TRASCRIZIONE 🔗
MIchael Bograd: Welcome everybody. Thanks for joining us on As Planned, a project controls podcast. I am Mike Bograd, a Executive Vice President at Stellic. with me as always is Pedro the producer who is in the background running the show. And today we have a very special guest, Taylor Moore. Taylor, do you want to give everyone a kind of brief introduction about yourself?
Taylor Moore: Sure. Thanks for having me guys. I'm Taylor Moore. I'm a vice president here at Stellic.
MIchael Bograd: Mm.
Taylor Moore: been, you know, in project controls for about 10 years now with a focus on on our life sciences clients down here in in North Carolina. So I appreciate you guys having me on. I'm looking forward to the conversation.
MIchael Bograd: Love it. So, you know, one of the big initiatives that you've been working on lately is dashboards, right? You know, you've been kind of taking our working with our dashboard team. You know, one of the big kind of value ad that we like to think we have at Stellic is we have a dedicated dashboard and analytics team. and all they do is focus on dashboards. And one of the dashboards you're building out is is very cool. It's it's for a a very prominent client and we and we love that. But the topic that Pedro has produced for us. is called the dashboard that nobody trusts. So yeah, that was my reaction as well. So what does that mean to you? When you hear someone say, hey, we have a dashboard from the project, but we're not using it, you know, what is that, where does that typically lead you?
Taylor Moore: Yeah, I mean we've you know, me and you, we've been on so many projects where we've seen these dashboards that maybe somebody thinks up at the beginning of the project and it's just maybe like automated to just spit out data every week. Nobody's really checking on the procedures, what the data's saying. you can very quickly, you know, lose sight of what you're actually tracking in those dashboards, and before you know it, it's just some tool sitting in a folder somewhere in SharePoint that Nobody really uses, and if you check it, none of the data's really that accurate. So, you know, we've like I said, we we see that all the time, and we that's something we we really do try to avoid on our projects because the ultimate goal of a dashboard is really to consolidate a ton of information, specifically project controls information in this instance, and really provide you insights and trends and things that you can actually use on the day to day to better plan your project, to you know, ensure your executing things on time, under budget. So it's really just, you know, making sure the tools you're using are serving the purpose that you intend them to.
MIchael Bograd: Yeah, completely agree. You know, you mentioned, hey, we've both been on projects where dashboards have been created and no one uses it uses it. It's kind of defeating, right? You know, when when a dashboard's put together, someone takes the time. I think a lot of it kind of ties back, my opinion is a lot of it ties back to, you know, who's the end user? Who what is this goal who are we creating this dashboard for and how can they use it? I think if you don't start with that question in mind, you know How do you get to the right outcome? So I guess you have kind of a process that you like to use that kind of gets you from, you know, hey, we we're thinking about a dashboard to we now have a really good dashboard in place that is getting used regularly.
Taylor Moore: Yeah, definitely. I mean, when you're developing these dashboards, it's really knowing your audience, you know, at a high level, what questions do you want to answer with this dashboard? and I always start with, you know, the philosophy that your dashboard should answer the simple questions first. obviously we want to, you know, get more advanced dashboards with trends and insights and things like that. But, you know, part of developing a good dashboard is ensuring that It's got data that folks can trust. And the fastest way to do that is really to show, hey, you know, the the the simple questions, the the milestone dates that we're showing, the the budget of a project, you know, how much have we built to date? You know, ensuring all that stuff is accurate. and I'm already going on a little bit of a tangent here, but it really comes down to, you know, talking with the client. What do you want to use this dashboard for? Who is going to be using it? You know, high level VP is going to use it a bit differently than maybe like a foreman on a project or some of the engineers. You know, what do they want to use it for? How is it going to be used for reporting? Who's it going to be presented to? really, like I said, taking all those considerations, getting that very simple, easy to answer information in there, and then kind of building upon that of saying, Okay, what insights can I get you from the information that we do have? And then basically taking, you know. a lot of these dashboards that we've developed in the past, looking at them in terms of widgets and saying, you know, this is what we've done for another client. They found this useful. what are your thoughts? And really just going through this iterative process of, hey, this is what we're proposing. What do you think? You know, is this something that you would find valuable? This is what insights that this this widget or this graphic or this, you know, metric KPI can get you. And really just ensuring that's something that they're going to use. you know, talking with the folks that would actually be using it. if they're going to extract the information and put it into slide decks for you know monthly or quarterly like fiscal reporting, you know, what do you typically see in those? What information d do we need to extract from our dashboard for that type of reporting? So just really understanding where the dashboard is going, who it's going to, and how they're going to use it.
MIchael
Load full transcript
Bograd: You mentioned something early on when you were talking about the data, right? So you know, you you are working on a dashboard right now. hypothetically, you know, on some of these projects, we collect data from not just you know multiple projects, 10, 20, 50 projects that we're collecting data to kind of ingest and look at. You know, it comes from different sources, you know, P6 schedules, Excel files. exports out of PMIS systems, direct APIs into PMIS systems. I think a big component of this is also quality control and fidelity of this data. You know, what what do we, you know, what are you doing to kind of put in place so that we ensure that this data is accurate, right?
Taylor Moore: Yes, certainly. And that's usually the the biggest challenge when developing one of these things, you know, 'cause I mean we've gotten lucky on a lot of our projects where we're the ones in control of the data that we're going to be using. You know, we're the the cost engineers or the schedulers or risk managers, change manage management team on these projects, so we can ensure that that's data that data that we're using is clean and formatted correctly. But to your point, when you're getting to these program level dashboards where you're getting trackers and, you know, like you said, schedules, integration into these tools from, you know, all over the US with teams that you maybe not have met already. You know, that that development side of, you know, how do we want to standardize this data? How do we want to establish processes and procedures so that the data that we're inputting is standardized and ultimately meeting with those teams and getting the buy-in from them to say, Hey, you know, we can we've got some pretty smart people on our you know, kind of software side. That can write the code that says, you give us your tracker however it looks and we can extract it into the format we want. But that comes with, you know, conversations. You know, am I looking at apples to apples or am I trying to extract something that's a totally different value from what I'm you know grouping it in with these other projects? So it's, you know, developing the dashboard, what's it gonna look like, who's gonna use it, and then you know, how do we get the information? How do we establish the processes and procedures? How do we get a normal cadence of getting these updates? and ensure that everything that we're mapping is correct and we're not, you know, putting the wrong data in the wrong bucket.
MIchael Bograd: Yeah, you know, I've we've I've definitely been on projects where where I've inherited a dashboard, right? It's you know, here's a tool we're using. And I feel like the second, you know, an error gets flagged on one of these dashboards, it you kind of lose trust in it, right? And then then it everyone around you also kind of loses trust in the dashboard and using it as a tool. You know, what's what's something we can put in place or what's something that that works
that can then say, Hey, you know, we recognize that there was this gap. We've now corrected it. How do we how do we course correct?
Taylor Moore: Yeah. I mean ultimately your your dashboard is kind of the like visual tool that's built upon a mountain of data, processes, procedures, people, teams, things like that. and I think it's important. And one thing we certainly, you know, emphasize and try to show in our dashboards is where is this information coming from? You know, if the the front page of the dashboard gives you a list of trends and insights and actions it thinks you should take, that dashboard should also give you the ability to kind of go one step deeper, you know. It I'm am using one vendor for all of this type of equipment and it's, you know, four hundred million dollars across three sites, you know, where where is that data coming from? I should be able to say, that's an interesting insight. Let me go one step deeper. You know, here's all the data, here's all the vendor data, here's all the cost, here's here here's the project data. And then I sh should be able to go even one step deeper than that. You know, have all the data listed somewhere, but I should be able to point to the processes, procedures and people where that data is coming from. So if you do have a question, you know, this this insight it's giving me looks a little strange. this KPI I'm seeing is a a number that doesn't feel quite right. I should be able to tra like, show my work essentially, trace that data back to say, this this value is coming from, this table. This table is culminated from these hands full of processes from these other tools that we're getting. So you can really go and see, If this value is off for whatever reason, maybe it's not the the process that we have of, doing the math, doing the calculations to get you that value. Maybe it's the data that we got is incorrect. And we should be able to point to that. and if it's wrong and, something is off, it it makes it a lot easier to to diagnose, fix that issue and, really show, we we know what we're doing here. This w we have everything documented. The data you're getting is real. It is from people. It is from projects. And we can show you exactly what those are.
MIchael Bograd: sure. you know, the bell of the ball, the bat feels like it's been a while now since AI has kind of come out and and taken construction industry by storm. you know, what are your thoughts on AI and dashboards? You know, d is it required? Is it not required? You know, where where do you stand on that topic?
Taylor Moore: I mean, I'm certainly pro AI, but like the the introduction of anything new, especially when you're in like industries like construction where things can be a little bit more old school. People prefer, pen and paper, showing your work, l w understanding the math. I I I think you do need to be sensitive to people's hesitancy. primarily like I said, on these big pharmaceutical projects, big life sciences projects
where you have a lot of proprietary data. You know, the the first and most important thing when using AI is ensuring, sensitivity to that, making sure you're not taking proprietary data and using it to teach public models. fortunately a lot of the large pharmaceutical companies around the world in the past year have started partnering with these these AI companies so that, you can use the AI tools without broadcasting information that they may not want to share. So obviously sensitivity to that. And then also going back to, like I said, of showing your work. you really want to try to avoid kind of the ghost in the machine. You know, AI can certainly do calculations and calculate SPIs for you and KPIs and CPI and all sorts of I's and indicators and things like that. but you want to ensure that any sort of like math or calculating that an AI tool is doing, you can back it up. You know, always check your work with AI. if you're Consolidating a bunch of information into one number, do it manually too, ensure it's it's doing the math correctly. and like I said, when when and not even just for AI, but like when you are using indicators and values. I know one thing we like to use on our dashboards are like tooltips or you know,
MIchael Bograd: Mm-hmm.
Taylor Moore: you can click on that indicator and say, okay, how how am I getting this number? Is this number good or bad? What does it mean? it's not just a number that says. you know, red, green, yellow. It's there there there's some backup there to say this is what makes an indicator green, you know. You're on time, you're you're within two weeks, maybe it's yellow, maybe it's something you want to keep an eye on. But you can really show your work. and I think people that use AI and and don't show their work, you can really kind of lose trust in the team. it's it always kind of goes back to just showing your work, making sure if you are using AI, it's been checked, you can prove to people that
MIchael Bograd: Absolutely. And I guess we we should probably qualify. What does AI mean in a dashboard? Right. You know, you can you can use it a lot of different ways. I think for us, the simplest way is really not even AI. It's probably just large language models, right? And I think we've done an interesting and hopefully a good job of of creating those inside of our dashboards where you can yes, definitely interact with the dashboard, talk with it, ask it questions. what I also think we've done a really nice job of is you know, like you mentioned, there's always a hesitancy to new tools. Where we have kind of pre-scripted prompts inside of AI and they actually build on top of each other. So you add, hey, what's the what's the most critical piece of equipment if I'm looking at a procurement dashboard? And then it gives lists out for you. And then it gives you a follow-up prompt. Hey, not only is it here's your your your issue, but here's your most critical piece of equipment, but here's the vendor. And and you
know, what action should I take? And it kind of leads you to water, I would say. So I think, you know, we've we've kind of have some training wheels on our LLM as opposed to saying, hey, end user, you go figure it out. It's more of a, you know, we'll we'll walk with you along this path of of how to digest your data.
Taylor Moore: Yeah, and I think that's a great example. 'cause like I said, when when it comes to project controls, like we are experts in this field, when you get a bunch of cost or schedule data, we know kind of what to look for, where you might see some some red flags, things that raise questions. and, there's a lot of smart folks out there that are using these tools and maybe don't have that sort of knowledge base. we we are able to teach these models, what questions should I be asking? 'cause if you don't even know what question to ask, you may not be able to identify the the trends or indicators that you you really do want to extract from these dashboards. so setting them up in a way where it kind of prompts you to ask these questions. if I see a bunch of schedule trends where the the end date hasn't moved, so it's like great, we're on time, but you see, total float compressing or intermediate milestones start to slip out and things like that, I know to ask, are we just trying to Are we snowballing a bunch of stuff and squeezing ourselves on the back end? Whereas somebody who's maybe not as privy to to those questions or project controls, scheduling, things like that, maybe just see that, completion date not moving and just be like, no, we're in good shape. But AI can really help bridge that gap of taking our experience, the questions we've asked, the things that we've seen on other projects, and help, like you said, lead them to water, get them to ask those questions. put those answers in front of them. so it can kind of give them more information than they were even, you know, prepared to ask in some instances.
MIchael Bograd: Perfect. do you have any final thoughts on dashboards, Taylor?
Taylor Moore: Yeah, like I said, I mean the biggest thing is really just ensuring you ha your your s your dashboard has a a backbone of processes and procedures where you can point to where you're getting this information and how it's being calculated. I the analogy I always use for folks is, when when America introduced the dollar, right? You can't just give somebody a piece of paper and tell them it's worth a dollar. No, it needs to be backed up. And that backup was a bunch of gold in Fort Knox. So, we may have this dashboard showing you all these pretty numbers and values, but your project controls plan, your project execution plans, these written procedures, tables of where all this data is stored, how we're getting it from various people. That's that's the golden Fort Knox. That's what that's what backs it up so you know that this val these dashboards and the values they're giving you are real and not just, printed money that we say is valuable, right?
MIchael Bograd: That was a good that was a good line right there. Final thoughts on dashboards are essentially, we use these internally, right? You know, we like to use dashboards to support our teams. The way we set up our dashboards at Stellic is essentially we want our our Planners, our schedulers, our cost teams, we want them focused on working with the clients, interacting with the clients, less about running reports, checking boxes. We we would use this data as a tool to kind of reinforce what our team is doing. So any chance we get to put together a dashboard for one of our projects, the goal is essentially how do we reduce workload? I know it sounds odd, but we want our again, our planners and schedulers focused on. working with the project team, helping drive construction, helping drive the project forward as opposed to going, I have a a month end report to that's due. We can have a team that'll generate this directly from the dashboard and kind of help automate and alleviate a lot of those those pain points.
Taylor Moore: Yeah. And one final thought there is, you know, your
MIchael Bograd: Yeah.
Taylor Moore: your dashboard isn't all like the end all be all of your project controls planning. Your dashboard is a tool that you should use to help make informed decisions and really understand where your project is and communicate that, whether it's your communicating up to stakeholders or communicating down to the people in the field doing the work. you know, it's it's really just A a tool to help you better perform on a project, ensure that it's done on time, under budget, and that you're planning accordingly.