Video: From Product Definition to Production Instructions | Duration: 2032s | Summary: From Product Definition to Production Instructions | Chapters: Welcome and Introduction (31.39s), PLM Data Fundamentals (217.28s), The Execution Gap (309.015s), Data Centralization Evolution (492.11s), Digital Work Instructions (619.825s), Automated Work Instructions (847.885s), Digital Work Instructions (1082.72s), Results and Future (1554.735s), AI Future Enhancements (1733.98s), AI Future Outlook (1823.81s), Q&A and Closing (1926.695s)
Transcript for "From Product Definition to Production Instructions": Alright. Hello, everyone. It's right on the hour here, so, some people may still be trickling in. But wanna thank everyone for joining today. Get to talk a little bit about, digital work instructions and and some of the work we've been doing with our partner, SSI. It's been, this is a presentation I gave down at, their world shipbuilding conference. Had a great turnout there and a lot of interest in it. So, doing a little bit of recycling, but for those that weren't able to join the conference down there, giving you guys an opportunity here to to see what we had to say, and and get to experience it for yourself. So, the topic today is, all about going from product product definition to production instructions. It'll highlight a lot of the work we did, with SSI, but we'll get into that as well. So, if you guys have questions along the way, I do have the q and a pulled up here in the background for me so I can see them. If there's something I can address while I'm talking, I'll certainly try to do that, but at the very least, we'll address all those questions, at the end. So I guess to hop forward here, a little bit about me. I work for Canvas Envision. I'm a solutions engineer for our team. I get to work with a lot of our customers during the early part of the sales process, to help determine, you know, scope of the project, and and really learn kind of how companies work today and how we might fit into their system, into their technology stack. So do a lot of work around that in the early sales process. I also get to work with a lot of our partners, and lead up our integration. So like I mentioned, partnered with SSI here, and work on an integration with Shipbuilding PLM. We've been working with them for the last... Almost four years now. So we've had a lot of work done there, gotten to have some people in the... In our tool, in our solution together, and give feedback and kind of drive us towards one that's tailored, more towards shipbuilding. So for today's presentation, it is gonna be tailored towards shipbuilding. However, you'll find a lot of overlap, into the early industry, into any industry. Turns out there's a lot of overlap there between manufacturing as a whole, not just in shipbuilding. I'm sure you'll encounter a lot of the same problems that I'm highlighting here today. So, this is tailored towards that a little bit, but by all means, it's it's probably gonna hit close to home, for those joining. So in today's presentation, I'm gonna go through this, kinda going from, the product definition, and and kind of everything that we know, today and then talk about where the gap and execution actually comes, and then get more into how we're solving that together, and putting our our data to work for us. So, getting into the product definition, we'll call this, what engineering and PLM already know. And what do we already know? Well, you'll see the ship here. And we know a lot about this, especially early and upfront, in the design cycle. With PLM, you get a lot of information. You get the, you know, three d geometry of everything. So what the, you know, what the product is, and what it should look like. Alongside this, you'll get things like engineering bills of materials, your manufacturing bills of materials, the assembly sequence. So this can kind of be derived from that MBOM engineering BOM type stuff. You'll get this information. Materials, you know, all of the product information itself, not just the assembly as a whole, but all of the individual components and the materials that make them up. Configuration data, you know, different SKUs of the product you're making, products change over time. Metadata, there's tons and tons of stuff in here, that shipbuilding PLM is able to contain, and then, of course, revision control. And so we know all of this stuff. We've got all this information, and luckily for some of us, we've got a PLM to store it. This is where PLM really shines is is making sense of this data, and storing it in a in a structured, centralized way. And all of this data tells us a lot of what. We've got, you know, how the product should physically look, which parts should be used, what order they should be put together in, all of the what what what. And so we have all of this data. What what really is the issue? Right? We've... We did our due diligence up in the design phase. We were very detailed in, you know, creating metadata and describing and creating and designing the product that we want to make. So what could possibly be the problem? This is where the envision side of it comes in a little bit more and where we deal. It's it's the execution gap that we really see. So I'm trying to take that product definition, and turn it into, you know, production information. And so I mentioned PLM contains all of this helpful data for us, and we're we're talking about what's the problem. We have all the data. Well, it turns out that the the data itself does not physically make the product. Right? These are our production teams. These are our operators out on the floor. These are our coworkers, our customers, you know, physical people actually doing the job, putting the product together, out on the floor. So we see a big gap here. And this isn't anyone's fault. It's it's the way things have been forever. But we still have to overcome this gap. Excuse me. So what we see to overcome this, we see people taking this data from PLM, this, you know, CAD data and and sequence information, and maybe they're reformatting it into something that's ready for production to, you know, to ingest. Maybe they're recreating this, you know, to to get data into a more, accessible, maybe easy to read, maybe more tailored towards what the production team actually needs. And, of course, searching for it and updating it. You know, these designs are are ever changing, and, you know, luckily, when we're searching, we're searching into one location here, but updates are are gonna happen, you know, no matter what. And what does this look like for for, you know, for our customers, for those out there, in manufacturing? This is a stat across manufacturing as a whole. There's a couple different ones here that I pulled together, but the one here on the left is, you know, 30% of employees spend time every single day looking for or requesting information. And we've done a lot of work upfront. We did all that design. We got it in PLM. We got it in a centralized location, so we're helping with that. But that's still, you know, potentially a third of your workforce every day looking for information. I always say the statistic, you know, five minutes a day, which doesn't sound like much, I'm sure we all spend that, is 1% of your workday. So every five minutes you spend looking for something, you know, it adds up quickly. It's taking up, time and and energy that could be towards actually making the product or improving the design or, you know, physically putting it together and and getting it out the door. And so, there's a huge inefficiency around this, in the industry. So what do we do about it? Well, we've got our data. I think it's time for us to, put that data to work. And so when we look at kind of the state of the industry of where it's been, it looks a little something like this. We've got this massive data, you know, stored throughout here on the right side. But you also have, you know, each of these individuals searching for that information. We talked about that inefficiency, how much time they're doing it. And they may be searching in different locations, getting different results. And once they have that information, they're making an interpretation of it, and then they're going and doing that work. And people doing that work based on different assumptions and different data that they find ultimately ends up in different work being done. It could be different quality of work, different... You know, maybe it's the wrong work being done in general. But it's hard to have consistency across your your enterprise when, you know, people are having to do all this interpreting and searching and then doing the work. It it leads to errors. Oh, so what do we do about it? Well, we talked about, you know, the PLM. This takes that massive data on the right hand side and pulls it into a centralized location. And this helps immensely. This helps reduce that number of, you know, 30% of of people looking for information because they're looking in a centralized location. But you still have this problem of people searching for the data and interpreting it and then doing the work off of it. And so what we've done as an industry, I'm sure you're all familiar with it, is to go to what we call SOPs or or work instructions. And, really, that comes down to what I would consider a smaller subset of our workforce is taking, is searching for that data and then turn it into these SOPs and DWIs, so that, you know, these people actually doing the work are working off the same source of truth. You know, they're working off the same SOP, the same DWI. They're getting the same instructions as everyone else out on the floor. And this has greatly improved, you know, the quality of work going out the door. So what makes a good work instruction? Well, we envision... We had envisioned... Think it looks like an envision document. But there's a couple specifics on it. So, a good work instruction to us is something that is specific. It is, you know, everything that you need at the time of execution and nothing that you don't. You know? It is not something that is generic and kind of just covers the process as a whole. Maybe it's for a different SKU or a parent, you know, a a group of products. We believe that it is specific to the job that you're doing at the time, and it gives you everything that you need, to do that job well and correct the first time. We also think a good work instruction in today's workforce is is digital. We've got a lot of people coming in, you know, a younger workforce coming in that, doesn't consume or understand or learn, data the same way that older generations have. You know, a lot of times, instructions live as PDF manuals or binders. You know, a lot of times it's printed out and delivered to the floor, you know, in a in a hard copy. And so we think it's digital. It's more tailored towards that changing, younger workforce coming in. But being digital also means that we get the benefit of a couple other things that you see here. It's connected is number one. So we talk about this being connected, back to the source of truth data that is the PLM. All this information that we're putting in the work instruction is is coming from somewhere, and we know where that somewhere is. And so as the work instructions change, I see there that, you know, they're living. You know, these instructions are are changing, and they're keeping up to date. You know, this isn't a file stored on a local desktop or something printed out, that quickly gets out of date. This is always up to date and current, for the job needing done at the at the time. And, of course, interactive. It's it's more than just going from, you know, the paper to a PDF. It's still a static PDF on your screen. Once we move digital, we can actually go beyond what is capable of just the paper. We get to leverage these digital native systems. And so you get things like interactivity. You can interact with the three d model and and answer your own questions. You can watch videos on specific processes. Maybe it's a a weld procedure or an electrical routing or a hose routing or something, something that doesn't show up in the three d model. And so all of this today makes what we consider a proper work instruction, for the modern times. And we've seen a lot of benefit of this in the industry. I think, you know, there's a couple statistics up here, but if you look at one, it is this one right here in the middle. 84% decrease in the total number of errors. It is, astonishing the number of errors that are attributed to poor instructions or ambiguity in the instructions leading to interpretation. We talked about people interpreting data. If your work instruction still isn't clear, oftentimes people are still interpreting that, in their own way. And so all of this can lead to, you know, a greater, increase or greater decrease in the number of errors, by going digital. And at the end of the day, this is this is your company's p and l. This is the the quality of the product that's going out your door. This is your customer satisfaction with the product in their hand. It is your your your worker, you know, your operators out on the floor. It's their job satisfaction. All of this kind of comes back to, you know, the work being done, and anything we can do to make that work being being done be better, easier, really kind of trickles down, all the way to the end customer. And so what's left? Well, I talked about how we've made strides as an industry across manufacturing as a whole, and in shipbuilding to, you know, consolidate this to a single, you know, or a small group of people designing and creating these work instructions. But if you remember back to the problem that I said, the problem actually lies more in, you know, the searching for, and and reformatting and and updating the documentation into a way that, you know, is ready for production to use. So we still actually have this gap here. And this is what we in SSI and Shipbuilding PLM have been working on, to tackle. And we're doing that by making our systems that we have in place today work for us. So this is, you know, essentially, instead of working with the data, we're making the data work for us, and using the systems, and the technologies that we have in place, to create these work instructions for us. So we are, you know, using scripted automation to take all of this valuable information from PLM and convert it or reformat it or compile it in a predictable way, that we know that we want for, you know, the people actually doing the workout on the floor. And all of this is leading to, eliminating that gap, you know, this this time sync of people, actually spending the time, kind of recreating the work. They created all the data upfront, and then they were just spending time reformatting it. It was very inefficient when these people could be, you know, spending time making making the design better rather than focusing on, all the documentation. And so now I wanna talk a little bit about how we're going about that, kind of what that scripting looks like, what it does, and what the output looks like. So, just conceptually, as we look at it here, we have all that information again on the left side, and we are turning this this what this... You know, what the product looks like, what order it is put together, you know, what revision this is, and we're turning it into the how. So you'll see this, three d representation is actually being delivered out to the front line to an interactive three d, to give them the understanding of of what the product looks like. The the bills and materials are being brought over. They're being connected in the document to add clarity, to which parts are which, and for them to understand, you know, what should be done on each step. Part material info, this is being brought in and delivered, so they can use it for understanding, how parts should be put together. Maybe it's weld information, and things like that. We're also bringing in, you know, all of this metadata of stuff like, kit list and ship list, hole and block and workstation information. This is all being delivered, into the work instruction, for them to know, you know, where things are going... Coming from, where they're going to, making sure that they're, you know, working on the right section of the ship that they're working in. Synchronized revisioning, we talked about this. We're making sure that it is complete and up to date, the work instruction that's being delivered. We're also taking this assembly sequence, that typically just lives, you know, in a in a data point. I'll show that a little bit. I'll show some shipbuilding PLM, but we're turning that into something that's actually visualized. We're taking that three d model and highlighting the parts that should be assembled on that specific sequence, to add clarity. And then the last part is the configuration info. This is a number of different things, but I touched on it. You know, work instructions are specific to the actual work that needs to be done. And so we have, this, you know, very, very specific, configuration. This isn't a generic thing. This was built exactly for the job that needs done. So I am gonna switch over just here to highlight some of that. So we will stop sharing here, and we will switch to, my... Let's see here. This one. Then I'm gonna go down here. So here is shipbuilding PLM. I'll zoom in here just a little bit for us to look at, but this is where all of that information is looking at... Looked... Or that we are talking about having. This is where it is stored. This is what it looks like. And as you're looking at this, you can kind of start to understand why this isn't ready for production teams. Right? Nothing here is really easy. If you don't know the system, your way around it, everything that you can do in it, it could be pretty overwhelming. But you get a lot of information in here, in this kind of top level container that we call the work package. It's got a number, a name, a description. Over here, you'll see, you know, the workstation, the classification, the discipline that's doing it, the hole and block information that we've got. Down here, this is all of the content parts. So you've got a whole list of kind of your bill of materials here, that is actually being included in this step. You've got context parts as well. This is, you know, everything to be looking around the parts that are actually being put together so you understand where you are in space. Related documents, you know, this could be drawings or PDF manuals. You'll see, you know, the Envision document down here. This is related. So one of the benefits of, being in this, you know, connected format is you see all of these relationships of stuff. In here, you see that this relate... This Envision document is related to this work package. If we open up the Envision document, we see that it's related to this CAD file that is a compilation of all those content and context parts. So if and when something changes, you know, when you go through your ECO process, you understand that these parts, or documents or related items need updated. And so we take all of this, you know, as we're here, and we have the ability to, create or edit a digital work construction inside Envision. So from here, it's gonna pull all of the information that we need from here, you know, the content, context parts into a CAD model. It is pulling, you know, some of this metadata here from our contents and creating tables. We've got, you know, to and from, WBS, planner notes, things like that. And it runs a script to generate a document on a couple minutes, depending on how large your CAD model is. But I've got an example here of what the output looks like. So from all of that information that you just had, this is the output that you can get from a single click of a button. Now there may still be, you know, more clarification that needs to be added, but our goal is to get you as close to finish as possible, you know, from that single click. But you've got this digital work instruction here. You'll find that same work package name, work package number, the hole, the block, the workstation, all of that information, you know, the description coming in here. And this visual representation of all of those content and context parts. So you'll see the content parts here in blue. That is what we're focusing on. Everything else is context around it. And inside these Envision documents I talked about, this isn't just a flat representation like a PDF is on a screen. This is actually interactive CAD. So I can zoom in here. I can rotate around. I could go full screen and, you know, hide and show and isolate these parts to get a better understanding for myself of the work that needs to be done. Throughout this document here, we've got 10 pages created automatically. You'll find the same kit list, delivered to get to the same way it was, you know, in in Shipbuilding PLM. It's brought in here. Maybe columns are are updated or or shown or hidden, to give them exactly the information that you need. But we've got the from, WBS here, to show, you know, where things are coming from. Same thing if we had a ship list related to this one, which we don't. Material list, you'll find in here as well. So this is all of our accessories and stuff that need, excuse me, to be brought in to, you know, to put this product together. And so you'll find all of that information here. You'll also find a full list of these content parts. So we've got this kind of, bidirectional table here. This is that that bill of materials that I was talking about. So, if I click on the three d over here, and highlight a part, it will bring me down to it in, you know, in the in the table itself. And same the other way. If I click here, it'll start highlighting parts, you know, in here, that I can look at. So being able to see exactly what's needed, which parts are which, answer questions, get a get a feel physically for where these parts should go, is immensely helpful, when you're actually out on the floor trying to put these products together. We'll go to the next page. This is where we get into the actual sequencing of it. So if I hop back here, we've got this sequence information here of our context parts. We've got, know, all of these are part of sequence one and sequence two and sequence three. And when you're looking at this data here, it's helpful if you know the part numbers, but you truly don't understand, you know, what the parts are until you see this visual representation of it. And so this table is the same. You know, I've still got, you know, these parts, but you'll notice some of these parts are now ghosted. These aren't relevant to the sequence. Maybe they're future sequences. You can kinda show this buildup of the product. You've got, you know, the the same interaction here from the table before. You've got some labels here as well kind of highlighting parts, so you don't have to go this level of detail into it. But we've created a page for each one of these sequences, for our operators to go and view. So we've got a couple here. And in this table, you'll see, things like the ability to contact planning or send a problem report back to shipbuilding PLN. These are all options that can be done directly in the work instruction. Contact planning sends an email. Problem report generates a problem report in shipbuilding, and so it's very, very helpful. And across any of these, you know, you can do the contact planning or a problem report, but maybe you have, you know, something in the document that you you don't think is clear or something is out of out of order. You can actually right click on any any page in here and give feedback on it as the operator. From an enterprise level, this really ties that digital thread all the way back around. We call it a digital loop, all the way down to your frontline workers. They can give feedback on it, on the documents that can be stored and managed with them so people, you know, your authors can go and update them. That's why we call these documents living. They truly should be living and updating as as expertise as gained and hopefully not lost, you know, as our enterprises are experiencing churn and, you know, the transition of tribal knowledge, we call it, kind of walking out the door with our generations walking out the door. So you've got all these different sequences in here. I'm pulling it together. And so you've got this very, you know, rich document for everyone to put the the, you know, product together, and hopefully hopefully, doing better work out the end of it. So, I will hop back to, my slides here. Got just a couple more to touch on, on kind of some of the outcomes that we've seen from this. So I'll stop sharing here. So what what results have we seen from this? Well, I talked about it. The, you know, the 85% complete DWI, the digital work instruction from a single click. We haven't gotten to a 100% yet. A lot of things holes or, dimensioning can be put in there, or maybe you're creating exploded views. Those are all still manual today, But this is resulting in a 90% decrease in the time to... An immense time savings on the upfront side of it, resulting in all of that kind of p and l error proofing down on the floor, when they when they actually get used. And, of course, the fully revisionable change control document in shipbuilding PLM is is invaluable, making sure that it's always up to date. Old documentation is one of the leading con... Causes of errors in in the field. And so where do we go from here? This is a good question. And as I was putting this presentation together for, SSI's World Shipbuilding Conference, this was their kind of preset slide that they gave me in it. I think encapsulated everything that I wanted to say at the end about what's next and where we're going. But it was also eye catching here. So we've got AI is not a silver bullet. And this can sound a little pessimistic, but I also think it's a little realistic, you know, in the era that we're in right now. AI is all new and exciting, but you'll see that everything that I just talked about was actually deterministic and and, you know, not agentic or reasoning behind it. It is a scripted automation that runs and creates these documents, so we get a very predictable output. So I think that that's very valuable, you know, to get this deterministic output because manufacturing as a whole, and engineering is in fact a precise industry. There are consequences. You know, there could be lives at stake, you know, if if the work is not done well. And so with AI still being a little unknown, on that side, we're we're doing everything we can to, you know, get a very predictable, consistent output. But as I say that, I'm also gonna follow it up with neither a script and automation. I just talked about everything it can do, but I mentioned we can only get to 85% complete, right, from the click. Still requires manual authoring. Some stuff still requires reasoning. It is which parts should be labeled, you know, which parts should be hidden and shown, what look... What saved views did the three d be in. You know, it is... Maybe it's on the sequencing side. You know, we don't fully understand how these parts should go together, but that's part of the understanding that we have to have and reason. And then I think that's where AI really shines. It's it's writing descriptions of the step. It's things that take reasoning. That's where AI is really gonna shine. So I see the future as some combination of these two. And we actually have, taken steps in that direction already. We've got, AI assisted authoring inside of Envision. So we have an AI assistant that you can ask to perform actions. It could be batch. It could be simple. It could be adding labels. It could be reordering pages or, you know, showing and hiding relevant steps, or parts on each step. And we're just trying to speed up that last 15% that I mentioned. But there's still more to be done as I go here. AI querying into shipbuilding, I think, is gonna be a big thing. So today, we get a very deterministic output. We have to call the specific data that we want to put in the work instruction. But what if you need more work instruction? You're not gonna go back and, you know, change the hard written code that created this document. You still need that data, but you're not a programmer. And this is really where AI can shine as, you know, connecting and being able to query into shipbuilding PLM and grab the data at the time of meeting. So maybe you maybe you need more information from the kit list or the ship list. Or right now, you've got the table in there that has, you know, the accessory list, the material list of everything, but you want a, you know, a summarized version of it that kind of builds up and takes everything that is the same kit and kind of, quantifies it so you can get a a much easier look at it. AI really shines in that area, and I think it's gonna be the next step for us. And then looking beyond that, is AI to AI communication. If you've been following Chip Billing PLM at all, you know that AI is on its way. I think AI is on its way for everything. And with us already having it, we've spent a lot of time tailoring our AI model to work instructions. We're making it an expert in the authoring and the understanding of work instructions and the manufacturing process, but it doesn't know shipbuilding PLM or any other data source. In fact, the agent responsible for that would. So if shipbuilding... If and when shipbuilding PLM brings in that agent into there, it understands the data. It understands the format. It understands how to query for that, and it really unlocks another barrier there, and gives you kind of two experts talking to one another to get the, you know, the information that you need when you need it. So, I think there's there's definitely future here everywhere. For AI, it's it's not going anywhere. In fact, it's gonna, you know, keep changing our our our industry, and we're here to adopt it and and be at the forefront of it, and and continue to give our customers, our users, the the best use case they can, around it to to keep them state of the art. Because, manufacturing is one, shipbuilding in particular that is, kind of lagged behind, in technology, and we're trying to get everyone up to speed as as much as possible. We see the demand for shipbuilding across The US here, and it's it's growing and it's becoming real, and we gotta be ready to support that. So everything here is is to that and for that, and so this is the end of the slides. If anyone has any questions, I haven't seen here any in the q and a, so feel free to send those on. But, other things to look forward to, always reach out to us, if you have questions. Even if you have questions about shipbuilding PLM, we can certainly get you in contact with the team there. Always happy to kind of be be the trusted source around this, you know, as industry experts as we deal with this day in and day out, you know, helping guide you on what maybe is the best fit for, you know, for your enterprise and how we can help. I'm also... I know SSI is planning on being at the work boat show down in, New Orleans here at the beginning of December. I'm planning on being down there as well, if I can be. So we'll be there if you're attending. Please feel free to stop by and see us. We'd love to talk to you. But if there's anything else we can do for you, you know, you can reach out through our website. I don't have my email here, unfortunately. I'll put that in the chat here for us to have. So that's there. Feel free to reach out to me via email or through our sales form on our website. We'd happy to get in contact and get some time set up. So that is... Thanks for just sending it in there as well. But any other questions that I can answer, while we're in here? I Don't see anything popping up. So I'm gonna call it there. I appreciate everyone, you know, coming in and joining us today. And, yeah, thank you. We'll we'll be done a little early here and give some of your afternoon back. So thank you.