AI Functional Use Cases, Sales & Marketing
August 4th, 2025
17 min read
This session focuses on AI use cases for sales and marketing, particularly in environments using JD Edwards and Oracle Cloud Infrastructure (OCI). It covers both high-impact "quick win" implementations and more advanced predictive applications. Topics range from AI and Orchestrator integration, document extraction, contract pricing, agentic automations, CRM synergy, predictive risk analytics, and customer churn prevention. Emphasis is placed on the value of freeing human capacity, improving insight and automation, and positioning organizations to evolve strategically with AI.
Table of Contents
-
Understanding AI: Foundation and Integration with JD Edwards
-
Advanced Predictive Use Cases: Pricing, Risk, and Optimization
Transcript
Introduction and Session Overview
This is the session I am the most excited for just because it really leverage a lot of my background knowledge as a distribution consultant. So for this session we're going to be talking about use cases for sales and marketing. So there were so many to pick from. So I picked some of my favorite ones. Again, as with all the use case sections, we've been through these. Some of these are, you know, in in production today. Some of these are being used by customers. Some of these you'll see demos of throughout the week, like for example, or quote to order or quote automation agent, right, our our solution there. You'll see some stuff around our our sales assistant, right. So we'll talk about that a bit and then you'll talk a lot about a lot of my favorite use cases around sales order is around predictive analytics. So we're going to talk about that. So going forward, you know, again, you know, we're going to talk around AI. We'll do a quick AI overview. I'll keep it brief because many of you've been in the same presentation or a previous presentation around use cases. So I don't want you to see the same information. We'll talk about AI for sales and marketing, and then we're also going to talk about very specific use cases. So this is, you know, when I'm really excited for.
Understanding AI: Foundation and Integration with JD Edwards
So just a quick overview, right? You know, one of the things that one of the recurring themes we talked about here is that, you know, a lot of people tend to think AI is just like one umbrella technology, but it's such a broad umbrella word and there's so many other technologies inside of it, right? That all comes from, you know, that AI is comprised of, right. So we'll talk a lot about what those different technologies are today and you'll you'll see them reference. And the cool thing is you'll never see me reference just one, you'll see me reference, you know, two or three or sometimes even more behind them. And sometimes I'm not even able to reference all of them that are being used, right? It's a whole host and suite of learning, right? One other thing that we talked about here in a lot of these use cases is that AI is the solution to every problem. Sorry, is not the solution to every problem, right? But it is a part of the solution to many problems, right? So it's not some magic pill. It's not going to solve your problems. It's not going to find you, you know, seven, $700 million in revenue tomorrow, right, that your company is somehow overseeing. It's not going to do those things. But there are ways of applying it that make sure that you make people's lives better, that you make people's jobs easier, and that you free up human minds to do what human minds do best, which is be, you know, you know, do ingenuity thinking, be creative, be strategic, be analytical, right? So it kind of frees the human beings up to, to use their minds more. And that's what I'm, that's what I'm really, really excited for.
One of the other things you'll talk, you'll see me talk a lot about here is AI JD Edwards and Orchestrator, right? So most of the use cases we're talking about here today are around JD Edwards. Some of them do incorporate other systems, right? So with us talking about sales and marketing, you can assume that CRMS are kind of in the realm of this in anything I'm talking about very commonly. And one of the things that I always emphasize is that orchestrator is at the heart of this, right? And that's because the AI technology that exists today that we're using that a lot of, you know, that's out there, that's very prevalent, is all existing in the cloud. And the ones that we choose to use because it's the one that we think works best with JD Edwards, is orchestra is Oracle Cloud Infrastructure OCI, right? Oracle has done a really great job of, of including a lot of really cutting edge AI technology in their tech stack. And they've made it very easy for orchestrations to natively authenticate with OCI services, consume OCI services, and that really creates an AI enablement layer. That's where orchestrator really is. It comes in AI enablement layer, right? AI kind of becomes our brains and orchestrator really becomes the muscle that's doing things inside of JD Edwards, right?
So in this example, you know, over to the right, you'll see how OCI technology can extract, format and create an, A JSON request for an orchestration. And here it will take things and images. It will do document understanding to understand a receipt, you know, and from that receipt, it'll know where to understand the, you know, where to put in the expense. And you know, what's funny is I've used this example so many times and I just realized that this is now a receipt for socks, T-shirts and boxer briefs. So that's, that's a pretty funny Easter egg. I never realized, I never caught that before. So that's, that's a, it's interesting. That's what the receipt is for, but it does, it doesn't work. It does pick it up. I wonder who from our team use that example. So we'll have to find out. So, you know, this is, this is kind of, this is where we, we see the, the, the marriage of those two robots, right? The the union of OOCIAI capabilities reading through the receipt, understanding that Walmart is a customer, right? Or sorry, Walmart is a supplier here or a vendor and, and what is their phone number and what's their address and what's the total on the receipt and what's the amount? And, you know, dates. And it picks all that up and translates that formats and sends it off to JD Edwards, right?
Types of AI and Their Relevance to Sales and Marketing
So we're also going to touch a little bit about the different types of AI, right? So there's different kinds of AI. So again, not to not to conflate what, what we were talking about, you know, just to kind of reiterate what we were talking about earlier, where AI is not just one thing, right? There's a lot of different things in AI and you want to know how you're going to apply them, right? The types of AI that we will talk a lot about today are specifically machine learning, predictive analytics and data science, digital assistance and computer vision, right? You'll notice I didn't talk a lot about robotics and augmented virtual reality use cases. I think there's actually a lot of great use cases for ARVR with sales and marketing, marketing especially, especially if you're in a physical product based business, right? Typically that does not interface a lot with JD Edwards, right? So hence why I didn't include that as a use case. But if you are a customer where your company sells a physical based product, you know, I think VR and AR have a really cool use cases for you, just not something we we do right.
So that being said, let's talk about, you know, let's talk about use cases for sales, orders and marketing, right? And then, and as we're going through this, if you at any point have questions or you have a comment or you'd like to talk through something or you'd like me to address something, you know, let me know. There's also a poll in the chat. So please make sure your, your, your answers are, are in the poll, you know, and, and I'm glad that people are giving feedback that, hey, you know, we've, we've already answered you. So I appreciate that that's there that, that gives us good feedback, right? We want to know where our customers are at with AI.
Quick Win Use Cases
So let's talk about the, the most straightforward, most popular use case that comes out of this, right? And this is one of our, our, our, our quick win use cases. And what you guys will notice the theme here is our vision and our agent use cases are usually are quick win use cases, right? They're easier to to put in. They're very iterative. It's easy to start small and expand on them, right? So you know, this is one of our our most common quick win use cases where what we really are doing is we're just uploading or consuming a a lot of customer purchase orders, right? So your customers P OS into an analysis tool. This is actually getting really impressive because now it doesn't really need like a structure PDF Oracle. We were just playing around with the beta program with this on Oracle and it can actually pick up handwriting, which is very impressive because it can pick up some really bad handwriting from someone who has really bad handwriting. It it's it's actually quite good. So if you're one of those, if you have a ES who are just like negotiating deals and in bars and writing on back of paper napkins and, and taking a picture and sending that to you, you know, don't worry, we've got a, we've got a solution for you, right? So essentially what this is, it's very straightforward, right? We can take a any type of really documents PDF, the pictures, Word documents, I think are supported now Excel files I think we're still waiting on and essentially take those documents, extract details, contextualize those details and then pass information inside of into an orchestration and enter it into JD Edwards. If you'd like to see a demo of this, please reach out. We're happy to show it.
It's funny kind of how you know how groundbreaking it is, but how underwhelming it is because like you just, we just uploaded a couple PDFs and then a sales order just appears and you're like, oh, that seems pretty straightforward, but there's a lot of computing power that goes on behind it. You know, it's trying to understand customers and items and, and things like that because we're building layers onto this and we'll talk about some of those layers here for you, right? This is another one of my favorite use cases, which is taking contracts, right? And extracting information from those contracts and creating advanced pricing set up based on those contracts, right. So I'm sure many, many of you do contract pricing and I'm sure you all love to maintain customer, you know, your specific contract pricing and updating that every year. And you know, I'm sure your, your, your pricing teams absolutely love doing that. You know, for those of you who haven't worked in the pricing realm with pricing teams, they don't like to do that. It's typically the smallest number of customers, typically a very large chunk of volume though. So we do have to do it often, but it's usually the highest amount of manual work, right? Or if you have like large pricing sheets or anything like that, we can take all that information, we can extract all that information and then we can update and enter that into JD Edwards in MVN orchestration, right? One of the things that you'll see as we'll go, as we go through this is I often end my use cases with an outcome, right? Someone is updating something, someone is doing something, right? And that's because, you know I'm, you know, I want us to think about how this can, can go end to end.
That being said, don't focus only on the outcome, right? The outcome can typically be specific to your business, your industry, your market, your, your environment. It's really the insights that are important and the time savings that come from not having to manually do any of this work, right. So one of our agentic use cases around sales, and we'll talk about two of these is an internal agent that can serve as both a source of information, but also as a source of automation. So we designed this, you know, with the thought of the, the outside sales Rep in mind, the outside AE, right? Can the AE have an agent that they're talking with, right, that can answer their questions that don't require strategic thought like they would get from their CSR, right? You know, to kind of again, free up some of the CSRS time and inside reps time to do, you know, more, more strategic work, right? So can I have an agent that can search through previous orders, can search through trends, right? Can it provide me information as I'm about to go visit a customer to be like, hey, you know, how many orders in the last quarter did we deliver late to this customer, right? You know, am I going to go into this, into the sales meeting and, and have to apologize and make this up and, and do some promotional selling? Or am I going to go into the sales meeting with, with being able to kind of toot our horn and talk about the great services we've been providing, right? What’s going on with the account? What’s happening? Why are there orders? What are we doing to mitigate it? You know, how do we make it better, right? All those things that information that somebody needs before they're talking to their customer, you know, you can get that information from this agent. The other side of this is, is being able to enter the order in JD Edwards without having to go into JD Edwards, right? So can the AE upload documents or send information to this agent? So the agent can enter the, the, the sales order for the A on the A ES behalf? Now, going back to the point I made earlier, right, was the way we built the foundation of our, our, our quote to order skills are essentially we're built in a way where we can use them as agentic skills or non agentic skills, right? So that, that that's kind of how we have to think about these things. So, so as you're thinking about AI, right? And I'm, I'm sharing this example with you of, of how the sausage is made to get you to think that when, when you're thinking about your AI use cases, don't just think about today, right? Think about where you're going to take this thing a year, 2 years, three years from now, right? And, and we, we, we have to do those same things because it's, it's, it you inevitably a customer's going to ask me for it. And if I, if I have to develop and at that time, it's going to cost a lot more money versus if I have something pre built, I can give the cost savings and pass that along to my customer, right? Which is what we we love to do.
Advanced Predictive Use Cases: Pricing, Risk, and Optimization
So you know another agentic use case. This is an externally available agent. So this is a customer service agent, not a sales assistant, right? Sales assistant assist your sales force. A customer service agent provides customer service to your customers. So this is the kind of agent. And again, the cool thing is, is you can have this agent be your front facing customer facing agent, or you can have your agent as exposed as an API. So your agent is essentially a consumable agent by a different front facing engine, right? Front facing agent. So for example, if you're going to use, you know, Salesforce communities or something like that to create your e-commerce site or your site and you're going to use Einstein as your AI, you know, kind of your customer facing AI on the e-commerce site, but you want Einstein to be able to get queried information from JD Edwards. You can kind of incorporate Franklin as as sort of a consumable service, right? So that's, that's one of the things that we can help you with that as well. Or you can again, use the orchestrations that are there. So this is an externally facing agent which can update orders, update order information. There's controls. So you don't have to you you can't, it doesn't necessarily need to be allowed to order update everything. But again, we build these things with like building blocks. So we can build it so it can order update everything like, and it can, it can enter a sales order, but you know, that's kind of the individual tweaking for each customer that that we do right.
So in this one, it's it's, it's very straightforward. Customer can, you know, engage with an agent on on order tracking information inquiring about order tracking, seeing where my order is. The agent can get that information from CRM Judy Edwards. You can get that agent information from a carrier API, right? And then it can provide that information back. Customers can send certain updates and say, hey, update my address or update this again, you can build controls around that, right? And those controls are specific to you. So then that, that that data then goes in and updates something inside of Judy Edwards. So before I move on, those are the last of our vision and agentic based use cases. And these are ones that are easier to implement. They're more quick wins. They can be done fairly. You know, I wouldn't say totally quickly. I'm not talking like a month, but you know, it's not like I'm not talking years either, right? So probably a couple months and then we can usually get one of these agents, our first MVP of one of these agents or one of these vision use cases ready and live and in use. So if there's any questions beforehand, please, please drop them in the chat because now we're going to get into some of the more predictive understanding use cases. And these are these are a little these are really cool, right? These are very neat. This is this is future facing stuff that we that that takes a little bit more work to do, but has a lot of value.
So this is this is a very interesting one and one that actually use case has been around for a long time, which is at least as far as I remember, you know, as far as I've, I've worked on it, I've worked on something like this about as early as 2022 and 2021, right. And this is where you are using anomaly detection algorithms to essentially find customers that are not adhering to contractual requirements or minimums on their sales contract. So if you remember, we, we had an agent earlier that can extract pricing information from contracts, right? So that's, that's the beginning part of this. And then taking that information and then going forward facing is like, OK, how do I find customers who are proactively not going to meet discounts for minimums? How do I incentivize them? How do I work that account more strategically, right? So you can you can start to get recommendations. So this is where the recommendation engine starts coming in and it's a little bit, you know, again, these require more data or requires more, more more prep data. But this is where the value starts coming in of AI, right? This is that that that predictive side of AI that that, you know, we all are really, really excited for.
So we further go along the same theme of pricing optimization, right. So, you know, can I use similar algorithms to identify customers where I can optimize revenue using incentives or baskets or promotions or something else, right? You know, can I start to, you know, identify how customers are purchasing So they're buying behavior, right? Are they purchasing certain items in certain seasons? Are they purchasing certain items within, you know, certain times of each other? Can I pair together some some promotion? Can I pair together items to provide some promotional pricing to my customer, like in a basket sort of a way? Can I start to? Find customers that are buying more but are not really being incentivized to buy more right through their discounts. You know, you know, is there an optimization we can run between their incentives and their, their, their, their purchasing their, their, their, their buying behavior, right so or their, their actual their actual purchasing behavior right. So, you know, really this is that kind of engine that really is very fine-tuned for what your pricing is like and what your customers are like. But it's very predictive because what it's doing is it's just recognizing patterns, right? It's it's finding patterns where I really have to optimize for ABC or D and my inputs are kind of, you know, 1234 and five. And it's starting to run every permutation of those and saying, Hey, you know, is there a world probabilistically where where, where scenario A&B kind of or one and a meet and, and, and what's the likely outcome of that? Right. So so it can run those types of analysis much, much faster than a human can, right. And the nice thing is I like to call it a recommendation, right? I'm not saying as these things are smart enough to make the decisions on their own. What I'm saying is these things are really smart and can give great recommendations and it can take a lot of the thinking time off the human being. So a human being can use their judgement and their creativity and their ingenuity to make the right decision, right for the account. Similarly, we have essentially in the same vein upsell opportunities. So there is the sales assistant inside of JD Edwards. If any of you use the P-421-O1 power form that's there, it's a little difficult to to really use. It's very in the moment, right. So it's a great in the moment type engine, but sometimes you need to strategically analyze your order trends and your order data to find upsell opportunities kind of strategically across your whole, your whole customer base or across a customer's entire ordering history, right? So, you know, same thing, same kinds of algorithms that can then extract, can, can analyze those patterns across an entire customer's behavioral history and, and start to start to, you know, predict based on that. The really cool thing is, is you can start to now pair this data, right? All this, this predictive data that I'm talking about here around pricing optimization and things like that and start to pair it with your, your, your, your CRM data, your, your marketing data, right? So if you're tracking anything around buying intent or, or if you're tracking anything around, you know, like page visits or site visits, you're tracking things around blog visits right now, you start to pair these things up and you start to get some really cool insights around, OK, you know, how are my customers really buying
Customer Churn Prediction and Closing Remarks
Another use case I think is very important. So this is a risk mitigation use case. So using anomaly detection, we can basically find orders that are at risk of fulfilment delay or non performance, right? So this is really important for customers where you have punitive, you know, agreements in fulfillment, right? So if you, for example, are late on deliveries or you send them the wrong EDI transactions or you, you know, something is missing in their paperwork, they punish, you're fine you for that, right? Because it's part of their, their vendor compliance program. These kinds of risk prevention formulas are really great at identifying orders that have high probability of delay or non performance or non conformance, right. So again, these are probabilistic risk algorithm. So it's not like it's going to be 100% certain. What it's trying to do is it's trying to, it's trying to predict risk for you, right? And, and, and it, it gets better over time, but it, it in, in many cases, sometimes these fines, especially the punitive ones, can can, you know, can really add up. So and, and, and actually more than add up, they can sometimes cause you to get delayed, you know, dropped as an approved vendor for some of these these customers. So it's good to, you know, implement anything that can help you identify that risk right before it actually happens.
And lastly, and I think this is the coolest of all my use cases. This is my favorite one. I've done this with, with Salesforce and HubSpot and, and, and, and I would very much, you know, like to do this based, you know, with, with pairing JD Edwards data. So if you're your customer who, who would be interested in, in, in, in pairing up and trying to do this use case, this would be a really cool one where I'd, I'd love to partner up, you know, because it's, it's what it's definitely possible, right? It's, it's not, it's not one of those things where it's, it's 100% certain, but you get better and better with it over time. So this is where you can, you can analyze customer, customer performance, quality data, complaint data, if you have a ticketing system, ticketing data, you can even start doing sentiment analysis on customer communications. You can incorporate a lot of things. This gets really crazy what all you can incorporate in this, right? And you can essentially build an index, a risk based index to start to predict the likelihood of customer churn, right? So, you know, again, this is it. It gets a little minority report E if you've ever seen that movie. But basically, you know, what we're trying to do is, is predictive a customer's going to leave you right? Or predictive a customer's going to stop buying from you. And, and what it comes down to is what behaviors do you typically see before a customer either stops buying from you or starts buying a less from you or just walks away from you overall? So if you are someone where your CMO or your, your CSO is really, really focused on how do we prevent customer churn? There's, there's competition that's peeling away customers. If you're really focused on how to prevent customer churn, this is, I think AI is going to be one of the major ways of doing that, right? And there's so much data that has this required to go into that, you know, so much of that data is going to come from JD Edwards and, and doing that analysis and that, that, that productivity, that predictable, that predictive analysis is, is I think going to have huge benefits for many, many, many companies where you're trying to do something like this, right? So this is a really, really cool use case. And like I said, if you're interested in, in, in helping partner with something like this, we, we, we love to partner with, with, with a customer on this one. I've done this before with, with other systems and other types of data. And we'd like to try to do this again as well, right. So that being said, thank you all for spending some time with me. Appreciate all of you who joined and stuck around. Thank you, Scott, and appreciate you all and hope you enjoyed walking through sales or use cases with me.
Topics: