Skip to main content

«  View All Posts

How JD Edwards Users Are Automating Manufacturing—No Coding Needed!

August 1st, 2025

30 min read

By Nate Bushfield

This session explores the use of an AI digital assistant to streamline work order routing in JD Edwards, focusing on improving efficiency and reducing errors in manufacturing operations. Drew Robb, AI Advisor at ERP Suites, and Greg Edie, a manufacturing expert, present a real-world use case where manual work order routing was automated using AI. The session covers a phased approach to AI adoption, from automation to autonomous ERP, with a live demo showcasing seamless integration with JD Edwards. Additional AI applications, such as improving MRP processes, sales order fulfillment, and quality assurance, are also discussed. 


Table of Contents    


  1. Introduction

  2. AI Digital Assistant for Manufacturing in JD Edwards

  3. Crawl, Walk, Run Approach

  4. Master Data Management and Automation

  5. Digital Assistant Technology and Integration

  6. Live Demo of Digital Assistant

  7. Digital Assistant Q&A
  8. Manufacturing Use Cases and Future Applications
  9. MRP and Lead Time Management 
  10. Quality Assurance and Machine Learning in Manufacturing 
  11. Final Questions and Wrap-up


Transcript

Introduction

So today what we're going to talk about is our AI digital assistant for manufacturing, JD Edwards. JD Edwards efficiency specifically in the work order routing space. My name is Drew Robb. I am an AI advisor at ERP Suites. So what I do is help customers really realize the value of AI in their business. How to really get started. As you've seen, if you've joined our keynote as well as other sessions and, and you looked at our website for this event, there's a lot of there's getting started track as well as some other tracks as well. But that's mainly where in the space I'm in really help customers begin this process of AI. Whether you're just jumping into it now, you have a good feel about it. You started to think about some use cases potentially or you've looked up some different technology. No matter where you are on the spectrum, I kind of will help you in your own AI journey as we'll coin it here at ERP Suites. So real quick background, I've been with ERP Suites now for five years. I've worked in the product space, was a data analytics consulting for a few years. So more on the technical side. And now I'm now I've jumped in this AI advising role, working along with OCI and AWS services as well. And Greg, if you want to go ahead and introduce yourself, my colleague Greg's along here. So as I've said, I'm more in the technology space. I wanted to introduce Craig who's more in the manufacturing or kind of our subject matter expert here at ERP Suites. And he's worked in a lot of different areas. So I'll go ahead and let him take over here. Yeah, Thanks, Drew. Yeah, my name is Greg Edie. I've been involved in the JD with space for over 25 years now, currently been working with ERP Suites for a little over a year. And my focus is mainly in the manufacturing, although I have extensive knowledge in the distribution and finance area. Thanks, Drew. OK, no problem.


AI Digital Assistant for Manufacturing in JD Edwards

So I'm gonna go ahead and get started here with the kind of a quick agenda, just high level what we're gonna talk about. So first we're gonna go through kind of a manufacturing digital assistant use case one, we spoke specifically with a customer of ours, right. We talked about and the best use cases we get from AI are really from the customers. So we're going to talk about that one. And as I mentioned before, it's really work order routing, moving work orders to specific work centers if a certain work, work centers at capacity. So we'll kind of go into that one and how we've kind of formulated how we would approach solving that solution. Next we'll go through a digital assistant demo. When I want you guys to keep in mind here is really order the possible where we can take it. Right now it's just automating work orders, the work centers, but where we want to go, we'll kind of talk about that and it kind of our three phase approach and Oracle will coin it as well throughout this, throughout this event, it's really crawl, walk, run kind of approach. And here we kind of label it phase one, phase two, phase three. So really talk about that and get into how to really get started on your AI journey without doing it all at once. And then lastly, we'll talk about some manufacturing additional use cases. Greg, again, extensive background in this space, really want to talk about again all of the possible where we can implement AI technologies. Now this one's digital assistant specific, but I'll throw in some other technologies we used in different manufacturing areas. So you can see how leveraging these AI tools and technologies can make your lives and day-to-day jobs easier. So just a real quick agenda there.

I'm going to go ahead and move on to the manufacturing digital assistant. So when we talk to customers, it's really important that we start with the problem, right? And, and there's other getting started tracks how we'll talk about our AI journey, but it's really that discovery phase. What's the problem that that customers are facing when they're jumping into an AI solution? What problems are they trying to solve? What pain points are they trying to solve? So it's really that discovery and talking to a customer in that space and seeing what their biggest pain points are and how AI can help and really define what is AI versus not AI. And then the second part is the solution. What does the solution actually look like? How can we, you know, build out a road map for our customers and build that out to have a really good solution in place that will provide good ROI and then we'll have the outcome. What does it actually give you? What are the efficiencies gained, the technical additions, the technical gains you can gain from an AI solution, specifically this manufacturing digital assistant today, This is kind of the framework we use when we really talk to customers and get down into specific use cases.


Crawl, Walk, Run Approach

So jumping in again to this specific one, and again, it's automating work orders to specific work centers. We're kind of going to go over real quick what this use case was with a particular customer here. So with the problems, the planner was manually moving work, moving around workloads. And this was really just they were being kept in Excel sheets written down and then actually uploaded into JD Edwards. And as you can see with all those manual process it, it's it's in tune to have error prone steps. And as we know and, and, and Greg can even signify to this, you know, it all links, all the manufacturing space links back to manufacturing accounting and accuracies as well. So it all ties together here. And the JD Edwards system was not really reflecting what was actually occurring. So again, you can also see it to this point like certain work centers and work orders were getting moved that shouldn't have been moved. Maybe a work center was at capacity at work center A and they weren't having the correct capabilities to move those to B, They didn't have the right data to kind of predict that movement. And that's what AI really brings is the ability to look at all these automation issues we have kind of start with that automation piece and start to implement these solutions here.

So with that, yeah, just really looking at kind of the manual steps, the the aeroprome steps here in the manufacturing accounting and accuracies that came from talking with this customer. So jumping into what I kind of talked about before was kind of a crawl walk around approach kind of our targeted phase adoption with implementing a digital assistant for this customer. So first we we talked about it already was the, you know, the ability to automate the update of work centers and work orders and really getting data out of JD Edwards as well and seeing, you know, what was inside of that all inside of a digital assistant. So that's really that phase one automation piece. Now again, kind of we want to create a forward-looking view as well.

So what's that phase two look like? So basically asking the digital assistant itself to provide recommendations of where to allocate work orders to specific work centers. You know, being able to have the digital assistant learn from historical inputs, which we'll talk about later is an important piece of why a digital assistant is so impactful to implement in your business. And finally, where we want to get to kind of that autonomous ERP is the machine initiate recommendations. Hey, work center A is at 100% capacity. Let's move those work orders over to work center B. And with that, the digital system will actually reach out to you. And with human in the loop that humans can actually say yes or no. And that's another big thing I just want to quickly highlight on is even though we're implementing AI as a phase adoption here, it's not really going to, you know, take over jobs. We still want human intervention. And me and Greg have a couple use cases later on to really talk about that specifically.


Master Data Management and Automation

So yeah, just kind of talking through kind of our target solution when talking with this customer, what we kind of looked at for phase one, phase two, phase three, I mean with that the outcomes of the solution. So biggest thing when you talk about kind of that phase one was the ability to improve master data management and automate the creation of alternate routings. And we'll eventually look in the bombs as well. That's kind of on the road map. But that was really the key initiative when starting this customer's AI journey. Again, as you can see below, kind of business benefits, increased production reduction, operational efficiency. We harped on it a ton with the problem. The biggest thing they were dealing with was definitely that data quality control inside of JD Edwards from all the manual work they were doing and the manufacturing accounting as well. But another thing we got to really think about too is we got to understand that when we start to automate these tasks, and again, I'll bring this back up, it's not about the job loss in general, but it's really allowing our manufacturing people, the people who work on the plans, the planners to actually make more strategic business decisions and get more in the predictive recommendations I was talking about before. So that's really a key benefit that kind of gets lost when you really think about implementing a digital assistant, especially starting with automation.

I'm not going to read the entire technical piece because we'll go through it, but basically we're talking about saving clicks inside JD Edwards. And actually the cool thing about JD Edwards that we'll talk about it is it's embedded inside of your JD Edwards application. And you actually don't really have to do any work inside of JD Edwards. It's all done inside the digital assistants and just connects back to JD Edwards. So you can just read their kind of the steps of, of, of, of what it takes and how it can, can make it easier, right? Limiting the amount of steps taken, right? Limited amount of clicks that people will go through to actually upload these work orders to alternative work centers. And then talking about future here and, and, and stuff we're working on specifically at ERP suites is real time insights to maximize business opportunities and drive efficiency. So what I mean by that, and we'll get to it and we'll show it a little bit actually in another session that's called Smarter analytics is actually connecting the digital assistant to dashboards to show you real time metrics and efficiencies and things that are happening inside of your business. So that's definitely a place we're looking to go as well. And then effectively determine how much to produce of each item. And that's again going back to the phase two approach of really recommending the right amount of items. So you're not, you know, keeping too much inventory and backlog and whatnot. So yeah, anything, anything you want to add there, Greg? I'll stop right here.

Yeah, sure I do. And so just a, just a point to note, you know, in order to have all these benefits and taking there's, there's one area that really needs to be honed in on and that's your master data in terms of work centers, the number of employees, your routings, your branch plant information that's all connected to the manufacturing process. And on the basis that data is accurate, you will start to see the benefits and you will start to see the correct planning messages, the correct works order messages, and ultimately a more efficient system. Yeah, I think the key to the the key to this is master data in, in all the areas. Absolutely. Yeah. And that's, that's really just a good place to start too. And then you can drive efficiencies in other places as well.


Digital Assistant Technology and Integration

Now we went through the solution I want to really talk about real quick, just highlighting the digital assistant technology and why it's so impactful. So really what I want to point out here. And there's a lot on the screen that you can read is the ability to leverage orchestrations, especially when we use OCI services, We use Oracle Digital Assistant when we build it, build out a digital assistants and it has direct authentication with orchestration. So orchestrations are really the backbone of any AI tool that we push out, whether it be digital assistants, anomaly detection tools, document understanding, which we'll get into in later sessions. Orchestration really has the ability to get data from JD Edwards or other third party systems, update information in JD Edwards and add new information to JD Edwards. And we'll, we'll highlight that today in the digital assistant demo today as well as the back end systems, right? As I said, it can connect to many back end systems, connect to large language models. We'll talk about that a little bit and how we're trying to integrate large language models into a digital assistant as well. And a key thing under that natural experience column, you'll see there is natural language understanding, the ability to understand what users have historically put in and understanding.

One thing the digital assistant's good at is actually understanding misspellings and actually understanding what the user's trying to and and learn from historical inputs as well is what you'll see there. Another thing we're really doing is trying to to make it more conversational. We like to use the term digital assistant, not chatbot 'cause you feel like chatbot's really robotic. We want our digital assistants to seem more conversational that you're actually talking to someone.

One of the examples we use is really a college level intern. When we talk about digital assistants, kind of the person who can just do the easy things right away, but still isn't kind of a robotic process when you start to talk about it. And then another one and, and just going to the next. So we can kind of it's a multi channel support. So with this one, our digital assistant is located on E1 web pages. We just found that with a lot of customers being JD Edwards, if not all, it's really important to understand that how we can use a digital assistant as an enhanced feature with the JD Edwards applications, but it doesn't have to be located in JD Edwards. It can be loaded located in SharePoint sites, Webex, internal websites, mobile apps as well. So you can really place this digital system really anywhere. One that we hear a lot from customers is customer portals, right? So putting them on customer portals so customers can start to interact with them when they're buying inventory or or products from your business is 1. We've also really, really thought about. And then personal personalization. It's really just a productivity, cool conversation flow really about all about business. Business process integration is what we really focus on with our digital assistant here. So again, just wanted to highlight kind of the technology background, the capabilities of our digital assistant. And when I go through this demo next, I want you to keep an eye on how we start, how you start, how we've started to integrate these into our digital assistant.


Live Demo of Digital Assistant

So with that, I'm gonna hop over real quick to a live demo. So just give me one moment to pull that up. So what you're gonna see is I am just inside of AJD Edwards application. I'm inside of our lab environment here. You can see up here. My name's Drew Rob. You can see the user up here. So I signed it with my JD Edwards account and then yeah, again, as I mentioned before, it's embedded inside of OJ Edwards application, but or any OJD Edwards instance, but also can be embedded on Webex SharePoint customer sites as I mentioned. So this is Franklin, our digital assistant. We like to call him the godfather of JDE here at ERP Suites. We called our digital assistant here, Franklin. So I'm pulling up Franklin right here. As you can see, it just comes up in a pop up window again, embedded inside JDE, but doesn't take up the entire screen. We have other capabilities of Franklin. We can make it bigger and smaller with web extensions and and we can show those at future demos.

But with what we're focused on here, this is Franklin. So I'm going to invoke Franklin here. So I'm going to type in the welcome to the webinar. It's just how we coin this. And what we're actually doing here is it's actually connecting to a cohere large language model, reaching back out and, and getting a response. We started to do here is really kind of what I mentioned before is kind of that what we call prompt engineering, really trying to make the digital assistant more conversational in the end. So just give it a moment. It takes a second to get invoked. I just opened up the JD Edwards digital assistant here. 

So this is Franklin, our digital assistant. As I said, what it was doing here was actually connecting back to a large language model and you'll see it, it reaches back and knows my user here. So hello Drew, I'm Franklin, an AI assistant and I'll be showcased at In focus. Now we showed this at a prior conference, but with this, what we were showing here is kind of prompt engineering. We can actually go through and actually make the digital assistant more conversational. I won't get this response back every single time we start to try to make our digital assistant not as automated, but again, more conversation and more interactive. As you can see down here, this says response generated using artificial intelligence. And again, just start to implement more of those large language models to again, make it more conversational. Can pull data from other third party sources as well. Moving forward with this specific demo, you'll can see there's a drop down here. What can I help you with? And we have it split out the address book, financial forecasting, inventory distribution.

We're gonna focus today on the manufacturing operations down here. So I'm gonna click on that one. And again, here's a list of tasks that we focused on with this manufacturing operation. And as I mentioned before, it's get current master routings, creating an ultimate alternate work center and then lastly updating existing work order with an alternative routing. So with that, I'm gonna go ahead and type 1 here. So the first interaction, what you're gonna see here is we're gonna have a list of inputs that is coming from JD Edwards here. And you can see all our branch plants. Again, we're using our LAMP data inside of JD Edwards. This might obviously look different for you guys. It says you're reading this digital assistant, it says click one of the above or type in a value below. I'm actually going to type in a value and it's going to be Eastern. Now what this is showing here is the ability for a digital assistant to learn from different inputs and different languages. So it's actually going to understand Eastern as a specific input relating to the M30 branch plant. So just give it a second here to load it again. It's reaching back and again, I, I believe here we're also connecting to the large language model. So it's going to take a little bit to run here. OK, just click.

You look at some of these questions real quick. So would Franklin need to be run locally an in house hardware or is this an API webhook, etcetera? Is there an option for one or the other? Franklin actually is webhook API affiliated. It does not need to be run locally. So that's a very good question.

Now let me just sign out one more time and try this and then we'll hop over to the video and I can walk you guys through that. Just one more, one more chance here. So again, I'm gonna click 1 here and then I'm gonna I'm gonna click Eastern here or I'm gonna type Eastern here for the M30 branch plant. And this, Yep, there it is. So sorry about that guys. Looks like it was just hung again. But as you can see, I typed in Eastern and down here, it actually understands that it's the M30 branch plant. And the reason behind that is a digital assistant. You can actually train it on different user inputs. So it can see Eastern branch plan, for example, item number 220, we can type in red for a red bike and it actually understands that it's item number 220 as well. And then from that see it. And so it went through there. And again, it's, it's just using different synonyms, different user inputs to understand what a user is referring to with their input. So with that, we have the Eastern branch plant, the red bike. What we want to do here is actually select standard manufacturing routings. And what it's doing here is it's actually connecting back to JD Edwards instance and pulling this information. So it's pulling the work order, existing work orders listed below here. So you can see the work center up here, the description, the from through as well as other information.

What's really cool feature that we've added to our digital assistant is actually this application link. So if I click this application link and I go back here and I click the check mark, you can actually pull up the JD Edwards application here, the work with routing operations and see the item number 220 branch plan M30, the routing type M, and you can see the work center information listed here. We're going to focus on this fifth line here, the 200 dash 9/11 the test and inspect. So I'm going to exit out of here for now and go back to Franklin. And with Franklin, what I want to actually look at is to see if we have any rush routing types created for those work centers. So if we want to move a work center, get, get a get an order out more quickly, we can look at that. Again, I'm going to go, I'm going to type 1 here. I'm going to go branch plant M30. And with that, again, you can type in Eastern. We can train it to understand different inputs. We're going to do item type 220. We can click on that or again, type red. And with this one, I'm actually going to select the rush routing type here. And what you're going to see is no routing is found. So you can actually look up different routing types as well. And again, it's just linking back to the JD Edwards via the power of orchestration to actually actually see that information inside of JD Edwards. With that, I'm going to move forward here and we're going to go ahead and create an alternate routing. I'm going to type 2 here. With that, we're going to need more information.

As you saw before, we're going to go branch point M30. Again, the item number is going to be 220. We are actually going to select the routing type, the standard routing type, as you saw before, from that list above. And then with that, we're going to select an alternate routing type. So again, creating that rush routing work center to move different work orders that we want to get out the door more quickly. So I'm going to click rush routing. And So what you're going to see is the old work center. And actually let me scroll up real quick. So we're going to focus again, as I said on this test inspect so you can see the work center. The old one is 209 eleven. I'm going to scroll right back down here and I'm going to click on the old work center, which is 209 eleven and I want to move it to this new Rush work center, which is 208 sixty. So I'm going to click on that again with the power of orchestration mixed with a digital assistant, we're going to we're going to use orchestration to actually input that alternate routing information inside of JD Edwards. So as you can see, the alt routing master was added and then from there we're going to hit a confirmation link back. And as you can see, the rush routing type has been created for item number 220, branch plant M 30. And then you can see again you got that that rush routing type created.

Finally, the last step here, we're actually going to update an existing work order with an alternate routing. So the rush routing type that we just created, we're going to move a work order onto that rush routing type. So with that, let me let me type 3 here to begin that process. And again, we're going to need a branch plant, so we'll go with branch plant M30. We're actually going to enter existing work order number here. Now one thing I want to preference is we can actually enter multiple work work orders here with this demo. I'm just going to show one just for demo purposes. Our lab let me pull this down real quick. So the one we're going to look at here is actually order number 45705 and as you can see, it's currently on the work center 209 eleven down here. So I just wanted to show you that inside of JD Edwards. Now we want to have capabilities in the future to actually look up the specific work order information, kind of like what we did with the the work center routings that I showed you above. So that's that's one new capability we're definitely going to be looking at there. So you actually can see that information with actually not having to log into JD Edwards and do all the work inside of the digital assistant. So as you can see, it was the 45705. I'm going to type that one in going to hit enter here and it needs the old work center. I just showed you. The old work center was the 209 eleven. And then we're going to want to link that back to our new work center, which I showed you before with that rush routing type was the 208 sixty.

So I'm going to click on that one again with the power of orchestration and it's going to be able to actually create that new work center. And again, using the features of application links, you can see that we have created that new, we have moved that specific work order, the 45705 to work Center 208 sixty here. So you can see kind of the start of this automated process, what it would kind of look like in JD Edwards and kind of start, you know, really doing that kind of data quality data management and eliminating those steps inside JD Edwards and doing all the work inside of the digital assistant. So I'm sure we have a lot of questions real quick. Greg, is there anything you want to add from the process? No, no, no, not at this stage. That's actually very good. Thanks, Drew. Sounds good.


Digital Assistant Q&A

Some quick questions here. Yeah, go ahead. Now that's a good question and and you definitely, and we're looking at the capabilities around this, but you definitely have to start back over when you get out of the window. And that's why, you know, we really have this embedded inside of JDE. One thing I didn't show because I was doing everything inside of the digital assistant is you can, you know, you can hop through the other windows and, and I believe it would still show up inside of the window as well. But yeah, it, it, it just, it just really depends. I believe you can go all the way through, But that's, that's a good question. Let me, let me go back with my team and confirm that one for sure, though, that, and that's, that's a good question. I know we've run into a few situations where you'd have to log back in. And that's a really good question though. So let me, let me get with the team and follow up on that.

Okay, continuing. And the answer right now is, is no, but we're looking to build out more modules, especially, you know, what's, what are those specific use cases, you know, you want to focus on right now. As you saw from that window above it, it was really the manufacturing digital assistant. There was a financial digital assistant, there was some others as well. It really depends on the process. And the one thing I wanted to preference with the digital assistant is we kind of try to split up kind of its functionality and a lot of that is around security reasons. Like we don't want, you know, different JD Edwards users, for example, ones working in the manufacturing space going in the financial space for security reasons, right? So we like to split that out, make it more process space oriented. And that's the big thing. You know, we really hope on with the digital assistant, it's able to do a lot of different things and not just one action like a chat bot would do. You know, it has the ability to do a lot of different things, but really breaking that out is the important thing. So it's really a process by process basis and right now its capabilities cannot go out through the whole whatever JDL was module. But definitely we'll work on that in the future. And it's really important to talk to customers to really understand what sort of modules would be the most impactful for building out a digital assistant. So hope that answered your question.

So this is actually it's squaring the JD Edwards applications, the tables inside of there and getting information back from that. When you talk about the Oracle modules and Oracle databases or maybe other databases you have, there's actually a capability called SQL dialogue where you can actually pull information from JD Edwards databases and actually use natural language to query on those, on that data, for example, you know, and it's actually in another digital assistant that we use called the financial digital assistant. But yeah, for example, like you can, you can pull sums, counts, you can start to do sort of that manipulation with your data. Again, I showed just an automated process today, but that's another digital assistant and and it's our financial one where it actually goes into querying different data, actually connecting to machine learning, doing the budgeting, the forecasting, looking up at quarterly reviews. So you're able to do like summations or counts or all that kind of stuff. It's something we didn't illustrate in this digital system, but it's another capability of the Oracle digital system in general. So hope that hope that answered your question.

Yeah, it's actually, that's a great question. And actually, I mentioned it before, kind of highlighted it, but it's actually 9.2 dot 8.2. What, what this is, is actually authentication with any OCI services. I mentioned mentioned digital assistant today, but there's others such as document understanding, so the ability to read documents using vision, so the ability to read images, you know that you might pull through a warehouse, different things like that. So it's 9.2 dot 8.2 has the ability because because as I mentioned before, it's really about leveraging, leveraging orchestrations and kind of mirroring the two together to make the digital system more impactful and the ability to use your data, as I said, get data from JD Edwards update. You know, as well, as I mentioned before, the ability to query databases and 3rd party data sources as well is really important. And that's all really the power of orchestration. So Oh really?

But can you run multiple digital assistants simultaneously? The answer is, is the answer is yes. But it comes back to what I mentioned before with the digital assistant one just being located. Let's take it this way. I was going to say based off the user, right? So when you talk about running multiple digital assistants and you talk about multiple users running different digital assistants, the answer is yes. For a specific user running multiple digital assistants. That's, that's, that's a great question. It's something we definitely we're going to, we definitely need to explore more. I want, I don't want to give you the wrong information. So let me go back and talk to my team. But like how we want to how we kind of picture this is more of a, a flow type process. Think of it as like a process flow. So we build out different flows for digital assistance. So you have your work order flow, you have your financial budgeting flow or what have you. Like just different flows that you can think of for the digital assistants.

As far as multiple users using a like the digital assistant, yes, you can do that. As far as a single user using multiple digital assistants at once, probably just going to have to look at the capabilities and how they all interact. Now, I know there's something called conversation flows where it actually knows what digital assistant you're reaching out to. So you showed the, I showed you before the list of five digital assistants in the manufacturing, the finance. There's a way to actually invoke, it's called like, I don't want to get too technical. It's called like a can. It will actually understand which digital assistant you're referring to and go down that flow through the digital assistant. Now there's a way to actually back out of it and and go into a different flow as well. Now running to simultaneously is the one we definitely need to look into though for sure because I think I believe you got to finish the whole process of a flow.


Manufacturing Use Cases and Future Applications

Let me just hop back in to just other additional use cases. So actually there's one more slide before that. Let me go ahead and share again. There we go from current slide. OK, there we go. So just jumping forward just real quick, I wanna talk about some and highlight some of the demo topics. So again, this might reiterate some of the questions that came up. So again, it's integrated with E1 picking and pouching using the power of orchestrations, E1 pages, some custom Java development as well. It's really integrated inside your JD Edwards experience inside of E1. Again, using large language models. You saw it earlier, just really starting to use those prompts to come back and making it more conversational. We've connected to a cohere command large language model there and start to use that a lot more to not make it as automated and then a big one to point out as well as the application link confirmation. So having the ability to do all the work inside of JD Edwards and kind of have that human in the loop or data validation to check that all the inputs are made correctly or that all the data inside of JD Edwards even is even correct to begin with. So that's very important. And as as I showed before, multi train, multi skill and I got a little into it with the question again, the ability to do various different things inside of one specific digital assistant.

I point out the manufacturing digital assistant, the financial digital assistant, the forecasting digital assistant. And then again, this answered another question as well. It's, it's, it's, you know, it's, it can be in the cloud or on Prem, it doesn't really matter. It can also be a hybrid, hybrid approach as well where all of the all the digital assistant capabilities are built inside the cloud, but all the work is done inside of the on Prem instance, right and it can be housed in there. So I believe that answered one of the the questions before from one of the attendees. So again, those are just some high level capabilities you saw from the digital assistant and where we're looking to go, what else we're looking to build out. And I mentioned as well like the data visualization, data visualization, the dashboard and connecting more to 3rd party services other than JD Edwards, you know, the ability to check the large language models of pull data from there as well. All new capabilities that we talked about. We talked about the digital assistant as well as the sequel query and that I mentioned that we do a lot more in our financial digital assistant as well. So with that, I'm going to turn it over and actually talk about additional use cases inside of of the manufacturing space.


MRP and Lead Time Management

So what we really talk about here is really the make the order, make the stock and the ability to actually provide information and context, automate processing messages in MRP and they also the ability to recommend changing with those processing messages. And Greg, if you want to provide more insight on this. Yeah, sure. That would be, that would be fantastic. Yeah, sure. So on the manufacturing in terms of the MRP messages, I think there's, there's one message I'd like to get across today and that's lead times. There's a number of areas in terms of production scheduling in terms of accurate lead times and, and the risk of delays and bottlenecks. So lead times in terms of the, the bombs and the writings are, are very, very, it's something that you need to be very careful of in this particular area. As far as the inventory management's concerned, obviously you're looking at up to up to date lead times, which will prevent stock outs and excessive inventory. From a supplier perspective, I'm sure you will all realize that accurate lead times and, and collaboration with the supplier is going to be far more efficient and, and, and more reliable. And then the last one is customer satisfaction. Obviously timely order fulfilment and on time deliveries will obviously create, you know, customer satisfaction. So in terms of MRP, you know, in terms of generating these messages, all very well having these messages come out. But if your bombs and writings and your transit lead times is the other one that I want to point out to you, if those are not accurate, these messages that you receive are really going to be inaccurate and probably not not much value. So, you know, the point here is really lead times.

And then in terms of phase two that you've got the drew, I think there's a human element to that. I mean, you may want to expand on that. You know, it's all very well processing automatically processing messages, but you know, I think there is a human element in terms of planning and you can't just automatically go ahead. So, you know, we need to have that little element in there and you may want to add it. Yeah, absolutely. And and I may have mentioned it before, it's it's really just having human in the loop, you know, when you're really training the digital assistants, training other models. If it's not digital assistants, it's another AI technology having that human in the loop for review before you actually make a decision. Because one, you know, when you have that human, human in the loop, you're making sure the correct decisions are made, those historical inputs from those messages are correctly inputted and, and it makes the digital assistant even stronger and better in the future. So that's really important when we talk about building out these digital assistants, having the human in the loop aspect and working with customers and to build out these, it's not just a product we go and build out Willy nilly. We need, we need to understand the data behind it. We need to understand how the processes work because that's the biggest thing. It's processes and data when building out these digital assistants and having the human in the loop and the people who understand the data.

And I might just add another comment here. You know, I, I think that the, the, the big, one of the biggest areas of concern that I've seen in my, in my travels in this particular area is the, is what we refer to as the item master data. And here on an in, on numerous occasions, we're finding clients and, and, and people using the system with inaccurate data. So inaccurate safety stocks interact inactive, incorrect planning codes, incorrect planning families and incorrect lead times. And all of this leads to, to challenges. So, you know, I can only stress that your master data is looked at and, and it's actually and it's accurate and to the extent that if you have correct SO PS correct templates, you can avoid a lot of challenges and ultimately create a lot of satisfaction within your environment. Thanks, Drew. Thank you, Greg.


Quality Assurance and Machine Learning in Manufacturing

Moving on, just another one we talk about is, is quality assurance with this one. It's, it's, it's, it takes a lot of different ways. There's a lot of different tools we can use it we can use in quality assurance Real quick scribe, do you want to kind of highlight the AI monitoring tool and how that would kind of yeah, sure. For the audience. So a little bit of experience, some experience on this side, you know, with the with the monitoring type tools or, or tools that you can, you can attach to equipment. So I'll just use the example of bearings on particular motors. You know, if you're able to pick up the vibrations on bearings or motors with that, we're able to send messages through to the equipment management system or the planting equipment system, which will enable you to create work orders based on a motor that's making a noise or a vibration or a bearing. And ultimately emailing using the orchestration tool as well emailing to the respective supervisors. How does that help us in terms of mining the data as we move forward, we've, we're getting more and more, shall I say, triggers from the AI monitor that is saying to us, hey, all these bearings are breaking down. Who did I buy this bearing from? Was it the same supplier? You know, how many times did this particular bearing or motor breakdown? So, so there's a number of areas where we can start to mine the data that we receive and ultimately and ultimately improve the uptime of a machine. And I guess ultimately you have better customer satisfaction because you obviously producing the goods on time. So, so that's an example at at a high level and certainly can dive into more detail, but it's certainly an area to to look into in terms of your equipment.

And, and just another real quick area that's not in presented on the screen. We talked to you about, you know, Greg talked about the monitoring tool and the vibrations. Another AI tool, just pointing one out is the ability to actually read images and upload them in the JD Edwards for quality assurance, right? So one quick example we worked on in JD Edwards and it's actually another session that we have during AI week called, I believe it's just uploading images or, or anomaly detections. And, and, and really it's one I should, you should definitely look towards because it's actually uploading those, those images into JD Edwards and, and saying, if you know, basically we use the pulley to understand if it's complete or defective is, is 1 route we took. So training a model based on images to understand and do quality insurance based on those images. And what it actually does is actually input data inside JD Edwards as well. So kind of mirroring that that's got that could be a phase one approach as well, just using a different tool other than machine vibrations or what we call Monotron here at JD Edwards that we use. So just wanted to highlight that that area as well. Just another different AI tool, another different capability, you know, for quality insurance and a different route we can take when talking about the manufacturing space with that, I know we we went through a lot there. Again, sorry about the demo hiccups happened with live demos, but I hope it was impactful. You got, you know, you guys got something out of it and you understand kind of our process of the phase one, phase two, phase three of starting to implement digital assistance in your business. And also so other additional manufacturing experience that we really thought about. It's really talking to customers to see where we can take this right, understanding your pain points, your different use cases and how to really start implementing AI into your business. With that, I'll leave it with anything, anything from you Greg, before we sign up. No, no, at this point I just see there is a question on the sales orders and creation using dropped PDFs, you know, on the basis of PDF. Yeah. PDF. Yeah, on the basis you've got a tool to read the PDF, the answer would definitely be yes. I'm not sure if we've done it yet or we've created it, but absolutely, yeah.

And that's, and that's what I was highlighting before. And Scott, I don't, I don't know if you have it handy, but you can drop the session for that one inside as well. The name's escaping me. But actually, yeah, we'll have a session later on during AI week where we actually show the upload of an image into JD Edwards where we just add a button into the P 4210. And actually we have the ability to drag and drop an image of multiple sales orders via PDF, JPEG, TIFF and upload that image with all the correct information. And actually one cool feature and is, is actually be able to cross-reference item number. So a customer item number based on item number inside JD Edwards, but it's all in one click of a button. And after that it would actually pin the document inside of JD Edwards as well for quality assurance validation. Is it the enterprise document intelligence one or? Yeah, Enterprise document intelligence, Yeah. So tomorrow at 4:00 Eastern, recommend joining that session, seeing the demo, seeing how seamless it is. I will say it's about a 5 minute demo with how seamless and integrated it is with your JD Edwards system and really doing all the work again as you saw with the digital assistant inside of your JD Edwards instance. Any other questions?


Final Questions and Wrap-up

Well, right at time. So we might have to come back to these, but one of them is have you considered JDE or other configurators with the sales order entry assistant? Yeah, that's, that's a good question. And and and not particularly me as looked at the JD Edwards Configurator. I can talk to some more of my team, but we really understood what the authentication, as I mentioned before with those OCI services and using document understanding as the technology we saw as well as being able to just embed it inside the JD Edwards. We found that, that was the easiest sort of integration that we found. Now again, it's an, it's another good exploration that we can look at and using JD Edwards configurator. Again, that's not really my area. That's the more consulting AMS side of things, but we can get you talking to one of them. There's a few sessions from them as well during this where they can talk more about the Configurator. That's a really good point and a really good question there. Yeah, sure sounds good. Yeah.

So real quick and I think I answered this, the technology being used behind the PDF scraping is it's called OCI document understanding. Just wanted to clear that up. It's our OCI service. I'm not trying with JD. I was figure sorry, my questions are coming in a little late here. This all right with that, I don't see any other questions. Thank you all for joining today. I hope it was impactful. Again, visit the website to see other JD Edwards. There's other things we're working on digital assistant document understanding. And again, we have more sessions this week. So definitely, definitely tune into those and you can see other AI capabilities and other different areas that we've explored here. Yep, thank you guys again and enjoy the rest of your AI week. Great, thank you very much. Thanks, Drew.

 

 

Nate Bushfield

Video Strategist at ERP Suites