How JE Dunn Uses Forge APIs to Improve Efficiency and Manage with Clash Process

PHIL JOHNSON: OK, let's go ahead and get started. Thanks, everybody, for being here. So we are excited to show you some of the integrations that we are doing today. We are from JE Dunn, as well as Autodesk.

So my name is Phil Johnson, Virtual Design and Construction Director, actually out of Kansas City. And when it comes to what we're talking about today, I'm more really the end user. I'm surrounded by these very smart gentlemen that's really done the development side. So I'm coming from the user side.

MARK STOCKS: Mark Stocks at JE Dunn-- I'm the Director Of Application Development and Data. And so I obviously handle all the data integrations, the application integrations.

KEITH WHITE: So I'm Keith White. I'm with Autodesk Consulting. Kind of like Phil, before joining Autodesk, I was a licensed architect. So I approach things from a user perspective, as well. So I've been working with JE Dunn for quite a while. And in the back, we've got Debraj.

DEBRAJ BANERJEE: Debraj-- I'm the Senior Information Architect. I work with data and third party integrations.

MARK STOCKS: So he's the smartest person.

PHIL JOHNSON: Yeah, I was about say, so just because Debraj is sitting in the back of the room, he's literally the smartest person I've ever met. So yeah, Debraj has pretty much architected everything up here that you're about to see.

So JE Dunn-- I don't know how many of you guys know who we are. We're a Kansas City company headquartered in Kansas City. We started in 1923. We're family owned. We got about 3,000 employees right now. And in the past two years, we've grown considerably.

MARK STOCKS: OK, so safety is number one, especially in construction. So safety is always on the forefront of our minds, making sure, you know, we come home safe to our families and friends. So definitely want to talk a little bit of safety today. So real quick if there is-- in case of a fire, we're going to exit right out here at the back door. We'll take a left right out through the double doors. And there's a stairway right there. And follow f exit signs. Great.

OK, does anybody have a safety moment that they'd like to share? I'll just even open it up. I know that's kind of crazy. We got Thanksgiving coming up this week. So anybody doing fried turkeys-- deep fried turkeys? OK, not a lot of fried turkeys. All right, that's OK. Well, if you are, what are some things about fried turkeys that you've got to-- that you got to make sure that you do before you fry them?

AUDIENCE: [INAUDIBLE].

PHIL JOHNSON: It's got to be dry. It's got to be dry. And thawed out.

MARK STOCKS: Exactly.

KEITH WHITE: Yeah, don't burn the house down. Awesome, all right.

MARK STOCKS: So you know, we wanted to show this slide. This is our guiding principles. This is a quote that our founders started. We exist to enrich lives through inspired people and places. This is a mantra that we live by in our company. And it starts at the top and get the pushed down. This is everywhere in our company. It's at every office. It's printed on the walls. It's plastered everywhere. So it really kind of shows how everyone at JE Dunn feels about the company, and what we do as a company, and then how we want to enrich the lives of people outside of JE Dunn.

So we've got about 20 offices right now. And we're actually working in 40 different states right now as we speak. So this just kind of a quick landscape of JE Dunn. And then, I wanted to show, just really quick, some of the projects that we finished in the past year.

So to BOK Plaza-- 27 story-- this is in downtown Oklahoma City. Gulch Crossing-- this is a new up-and-coming district in Nashville, so a high-rise retail mixed use. Populous-- we did the headquarters for Populous. I don't know if you guys know who Populous is, but they've literally designed every football stadium, every final four stadium, every NASCAR event. Literally, every sporting event that you've ever seen, they've designed. It's a pretty awesome company.

And then, we had a local Kansas City shop. This is the Cerner Innovation Center. At the time that it was being built, it was the largest economic development in the state of Missouri. We have two campuses for Cerner, one in Missouri and one in Kansas. This is the Missouri one.

And then, One Light-- this is just a residential building that's actually very close to our headquarters in downtown Kansas City. It's a pretty fancy apartment-- 25 story building. It's pretty awesome. And it was so popular that they had to build Two Light. So Too Light's the exact same-- 25 story. It's pretty close to the spring campus. And because of the popularity of this, they're actually about to start a Three Light and Four Light. So yeah, we're a big surge of residential in downtown Kansas City right now.

PHIL JOHNSON: And then, some of the topics that we're going to talk about today-- this Two Light was actually integrated in using the solutions. So it kind of piloted some of the beginning things that we developed.

KEITH WHITE: I think this will help your downtown too. Because every time I've ever been to Kansas City, Missouri, I'm amazed that no one is walking on the streets. I never see anyone. So that's just fantastic.

MARK STOCKS: Yeah, our city gets pretty quiet after about 5:30. Yeah, the downtown area definitely it's really quiet. So this is our ENR rankings right now as a general contractor. I'm not going to go over all these. But we're top 20 in almost every category right now.

And then, so we're going to talk about our relationship with Autodesk. And we're going to do a quick landscape of our ecosystem. And then, we're going to dive into the class. So our relationship with Autodesk started about six years ago. We have an estimating platform that we wrote in-house. And we wanted to do an integration with Revit with the actual model. So we partnered with Autodesk six years ago to do that. That project was a complete success. And then, after that, we essentially started transitioning into multiple projects.

Because of this, our test team has not changed in the past six years at all, and neither has the JE Dunn team. So we have this long history and this deep knowledge of our integration platforms that we're able to keep building upon without having to relearn anything. The middle quote is by our CIO, which I'm not going to read, because he's not here. So I don't need to read it.

But the last quote is from my daughter. If dad is on the phone when he walks in, dad's talking to Saikat. So Saikat is on our project team and that literally speaks to kind of the depth of our relationship with Autodesk. I am always on the phone with Saikat My family knows Saikat I'm actually going to San Francisco next week to go see Saikat. So it's not just Autodesk and JE Dunn. We're an actual team.

PHIL JOHNSON: Awesome, all right, so our agenda today-- what we're going to talk about is really diving into a process that is pretty well defined. This process happens all over the world. And it's something we do every day. But we're looking to evolve it and improve it. And we want to show you some of the integrations that we've done. We're going talk a little bit of back of house, so some of the integrations that were key for success. We'll talk about our data flow and our process.

We had many gaps to fill. And Autodesk helped us fill those in to really complete the picture, but a lot of gaps that we had to overcome. Clash management side is our new approach, and it integrated into our data management process and BIM Coordination process. So we'll talk about some of that, how we're tracking that data. And then, looking to the future, how we're looking to adapt, and evolve, and improve.

OK, so there's been a lot of effort and a lot of strategy to think about how to connect all the data on our projects. And so that everything flows really well. And so there's a lot of thought that's gone behind that. So from project administration to cost estimating, during pre-construction, are solutions for our field, as well as documentation.

But what today we're going to talk about is our cache management tool. It's our new workflow for managing that process and the data that's behind it. All right, so looking at the workflow and how to improve it, the [INAUDIBLE] specialists that we have all over the nation are working on upwards of 100 projects per year. And they're leading these efforts. And they're going with our operations teams, and we're engaging with our owner design team and our trade partners to really evolve into a process where we're weeding out issues. We're finding those identifying issues early on and weeding those out.

And so we'll go through a cycle. And weeding out those issues is beneficial for the project and enables our trades to be able to work through pre-fabrication and be able to use [INAUDIBLE] in the field and that sort of thing. So those things are not new. But that is a process that we go through many times, and it's well-defined.

So [? our ?] main projects actually stay on schedule and within budget, but there are some that are very problematic to where we have a lot of issues to work through. And it slows down the process. So coordination really moves at a snail's pace. And so a lot of time is spent from all parties. So budgets are busted. We're behind schedule, and coordination becomes critical path.

And so at the end, we're left wondering, how did we get here? And so with all the data that's available to us, we're still asking ourselves, how do we get here? So with all the data that's available, behind the process that we do every day, we should be able to utilize that. And so there has been attempts to track that information along the way. So VDC specialists will actually input this data that's clash related into spreadsheets, Google documents, all that information that's a very manual, arduous, time consuming process. And it's very siloed, as well.

So really no one ever sees this information. How many clashes are we working through? How arduous is it? Until it's really sent into an email. So what we really needed was to want to automate the collection of that information, but also tie it into our existing ecosystem. And Mark Stocks is actually going to talk a little bit about that ecosystem in the data behind it.

MARK STOCKS: [INAUDIBLE]. So this is a real high-level overview of the way we architected our ecosystem. The line at the top is obviously all of our cloud providers that we have integrated with. We have Autodesk, Zendesk. Right now, we actually have a couple more data integrations with LMS, with employee data.

But everything that lives in the cloud, we run everything through an API layer. And this is one of the layers that Debraj actually executed, which that API layer-- what it does is it handles all of our business rules. So anytime that we have data that's flowing into our system that has to have events, or triggers, or anything else that have to happen, because of a object coming in that happens in that blue layer.

And those business rules then get applied to the data and actually either create more events. They create workflows. They create secondary API calls-- whatever. But that business layer is really where that is where the intelligence layer happens. And then from there, we have in the green layer is all of our data, our internal applications. And so really, you have API feeds going both ways outside of that blue layer that's pushing data down into our applications. And it's sucking data down from the cloud environment.

KEITH WHITE: Let me add something here real quick if you would. This is kind of a scary slide, especially if you're thinking about starting something like this yourself. There's a lot of stuff up there. But this was started really more than six years ago. Right? And it started with pieces in parts of this, and then over those years, we keep extending and adding to this architecture. So all of this wasn't in one go. So it's an evolution.

MARK STOCKS: And to be fair, we deleted 3/4 of the boxes on this slide to slim it down for this presentation. So OK, this is a pretty funny story. So we have an internal business application that actually creates all our projects, creates everything that we do at Dunn.

And so we, obviously, have to have a project before we can even have any clash data. And so we created this application that was going to facilitate creating projects in our ERP, Active Directory, Outlook, our clache management system, projects and Autodesk. But we wanted one location where you could type project name in, get a simple workflow, hit Submit, and it just pushes out to all of these separate environments.

And it was actually called the job startup application. And then, someone realized, well, we can't name it that because the JSA is the safety form. And we can't call it a JSA. And so as a joke, I said, why don't we call it TPS-- total project solution-- so we can have cover letters.

And my boss at the time had never seen office space, so everyone else in the room chuckled. He didn't, and he was like, hey, we should totally do. That and then after lunch, we showed him the video of [INAUDIBLE] talking to Peter about the cover report. And he laughed and was like, yeah sure, let's call it TPS. So we actually have a legitimate TPS application.

But what's cool about this is that all you do is you fill out a form. You fill out some information. It's got a simple workflow. It's got to go to the project exec. He just approves the screen, says, yep, I'm good. You can fill out about five small pages. Some pages have three questions. Some pages have 10. You hit summit, and it pushes out into our ERP, creates the project in a clean manner-- so our data is 100% clean-- creates the project and Autodesk, creates an Active Directory group, creates an outlook mailbox, creates a done dashboard website. Everything that you need to run that project within JE Dunn gets created by a click of a button.

And then, once that data is created, because everything is in a database. And it's all event based. Then, we just called triggers to create APIs, and that's how we actually create the Autodesk project.

And if we have a brand new company, that also gets created, so it's always there. All of our data that we need to build a project is populated before the end users need that data. So we want to make sure that data is there when they want to use it.

And then from that, again, everything's API based with us. So after our data is created and forged, and we have the project in our tenant, we have the Glue project, we use APIs to pull that data back down to on prem, so that we can do predictive analytics and better reporting. And we can use that data to provide more metrics for our projects teams to make better decisions.

So this is some of our stats right now in our HQ tenant. We have 3,000 active projects in our tenant. We actually have more active projects than that. We really archive everything that we're not using to keep our data set clean. We get about 10,000 users that we're syncing into Autodesk. We got 85,000 external users in our system. We're a company of 3,000 employees. And we have 85,000 external users actively logging into our system.

Now, 10,000 are in Autodesk. And we got 49,000 companies synced between both of our systems. And then, right now, we have five custom integrations. And the reason why we have five custom integrations is we like to track the usage of every integration across the board. And so we segregate every integration that we do into its own silo. So we can look at metrics, usage, analytics, and identify how that application is actually getting used.

So this is a screen of TPS. Talked about it before. And really, all you've got to do is click those blue buttons. You click a blue button, it creates the Autodesk project, creates a blue project, creates whatever you want.

But the benefit of this is because of the work that Debraj has done through our business layer. When you click a Glue project, and your project is in South Central, in the Oklahoma City office, or in the Atlanta office, we might have different rules that we want to have behind the scenes. So we want these three users to go to the Oklahoma City office, and these three users are going to be admins at the Atlanta office. So all of our business rules happen behind the scenes so that we don't have to write any real custom code to push out to Autodesk. But everything is seamless. So user creation, projects selection-- everything always happens through one interface. So our users don't ever have to learn anything. That UI is identical across the board for them. And then, we just handle everything behind the scenes through API calls.

So again, kind of a quick overview of the landscape. You know, we aggregate all of our-- we aggregate our P6 data, our ERP data, operational data, Autodesk data, and even our model data. And everything gets funneled in through APIs into a centralized database that we're then able to really start connecting the dots and start providing more value for our teams.

So translating LMV, connecting LMV to our estimate, connecting clash data to our trade partners, and then having our trade partner backlog information, the financial information-- getting all of that data connected is what's really providing the power for our project teams.

MARK STOCKS: One quick note, they've been using Forge for a long time, right? They were one of our very early partners. LMV stands for a large model viewer, which is now the Forge viewer, in case you're curious what LMV is, that's how long they've been using it. We used it under a different name.

PHIL JOHNSON: Yeah, I've never not known it as LMV.

MARK STOCKS: Yeah, right? So just FYI, in case you were wondering what that was.

PHIL JOHNSON: OK, so an area that we really need to focus on was if we were to collect this data and automate the collection of that data, how does it impact our workflow, the workflow that we have today? And there's a lot of individuals that come into that that play that out. And so we really had to map it out and look at it. It was a careful balance of automation versus decision-making.

So if we were to lean towards automation, then we're taking the way the decision-making of our VDC specialists, which is really the heart of the coordination process. So if we were leaning really more towards decision-making, then we take the automation away and we're creating manual tasks for people.

So this is where Keith and his team really came in and helped us out.

KEITH WHITE: So as Phil said, there's a balance between decision-making and automation. And if you move toward automation, it becomes rules, right? And rules apply to everything. So that's the balance that they had to decide upon what's the best path forward for them for them and their company.

So we came in, and originally we started in this vertical pink box, over here on the side of this peach color, where we worked with Debraj in the background to harvest data from our cloud solutions into the Done dashboard so that they can visualize and analyze the data and trend it over time.

Here lately, the last integration I worked with was in the other two smaller dash boxes. And that's where we automated the processes within NavisWorks and within BIM 360 Glue.

So as we move toward a more automated process, the goal for us was to reduce picks, and clicks, and other stuff. So we were working with the NavisWorks API, the BIm 360 Glue API. And you'll see something in just a few moments that follows that kind of illustrates what that is.

So that's where our focus was as a partner. We came in and we either coached them and mentored them in using the code in their own system or we wrote add-ins that did work for them by user clicking some buttons.

So this is an example. You'll see an animation in the lower left. But the goals of the add-in-- I said earlier, right-- let's reduce user picks and clicks. Outside of decision-making and rules, let's see if we can reduce the picks and clicks.

And then we also needed to automate the processes that were between NavisWorks, BIM 360 Glue, the four JPIs, and the Done dashboard. So as the animation is playing on the left, you're going see this dialog is really where my work was. And this is where we're starting to make use of BIM 360 Glu, NavisWorks, and four JPIs.

So on the right, you're going to see a couple of number boxes. And when the dollar box opens, there's some pre-work done, which has kind of been number one up above, where we harvest some of the BIM 360 Glue View information, we do some work with the Glue APIs on the back-end.

And then when the user is pushing, interacting with the box, it's pushing views to Glue. So if you've ever tried to push a view to Glue, you do it one at a time. So this is a sample project. But on an average project that has real issues, they may have hundreds or thousands. So click a button 100 or 1,000 times. And how much time does that take?

So that was a part of our goal, was to reduce that time for users picking and clicking. So within this dialogue, they pick up to however many n number of views. And we can push them to view all at the same time. We can also remove views from Glue at the same time as well.

In just a second we'll get to a question. Just a second.

And then after that, when they're ready, when all of this work with Glue is done, they push to Forge. So that's what the numbered boxes are below. So once we've interacted with NavisWorks and with the Glue and the NavisWorks APIs, we then save the file, right? Because we're working from a file that was open from Glue.

So we open the file from Glue. We then need to save that. Because any changes made to the Glue views need to be pushed and made sure they're synced back to the cloud.

After that's done. We save a local MWD of the models. So we basically do a Save As to MWD. And then we push it to a console application that runs on the background that uploads to Forge.

So it's returning to the user the ability to keep working in NavisWorks while some background processes are done. So there's a lot more code kind of in the background in each one of these five steps, that are little sub-processes. But it's not that difficult. We just consume the a public APIs that are out there on the web, that are part of the SDKs. And we just automated the process based off of the workflow that we discussed and we developed together.

So they were happy with the results of this. And it saves-- you know, I don't know, we don't know what the metric is yet, that I know of-- but hours on every single project just through an automation process like this. You had a question?

AUDIENCE: Yeah, just real quick, are you guys using BIM 360 Classic or the--

KEITH WHITE: OK, so we're repeating the question because it's being recorded. So the question was are we using BIM 360 Glue Classic or Next Gen. Currently, we're using BIM 360 Glue Classic.

APIs are a little behind the UI in the Next Gen field in Glue products. So this is making use of classic BIM 360 Glue.

OK? OK.

OK, so this was really exciting for us and really a game-changer. So the data that's being tracked and provided behind the scenes, is now visible into really our single source of truth. Our Dunn dashboard. So where our operations and our project teams are logging into every day for plans, and schedules, and submittals, also is our clash information.

So the issues that we're working through and getting solutions to and assigning, is coming up through our Dunn dashboard. So what's great, is that at any point, any time, anywhere, someone can log in and have access to this information and see it and also interact with it. So our trade partners can come in here and interact with the information, so they are a key player in the cycle of this information. So they can view it and then they can also resolve it right from there.

All right. So this is a screenshot of our clash management tool. And again, Debraj is the brains behind this one, for sure. And this is all the data that's coming in, and flowing, and creating graphical information for us. And so we're able to see the active number of projects that are actively providing data to our class management software, and then also, the number of issues that are being created or generated and assigned, number of issues that are being resolved. So we're seeing progress over time and we're starting to see trends, and that those trends are telling us some interesting information.

A lot of log information that's happening behind the scenes, too. So there's troubleshooting along the way. So hey, why does it look this way or why did this happen? We can always go back to the history of that information, understand why. OK, so the information we thought we were pretty good. Most projects we were thought we were pretty efficient, but what the data is telling us is that it wasn't uncommon for issues to take upwards of three weeks to resolve. And that just gets passed through and really isn't seen or visual.

And so now we have a good visual of this, and it's actually taken us longer than we thought. And so we're able to use these visuals as a communication tool to improve the process. So something interesting about this graph on the upper left is that our active number of items that are coming through, those are going up and down, they're spiking at times, but overall, our resolve, which is that gray, is getting closer and closer to our assigned items. And that's what we want to see. We want to see our team members actively pursuing those issues and getting them resolved.

In the beginning, we were pretty terrible, right. So the process was just beginning, we were just getting off our feet, and then as time goes by and we're coaching our teams to use this, the data is getting better. And as long as that gap between our active items and our resolved is narrowing, that's a good thing.

OK so here's some of our data. This is specific to BIM 360 Glue. So we have 300,000, or I'm sorry, 3,000 active projects. 1,700 of those actually have a model that's associated to it. OK, so that's not just all coordination projects, but that could be field or other solutions that are there using the model. So for quality and safety as well. 122 active projects that are consistently providing information to Forge at Dunnpedia. 29,000 business partners are in our system, and that's growing every day. With 361 of those are actively being assigned issues on our projects nationwide.

All right, so this is some of the data that we can take advantage of here. And so we can start to take the data, really break it down, right? So data is just data, initially. Until you can organize it and use it for a purpose. And so now we can look at it per project there on the left, per area, or level, in the middle. And then portray partner there on the end. We can see how many items are being assigned, how many things are being resolved, by a day-by-day basis.

All this information is being synced every night. So every morning we come in, it's the fresh updated information.

MARK STOCKS: This is-- you're displaying this through Tableau, I think, is that right--

KEITH WHITE: So yeah, we actually, because of the way Debraj built our data layer, we can display the data in multiple systems. So the bottom ones are we have Tableau dashboards that can be consumed by everybody in the company. The top one is some of the clash management tools that really all the VDC guys, it's the tool that they use to manage the data. And then we also have another data layer where we can pull into the Dunn dashboard to show every trade partner that's on a project, so there's multiple places that were reporting off of this data with multiple tools, but again, back to the architecture, it's a single source. We have one source of data and we report off of that source of data.

PHIL JOHNSON: Awesome. So when you take this data and you put it over time, we can start seeing performance trends and that's huge. Instead of waiting till the end of the project to react or to wonder why, what happened, right at the beginning we can see how long it's taking to get these issues resolved, which really tells the health of the project.

So I left off the trade partner names there on the left, but you basically you can see the number of issues that they're working through and the average days to get those resolved. And so we've been able to pull that down and coach our teams, and we've gotten much better results through the year, really two years back and since we started this thing, and we're getting closer down to averaging a week to get things resolved. So that's a positive trend.

OK. So we select our trade partners based on their performance, right? So safety is huge. So the EMR, recordables, is financial information, that's all important, but just as important is their ability to perform in the preconstruction process, or in the preplanning more so process. So they may have a great field install team and a great fabrication shop, great. But if we can't perform early on in our preplanning efforts, then the field install is never going to be as good, or efficient, or as productive as it could be. So that's very important and this data is being used in our selection process.

OK looking to the future. Some of the things that we're looking to improve and adapt, especially with our design partners and our trade partners with this system, so I'll speak a little bit to that, as well as there are some new visuals and better communicate to our project team so that they can react sooner. OK. So you've probably seen the MacLeamy curve, which really highlights the positive impact of getting started earlier to the project at a lower cost, but higher impact.

So that's what we're trying to do, especially if we have the opportunity to get involved between 65%, 95% [? CDs ?] on a non design build project and model information is available we're going to certainly plug those together and look for issues and do reviews. We already do plan reviews of our documents. Why are we not reviewing our models? So we are doing that. We're finding constructability issues early on, communicating that back to the design team.

So when we can get those things identified and resolved early, then that's going to be a better impact to the project, a much lower cost, we're eliminating our [? afis, ?] change documentation, et cetera. And this may be right before we bring our own trade partners, or if or if they're on-board, then we'll certainly involve them in that process as well.

AUDIENCE: When you say you are actively using data to hire trade partners, are you looking at the installation trade partners, or the design trade partners?

PHIL JOHNSON: I would say, yeah. Installation trade partners. Yeah, thank you. Yeah, so the question was, when we're using this information during the bidding and selection process, it's with our install trade partners. Thank you. What's that? Internet! Excellent. OK.

All right, so some other trends that we're seeing that positively impacts the projects is that Autodesk has done an amazing job of democratizing the use of model information. Anyone can jump in there, and participate, and review the information that's on the project and start weeding out some of these issues or constructability issues, so do we all do we all do that? Maybe not. But it's available to us. And so I think the general trend over the years is that we'll have more active people participating and better designs coming out through the end. So that's the goal, is to democratize it and get those issues resolved prior to construction.

All right. So this is our project Dunn Health Dashboard, and this is something that we're integrating into all of our projects. This is much better communication tool than what we have even months ago. So this is fairly new and so we'll be integrated in, but it really gives a good look to the health of the project itself.

How is coordination going? How long are we taking in our durations to finish out an area? And we can take this and compare it to our overall construction schedule. How that compares and how efficient we are. And then how many days is it taking? What's our average of days to get these things resolved? And so it's color-coded based on that information as well. So a very quick look. Our operations teams they don't have to go to enormous graft to try to figure this out. This tells them right away. And then instead of waiting to lead the project, we can start making better decisions and having those conversations early on.

KEITH WHITE: So this is the next version of the Dunn dashboard, where we're on SharePoint 2010 right now and we skipped 2013 because there wasn't a big feature-- there wasn't a feature jump for us. And we are actually going SharePoint Online now for the next iteration. And so these are all SharePoint effects, react, web parts using API calls behind the scenes because we love API as a JE Dunn.

And so all of our, from our P6 data, our ERP data, our clash management data, all of that and we delivered a very seamless unified and very similar experience for our end users so it would be less for them to learn. So someone who is not familiar with clash, or P6, or anything like that, can easily come into one of these dashboards and be able to identify and look at the data and understand the data.

PHIL JOHNSON: Excellent. OK. I think that's all we had, so now we'll open up to questions. Any questions?

AUDIENCE: What are you guys doing on the [INAUDIBLE] between CMiC and BIM 360 [INAUDIBLE]?

KEITH WHITE: OK. [LAUGHTER] Yeah, so BIM 360 Field, how do we-- if I want to understand your question correctly, how are we integrating the two, or?

AUDIENCE: [INAUDIBLE]?

KEITH WHITE: So the way we've taken our stance on BIM 360 Field, and this is company-wide. We had the option of going down the CMiC road or going down the BIM 360 Field. We chose BIM 360 Field. And we do a nightly data harvest of our BIM 360 Field data, we pull that data in every night, and then we blend that data in reporting with the financial data. So that's how we do our hierarchy of reporting.

So issues and checklists and commissioning does not happen in CMiC see right now. It all happens in BIM 360 Field. Did I answer your question?

AUDIENCE: It does. But that later translates to [INAUDIBLE]?

KEITH WHITE: Yes 100% cleanly. That's a weird noise. So everything that we do is based off of our job number. And that whole TPS thing I talked about at the beginning, that is the start for us. You can't do anything without that job number, nor can we integrate any application without that job number existing in that application. So in our HQ tenant, we have our job number in there, and then we say that project GUID into a system on our side so we can do data mapping across the board in any application and that's how we do that--

MARK STOCKS: Can I add something [INAUDIBLE]? So, as a consultant, right, I not only work with JE Dunn, but I work with a lot of customers. And they started very early and so their data is really pretty clean, and they did some data clean up themselves. But new customers, the data is all messed up, right? There's just dirty data. Their TPS system allows them to enter the data once, and so they can trust thereafter, project names, project numbers, all that data matches, so that whenever they harvest that data down and they push it back into CMiC, or they push it wherever it needs to be pushed, the data is going to match up. Project names, project numbers.

Just even in the spelling of the word incorporated. With the fully spelled out, or with a period, or however, right? All of that makes a difference when you're trying to map your data together. So them building their TPS up front was critical to this now and the benefits that they're reaping. So data and how you create it, if one person is typing in multiple forms for the same project, something's going to be off. And it we're just humans. It's going to happen. So that's one of the biggest benefits of TPS. Let's get to mic so that we can--

AUDIENCE: Just had a question about your TPS program, him down with those in the past. And I guess I'm curious more on the internal infrastructure workflow. Of all those platforms starting at the same time at one point, we get schedules that start in our scheduling department during the preceding days carry down the operations. But they don't have a project management site set up at the same time until we're actually awarded a contract, and all that. So I was just curious, if you could share how you, I guess, manage that and how all that starts all at the same time.

KEITH WHITE: So when we built TPS, we built it with the concept that the opportunity phase is already done and there wasn't any preconstruction going on yet. And I was like, you've already been awarded the job, it's already been contracted, and you need a job number to start charging time. About four years ago, we realized that that was, we needed to start earlier. We needed to start at the actual opportunity phase, and so we actually did an integration with Cocentral, which manages our opportunities.

And so now, we have a simple form so that if you're in a bar and you meet an owner and you're like, you want to build a $150 million building, you just fill out a form you hit Submit. Once you hit Submit, within an hour, the project's created in our system, the opportunity's created in our system, and then we can start essentially linking that. So we had to actually build what you're talking about, your question. We had to build that layer ahead of the time when we started the job.

And so now, we're at the point now with the new Dunn dashboard that we had to change our whole philosophy off of job number and we're going to GUID for everything to track everything because the GUID then ties into the opportunity number, it ties into the CRM record, and it ties into the job code. But again, our data is clean so we're able to really quickly be nimble about this.

So, yeah. Your question is valid. I mean we ran into that issue four years ago and we essentially had to write code to accommodate to allow for starting the opportunity, and then quickly start piecemealing the applications on there. So getting a Dunn dashboard earlier, stuff like that. As far as the architecture behind the scenes, it's a lot of database triggers, a lot of APIs, and to be fair, there's a lot of PowerShell. We do a lot of PowerShell for automation of tasks. So connecting to Office 365, connecting to SharePoint, connecting to any of those different objects, PowerShell is the event receiver behind-the-scenes for that.

AUDIENCE: [INAUDIBLE]?

KEITH WHITE: Pardon?

AUDIENCE: There's some projects we don't [INAUDIBLE]?

KEITH WHITE: Oh. Yeah.

MARK STOCKS: Yeah.

PHIL JOHNSON: Right.

KEITH WHITE: Yeah.

MARK STOCKS: So they just fade away.

AUDIENCE: [INAUDIBLE] platforms kicks off your project manager [INAUDIBLE].

KEITH WHITE: Yeah.

AUDIENCE: [INAUDIBLE].

KEITH WHITE: No it gets closed. It gets closed in our monthly, we have a monthly financial meeting, which is great that IT is actually a part of that. And so when-- Yeah, so that we close the projects, I get a list, I get an Excel file and I literally, once I get the Excel file--

MARK STOCKS: It's next door. It's a cart rolling down.

KEITH WHITE: It's the weirdest noise ever. I get that Excel file and I clean up everything. And everything gets deleted, flagged, it's all soft release, but everything it's deleted, flagged, cleaned up, and just pushed off to the system. And an our CEO and our CFO, they know what clean data does for the company. And so we're very-- even they are very diligent about ensuring that even at the financial standpoint, at the project management standpoint, at the site cleanup standpoint, that the data is clean. So we all work very hard together to make sure our data is clean.

MARK STOCKS: I'll add to this too. That everything, like all the add-ins that I was talking about that I wrote earlier, the code that we write, the code that Debraj has written, all the stuff is logging each one of the actions into their system. So you saw a bunch of trends and a bunch of data points, right, for the data that's relative to their construction process.

We're doing the same thing with our add-ins and with our code. Every event is being logged, so if something goes wrong, they can see where the problem was and it can be resolved. It can be re-triggered, right? But they're also seeing they can trend, these things are working, right? Over time as well. So that's another important part that we really didn't discuss, but that helps them trust those integrations that are being developed, are actually indeed working. They could see that as well. All right, well him first, yeah.

AUDIENCE: I think it's great that you guys have been able come on come across to these solutions, but I don't want to sound too harsh, but with this partnership, Autodesk is acutely aware of some of the deficiencies of BIM 360. And I'm wondering why it's taken so long to fix simple things like shared views not even having the ability to do markups or not be able to push shared views, save views, and then you have markups.