all doing great. We'll have a quick pause just to the audience to remind you that we are taking questions. So please submit your questions and we will try and get to as many, of them as possible at the end for, for Michael, for James and Brent, as if you haven't had enough of a grilling already chaps for me, we'll, we'll give you a second round towards the end there. So, yeah, please keep your questions coming in, and and we'll take a look at those and see if we can, jump on some of those at the end. Let's keep things moving. In the meantime, we'll take a look at the stage, I guess, from commissioning to to operation. Commissioning is next, as we said. Sometimes perhaps, we hope not. Don't all put your hands up in the audience, but treat it as paperwork at the end of a process rather than a distinct phase. Do admit it if you're doing that and Michael, James and Brent will have words with you after. But, James, yeah, look. Coming to you on on the subject of commissioning, why is it? Why is it sometimes seen as that? And, I suppose if you're looking at the rest of the project and, I guess, ultimately, operations, what what's the impact of not having that right mindset at that kind of crucial stage? Yeah. Look. I think the the point is it's it's a milestone commissioning complete is is a point in time, but it's absolutely not just a a paperwork collation exercise at that final point. It's the point where those years of design and delivery effectively come together into an integrated system. So if you plan this out correctly a long time ago and been collecting the right data as you go through, then you are in a much better position to be able to, to test the integrated solution, find the defects, and ultimately collect the evidence that allows you to to claim, commissioning complete. If you don't do that, the, the the operations business is gonna inherit risk. There's gonna be a lot of back and forth and and finger pointing. And, ultimately, there's gonna be challenges to your ability to deliver revenue and deliver service levels to to your end customers. So I think having that ability to track the process of commissioning and moving that up upstream as far as possible, And as we've already discussed, linking together these kind of systems of record that allows you to to have to effectively, mature those data points as you go on makes commissioning a lot less risky and a lot easier to, to ultimately deliver. And I guess your the point is testing the normal operation, testing the failure modes, testing the redundancy and alarms, and all these kind of things that we do in that that piece is the sequencing is is really critical as well. So you're also orchestrating a number of other parties to potentially support you in that in that process. So that's the other side of it is it's a it's a complex data pro and a process problem that needs to be orchestrated together and all to finish and come together at the same point, and then any defects addressed before you hand over into the application side of the business. Thanks, James. I I wanted to ask you why it's so hard to manage well, but I I guess you kind of set it out there with the complexity. But is that just it? Is it just that it's a very complex process? I guess maybe you want to walk us through some of those kind of key steps in commissioning, albeit people in the audience may know that and and be going through those at the moment themselves. But where does that real complexity come in? Are there any pinch points, or are there any kind of real moments during that phase? And and Michael and Brent, please come in as well. I mean, just I'll I'll let the others add add to this, but, ultimately, Sure. the the systems that you're kind of, you know, it's so many systems coming together at any one time around power cooling, controls, fire, security, network, compute all into one, all coming together. So that that's just difficult in itself. The scale that, that Brent mentioned in terms of the size of the the sheer size of these facilities as we get to sort of gigawatt plus, and, you know, multi gigawatt sites is is astonishing. So just the sheer number of of skews and things that need to be, in place and and and reviewed, it it makes the complexity increases massively as that goes on. And then if you look at the high density and liquid cooled nature, the sort of, cooling cooling on chip and those kind of stuff, that drives, again, another layer of complexity and another layer of new technology that all of the commissioning engineers need to sort of get to grips with. So I think, you know, one of the points we haven't talked about a lot, but I think underpins everything here, is ultimately the human behavior and the the challenge that the sheer amount of work that's happening in this industry right now, the the the the stretch that puts on skilled skilled engineers and skilled technicians and skilled trades who can can do these activities. So that that in itself is also a real a real challenge, which, you know, all comes to roost on, when we get to commissioning. I think it's also, from a people perspective, it's also about the cross party visibility. So you think about cross system visibility, but having multiple vendors in play and and no single system to ultimately coordinate and make sure that everyone's speaking to each other and singing from the same hymn sheet as it were. I I think that's that's one of the aspects that that we certainly see, with customers across all of our industries as well. Alright. And having that open API structure to make sure that all of these systems with all the third party vendors that are inevitably going to be involved, they all need to connect to that golden record. Right? That that single source of truth we've been talking about, there's going to be multiple systems and, you know, an arm literal army of people that are working on this project, and we need all of that to to find its way back to the single solitary, source of truth so that they can all work together. And so, that's where that open platform really becomes an advantage of being able to connect to all of these systems to to to see that that big picture. And that that's always been a a hallmark at Salesforce is, you know, we we don't care what system you wanna connect to us, whether we sold it to you or not. We want that data to be in one place for you so that you can manage the project effectively. And that's that that's critical when you think about the partners that you're bringing on to to to a project like this. Fantastic. Thank you, Brent. Brent, while I have you, is there is there a difference between between, I I guess, a building being that kind of commissioned stage and and being service ready? Is is there any distinction between those two kind of phases or or stages of a life for for you, for for for the guys that are on the webinar? For sure. Yeah. Commissioned means the building works. Right? The the power is there. The water is there. You know, the systems are connected. Service ready means that the organization can run it. Right? So every asset in that facility has a record. Every SLA has an owner and a failure at 2AM or 4AM or whenever it happens. Right? It always happens in the middle of the night. Right? It hap it goes to the the right certified tech automatically, right, that we're handling the post handoff, and then handling the problems that will happen when they happen, quickly and efficiently. So that's the difference between, like, a handover document and then the operating system that's running the system there on afterwards. Right? It's very important to understand the difference between building the thing and then running the thing and making sure it stays live. And that's where, you know, having the same system manage both is a huge advantage because then you have all that legacy of, well, why why did why did we reroute this conduit to this location? Like, because that's the one that's failing. Where is it now? When all of that context is available in the same place, the the solution becomes a lot faster. And whoever's managing the, you know, the facility, if it's different than someone who's building it, has all that context as they're operating. Just one other build on that, I think, is that we should always think we try and think about these metrics end to end. Like, what's actually the most important metric at the kind of, you know, c suite level or at the investor level, it's historically, people talk about buildings completed or built or, you know, commissioning done. Effectively, the practical completion is not really the the end goal here. What we're talking about here is revenue, which starts with usable service. So to to Brent's point, like, you know, you can't until it's customer accepted, until it's in operation, it's still essentially, service ready. You're not in a position to be actually, claiming that revenue. So I think, you know, for especially for the Neo clients as but but for all of the providers in this space, that's kinda critical. Getting to that as quickly as possible is the answer. Great. Thank you, James. Look, let's keep things moving. I I I can see some some activity in the Q and A. Thank you. Please keep sending your questions in. I want to give you all time to talk properly about this this connected delivery model. We we we've set the grounding. We we need to come to it, really. So I guess an open question to start and and we'll come into some specifics, but bringing all of this together, to provide some kind of key insight for for our listeners, for our for our viewers, if you really had to describe what what that connected delivery model looks like, what would you say? Where where would you say they see the most benefits right here and now? And and maybe we'll go around the the group for this one if if someone wants to jump in and and start us off. Yeah. I mean, I'll I'll start because I think that's where I left off. Right? The connected. delivery model, right, post build, that's where that delivery data and that customer commitment, they live in the same place. Right? What what did we promise the customer what they would get when they took you know, when we handed them the keys? And so the moment a milestone moves, whether it's a tenant commitment, a forecast to what we're delivering, what the service plan is, we need to be able to adjust all of that at the same time. So when you have fewer core systems that are running the build out project and then the the actual delivery of service, you know, we have one thread that can handle all of those, you know, curve balls, all of those edge cases. They can all be handled because we have that single source of truth. Right? We know which engineer put in which conduit that routed underneath a rock that we couldn't move. That that's an important data layer of context. As I said, I used the word context earlier. That is so crucial to making sure that we can solve these problems that will happen downstream. That that's why the connected system actually does matter, and that that we have as few systems that so that those seams don't become the the the canyons, between the systems that that prevent us from delivering continuity for the customer. I think I think to to build on that, it's, it it's about that that single source of truth as as Brent says, and we we've spoke a lot in in the webinar about handoffs. I guess I'd almost subvert the question in terms of don't do a handoff. Like, don't don't transition between systems. Work in the system that that is able to hold the thread right from the conception of a of an asset to to the proof of the business case right through to to when it's, built operational and and then, one day ultimately decommissioned. Because, really, the data doesn't change throughout all of all of those stages. It's just the team that that's changing. Mhmm. And so having them in the the right system where they can, have a have a seamless handoff, not not fall through the cracks as Brent says, I think that's that's that's that's how we we see success in this, in this space. Thank you, Michael James. Yeah. Look. The same is, I I think is is right, which is ultimately you wanna avoid those you you can't avoid certain handoffs because there's different people doing different activities, but what you can do is design those interfaces. So rather than designing a set of contracts that are completely separate and maybe run by different people, different contract managers, what you want what you do need to do is ensure that those interface and those handoffs are really clear and those quality criteria across those interfaces are are clear. And that that that's really much enabled by this this thread of data that we're talking about. That who who owns each decision, who owns each data object, both the the responsibility of filling it in and the quality of the data that's captured there. And that is that's really important when it comes to fixing faults at a later date and continuous improvement as well, which is really important. So you need to keep the specialist tools where they add value, but connect them through these common identifiers and ultimately a workflow end to end. And I think if you have that, then you are half the way there to to to resolving risks before they happen. Fantastic. Thank you, everyone. We'll we'll keep this moving. We'll we'll talk about how you're doing that, how these solutions do it. But we have had a question. We'll come to it now rather than at the end because it's, I guess it's about things we've been talking about previously, not to put you on the spot too much, but I do honor my call out for questions. So thank you for providing the question. Please keep them coming. I guess it builds on some of those kind of points we discussed earlier, but maybe one of you jump in. It's it's it's one to build on perhaps. So, we have the question here. What what are the most common failure points between planning, design, construction, commissioning, and operations team? I don't know if one of you wants to jump in on that or you each wanna wanna have a line or two on it, but, yeah, go for. I'm happy to have. a of a of a first go. The silence was is deafening. So Defending. yours. It it was it was getting uncomfortable, Yeah. Michael. Or we're just being polite. I I don't know who you you've just made a few. were all deferring to Michael, of course. So, Yeah. go for it, Michael. look look I mean, I I think the most common failure points. So the the the failure points are are not con not consistent in the sense that there will be different failure points between, Sure. you know, the the planning and design phase. The the the issues, I think, are are definition and expectation of those of those quality criteria at the handoffs is the is the most common failure mode. That's what we're seeing. That that coupled with the fact that, you know, that each part each siloed part of the business could that running a different best of breed solution could be doing a perfect job of managing their entire, the entire sequence of events and the data that's in that system. If the handover is not correct, you know, and the handovers might be something like between planning and design, you know, the the the ultimately, the the planning piece signed off to allow us to move to detailed design. So high level design signed off detailed design. Ultimately, then the engineering drawings and everything else that needs to be put in place, different people do each of those activities, different specialists, engineers, and and the key thing is that it's very clear which bits are being handed over to the next phase of the process. And I think that if that bit is done correctly, that's the consistent challenge that underpins the, you know, the failure modes of all the all phases. Thank you, Michael. Thank you. And thank you, James and Brent, for so graciously letting Michael answer the question. We will get you with another one at the end. Look. Let's keep things moving. I want to come to how how you three are working together. I guess, James, just very quickly to to to build on that. We we talked about people working in in silos perhaps. Let me ask you as perhaps a follow-up, but also leading into to some of this kind of connected delivery model. If you're looking at it in practice, people people doing this right now, let's say, how how do all of those teams, owners, reps, trades, OEMs, EPCs, all of the all of the kind of areas we've talked about, how can you really ensure that you're kind of reducing friction between each of those teams rather than, I I guess, working around each other, which which perhaps is what happens and and leads to some of those failures. What's the kind of best approach for for that kind of collaborative, method? So I think the that it comes down to processing data and and actually conversation. The ability to bring those together, those those teams together and not work not work as siloed teams effectively. And it requires the the it requires the owners, and and some sort of orchestrators to come together and do that with the with the help of the of the underpinning, process and data maturity. And I'll let the others build on that because I've probably said enough about handovers for for a while. Yeah. I I think it's about the the, I sometimes I I think, the idea of shared can can it can almost sound a bit, like, will your or I I've I've even seen it with our customers where it can almost raise kind of a red flag of, like, we're all sharing all of this data, but whatever happened to security, and and data privacy. And and, really, the the point here is is about the the data and the data model being consistent across this, but each individual stakeholder seeing what they need to see and only what they need to see. And, really, that's where the SiteTracker solution being built on the Salesforce platform is, is going to excel. Right? Because you've got the EPC has seen their scope, the OEM seen their equipment, and then the owner's got that, that single point of view across everything. So it's it's really the same data, but it it's looked at through, through different lenses. Yeah. I I think to to bring it all together, I I think if you think of it as one program but three lanes, right, SiteTracker owns the project execution, development, construction, interconnection, the milestones. Salesforce owns the customer and service life life cycle. So the commitments made to the tenants, the analytics across the portfolio, the predictive insights, aligned with all of that. And then the service model, once it's live, Accenture and James' team, they make it all real. Right? They integrate all those systems into the operator environment so those handoffs don't drop the data. Right? Those seams don't become chasms and canyons. It's it's the same platform thread that connects the delivery through the operations, and that's that's the beauty in the platform, in and of itself. Amazing. I love that. You you can see how you you all fit together so well across that project life cycle. James James, Brent gave you a great intro there. What why don't you kind of come in to continue and build on that? And I guess tell us about it from from that kind of implementation side. Yeah. Look. I think there's there's multiple parts to the implementation side, and we do get involved at all stages sometimes in, in a more kind of surge capacity kind of way. But also, the earlier we're involved, the more we can help to influence the the the approach to operate model, the definition of that data model, which allows for, later later execution excellence. So, but, you know, increasingly involved in engineering and life cycle expertise as well through a number of, acquisitions that we've made. And as as Brent says, like, systems integration is critical. The ability to connect with APIs that are exposed or link all of this stuff together in a way that, despite the fact this is a relatively low this we shouldn't be thinking about this as a huge systems integration program. It's more of a, a data definition and and actually change management approach, which is ensuring then that people adopt the tool and make sure that works end to end because there's no point having this, this ability to orchestrate and and and the visibility if people aren't aware it's there and aren't using it to make that decision, those decisions. that that expedite the program. So all of that part of it and that this of change capability is is also really important to embed this in the in the operator's business and scale it across the regions as well. Fantastic. Thanks, James. I I love it. I love we're at the meat of the conversation. Brent, let me follow-up and ask you. Leaders, senior executives watching, how should they be looking to, I guess, connect delivery progress with with the commitments they've made to to customers or or any other of those commitments? And how can they understand whether a site is is genuinely ready to operate, bearing in mind everything that that we've kind of talked about along the way during this conversation? Yeah. I mean, I I hate to sound like a broken record, but the the delivery status and customer commitments need to be in the same place. And that's that's how we're able to spot leakage before it happens, like I was talking about earlier. Yeah. Right? When we we noticed that delivery data was flagging and that the contract capacity risk was was coming, we were able to anticipate that and help them avoid that by having it all in that connected system. And, listen, you know, the technology is changing so fast in part because the data centers were were helped to construct. So as a result, now a lot of these status updates can happen with a with a simple, you know, message to your your LLM. Right? You can say, give me an update on the x y z project, and all of these interconnected systems can give back a a single simple readout to the executive or or manager in charge. And and that, transparency that's coming. Right? Right. They don't have to learn a new UI. They can just say, go go check on this, and that open platform, with all these interconnected systems can give them a simple, easy readout. And, hey, what are the couple of risk factor you can ask questions of the data once it's all in the same place? So the the data and the detail is getting so much more consumable now with the advent of these tools that getting, you know, genuinely ready to operate is is a lot easier than it used to be. And giving those status updates can be automated and easy to understand in a way that has never been possible before. But, again, you need to have that data layer in the middle of all of it, and they need to be connected systems. And, you know, we've talked about, you know, there are these specialty systems, but they're if they're not connected, then you don't get this single view that is so important so that you the owner has an idea of what they're getting the day they go live and not have this multiple months orientation process, that that happens later. So it it it is so important to have that platform and then connect it to a system that can can give these type of readouts that are so easy to consume, that that include all the data points necessary. Thanks, Brent. Michael, James, anything to come in on there? Yeah. I I I think one thing I'd I'd ask an executive to look at is, try and baseline a a simple number. It's what's the time from site committed to to service ready? Think about the actuals and not the plan. Try to figure out where each of the sites stand with that and where times accrued and and where the handoffs took weeks. And I I I hope I'm not putting the fear of God in anyone, but the the the fact of the matter is if I ask most executives sort of pre site tracker, pre Salesforce about that, you'd need two weeks to to actually answer that question. And so that that in itself is is the finding here. It's about having a a consistent operating model and being able to to really manage that, through the through the through the entire asset life cycle. Okay. Thank you. So, look, I'm gonna keep pushing you. I I will I will get my, my ten minute klaxon soon. I wanna make sure we're moving, and we give time for those questions that are coming in. I mean, the the dreaded key takeaway question is always hard. Right? We we've covered a lot of ground in this conversation, and you guys have been fantastic. I hope everyone listening and watching has already got many takeaways, but let me go around the room so we have have time to answer a couple of those questions that have come in. Whoever wants to come in first, but you know, people watching, people listening in the thick of this right now, what's a key takeaway? What what should be the one action when you've you've switched your computer off, you put your headphones down and stop listening to this? Where where should you go to really start making a difference or or or enacting some of this change we're talking about? I mean, my my view would be don't think about this as a as a major transformation. Think about it as a series of of interventions that need to happen. So finding a critical handoff, defining that very clearly, and connecting the data across that silo, you know, across the the the themes as as Brent says, you know, and then being able to say, well, actually, let's look at that boundary. Let's look at that ownership. Measure if that brings productive capacity online sooner as our kind of overall metric. And if if it does, then, you know, continue, continue, continue. It's a series of continuous improvement pieces where where work is already in progress. Slightly different scenario if you're starting from scratch. There's an opportunity, I think, to define all of that upfront. But, you know, there's it's never too late to to really kind of get that that level of data, definition and maturity accurately captured. And and the winners in this industry with organizations that can convert that capital into operational AI capacity fastest and and critically more predictably. And the way you're gonna do that is by making these kind of interventions. So don't think it's too, it's too too big a change. You can make incremental changes, which add a lot of incremental changes add up to a very big difference in both predictability, outcomes, and ultimately, the time to revenue. Thanks, I'll I'll. come in as as well. I think to to James' point, it's look at look at one of those handover areas. Look at the executives or or the leaders on either side of of that handover. And and as I said earlier and and and just a moment ago, have those executives try and come up with the same number? And and if you can come up with the same if both of those individuals can come up with the same number, then you've got, a connected view. Then you're you're starting to to have an operating model that's that's going to scale with, with the capacity demand. Thank. you, And. I, would, you're up. yeah. Yeah. I would add to that that, you know, for for the operators that are connecting the delivery to the customer relationship and being able to, you know, bring meaning out of multiple systems and and share the transparency across the entire enterprise, they're gonna win. They're gonna out execute. And, you know, we see, you know, removing friction from a lot of these human based tasks through the use of automation and AI that can really make your teams much more effective to focus on the things that human brains, right, can do, which is being creative and problem solve. And so, you know, SiteTracker is a great partner of ours. Our our platforms are are connect interconnected. They're using our agent force platform internally to solve tier one cases that would normally things like password resets and and things that are just bog a human down but don't really require a whole lot of thought to accomplish, right, and freeing them up to work on really, thought intensive projects. So what are what are those processes across the enterprise that that you can automate, use AI to enhance, take the burdensome things that people just don't like to do eye roll when they get that task, have AI maybe take some of that that cognitive load off of their points so they can focus on the real big problems of things like, you know, conduit relocation or or scope, you know, the the scope going up and and the project getting larger. Okay. Well, the humans gotta get back together and talk this out. We don't want them burdened with some of these lower level tasks. So that would be another thing I would I would challenge, you know, individuals watching this back to to think through what what are those low low level tasks that are that are pulling your people out of the project, that you can automate using, you know, AI and automation. Thank you. Thank you, guys. Let's move to a couple of questions. I'm on the clock here. I'm I'm keeping you moving. I've got one coming. Let's go to that. It'll be quick fire, but I'll open it to the room. Please come in, each of you. If you want a quick answer on it, go for it. Here's the question from the audience. Thank you who submitted it. Thank you everybody for your questions. How can data center operations use AI and predictive analytics to transition from fragmented manual spreadsheets, which you talked about Michael at the start, into an interconnected stakeholder ecosystem that proactively prevents project bottlenecks? What a question. Tried to catch me out reading that one. But, who wants to jump in on that one? And thank you for the question. I can I can take this to I I think it really builds on on what Brent was just saying? You know, we we spoke about connected connected systems. And if we were having this conversation perhaps two or three years ago, we we'd be talking about SiteTracker and and Salesforce as the, the tool that sort of lives across that, entire asset life cycle. When we speak now about the AI tooling available to us with Agent Force and and also with what site SiteTracker has with Scout, it's really that we're able to also connect the systems into, that single source of truth and come out with consistent answers, on on that. And I I think when when it comes to sort of that idea of critical business insight, you know, Salesforce as as as of SiteTracker, as of Accenture, have has made, a lot of money on, dashboards and and KPIs and being able to provide that insight. But that traditionally was a specialist job. To to get that kind of insight, it could take weeks, months to to answer that question to connect all of those data sources. And now we can do it in hours or or days or or even quicker even just with one prompt. So being able to have really powerful insights at your at your fingertips, it it's it's really about leveraging the power of the platforms that that are being given to you and and and sort of moving forth and and adopting what's what's available right now at the cutting edge. Yeah. And that that, Michael. AI doesn't have if it doesn't have the core data, right, if you're not working off a single source of truth, the AI isn't either. And I think that's that's a thing we've seen maybe in the in the market here, over the last couple weeks now of of people actually realizing that AI is only as good as the data it has. Right? And so having a platform where it's all in one place allows you to give it democratizes the answer. Right? Before, we had one data analyst or several maybe that were working on a project, and we'd have to ask them the question and wait for them to come back to us with a response. And while we're still using those data analysts for really important work, the AI can bring all those signals together and and give a response out to a leader and executive, somebody who's responsible for the project in a way that they can understand and understand the root cause of those situations more importantly so they can they can get you know, mobilize the teams to get it fixed, to get the right people to get it fixed, because the AI knows, you know, where that where that came from, where that signal came from. So, yeah, the the AI is the enabler, but only if the data is all in the same place. And it'll really show the seams, going back to the analogy, if if the data is not all in one place. Thanks, Brent. Guys, I I'm going to have to ask us to wrap up. I don't like to cut across anybody. It it's I feel we could go on. Right? Everybody, thank you for submitting questions. Any more that come in, we'll we'll give Michael, James, and Brent a little bit of homework after maybe to keep them busy, as if they haven't had enough of my time already. But, yeah, look, that's all we have time for for today's session. It goes without saying, I would like to give a big thank you to Michael, James and Brent for such a clear and honest look at what's happening on the ground. If there's one thing that's really come through today, it's that ambition and scale isn't the bottleneck, rather the handoffs. The moments where a project moves from one team of systems to the next, as we've talked about. Delivery models need to adapt for AI era growth, and those that embrace a model that connects teams, partners, and stakeholders, as we've discussed, will have a distinct advantage moving forward. A quick note, I have a quick note for you all. If you've registered for this webinar, you'll be delighted to know that you will receive access to a fantastic profile report in September, which will feature deep dive interviews with our panelists today. So we'll keep them working, for the foreseeable future at least. So please keep an eye out for that one and it will be coming. Thank you, of course, to SiteTracker, Salesforce and Accenture for partnering with us on this session. Thank you again to Michael, James and Brent, and thank you all for joining us. Remember, the recording will be available shortly, so please share it with your teams. Check it out again and and spread the word far and wide. But that is it from us. So goodbye for now, and we will see you all at the next one. Goodbye, Michael. Goodbye, James. Goodbye, Brent. Thank you. Thank you.