How AI Could Empower Any Corporate Business?‍

‍

Raghu Banda is the Senior Director of AI Product Strategy at SAP Labs. With over two decades of experience in the software industry, Raghu has held various roles at SAP since 2001, specializing in predictive analytics and machine learning since 2013. He is responsible for integrating AI technologies into SAP's cloud ERP products, focusing on enterprise benefits.

Raghu holds a Bachelor of Science in Computer Science & Engineering and is an alumnus of the INSEAD Business School's Management Leadership Program. He co-authored a book on implementing machine learning with SAP during the COVID-19 pandemic and hosts a bi-weekly AI-focused podcast, "XTrawAI". His combination of technical expertise and leadership skills places him at the helm of AI product strategy driving innovation for businesses.

‍

Host:

Hello, Raghu, how are you doing?

Raghu Banda:
I'm doing awesome, how are you?

Host:
I'm good, I'm all right, I'm doing my best. I'm gonna introduce you here. Uh, this is—and they correct me from around—Raghu Banda. Yes. And, uh, okay, he is, uh, he actually does a lot of different things but I think what he's been doing a lot, uh, is working at SAP and, um, uh—and I, I, as one might assume, quite a litany of roles, almost like maybe even take seven, eight roles, but a lot of it luckily for us has been in big data and, uh, sort of, you know, bringing AI models together. And right now, he is the senior director of AI product management at SAP Labs. Uh, I think he's gonna come on and talk to us a little bit about what he does and a lot of a little bit about AI and other things that are advancing very quickly and, you know, his ideas, how it was, and, you know, what they possibly can be in the future. So, you know, thank you for coming on.

Raghu Banda:
Thank you, Host, for giving me this opportunity to get on board and speaking about some of these AI technologies. Maybe I can talk in the context of how big enterprises are building it. I'll talk more, uh, with my experience, uh, uh, with SAP. I can talk more about how, uh, from the enterprise standpoint, how AI and analytics are being built into it.

Host:
Okay, so let's dive right into it. First of all, I'd love to know more about what—I mean, machine learning and everybody is sort of thinking about it now. I think about AI learning now, but you got into it very early. It's always interesting to me, the people who started early. What did they see in this area that made them feel like this is a direction they want to commit their careers into?

Raghu Banda:
So there is one saying here, Iheatu. Machine learning and AI, they go interchangeably. Of course, machine learning is a subset of AI. This is where for me, the experience or interaction with AI started way behind during my final year semester days in my engineering back in India. These were the mid-90s and there were not many resources, but obviously, our human mind is always there, like trying to understand what are the different possibilities out there. How can I do some kind of predictions? And how can—there is always this fantasy about how robots can come into play and how they can design and how they can help, uh, become human assistance. So that's how my interaction started. There were not enough resources. I just wrote a white paper during my final year semester days and then I got into, like everybody else, into the IT world. Teddy is back. I got an opportunity when I've been working with SAP. Uh, just if he was venturing into the AI in business operations world. Got an opportunity of, uh, how we can, uh, help our customers provide better experiences through automation of business process. So that is how our journey started. And for me, it was a great initiative because I could—if you go back into our programming days, there are a lot of these—you do manual IF statements, IF-THEN-ELSE statements. This is how the old, uh, I would say, AI started. It was like a manual AI, I would, if I may say. You understand what are the different things and how do you act to that? So based on that, my interaction started with that. So I would say that yesterday's mere analytics, uh, today's analytics is yesterday's machine learning or predictive analytics in business. And today's, uh, machine learning will be tomorrow's mere analytics. The reason why I say this or the code is that technology has been improving at a tremendous pace in a fast pace, and this is how you can build better models, better algorithms. And as AI progresses, we can expect to see the future of business AI being fully integrated into all aspects of decision-making, providing even more advanced solutions to solve complex challenges.

Host: 

Fantastic! So, um, you were from what I'm hearing, excited about what it would be like to have a sort of a digital assistant, right? And what it would be like to have computers assist you in the way that a human can. And that's, you know, I like, at least at that time, quite a fantasy. So, got this new opportunity, and you decided to take it. So, moving on to some other ideas... where do you see— and you should have alluded to this before— how do you see AI's integration into larger enterprises, into bigger companies, thousands of people, lots of um partners? How is AI going to immerse itself in that space? And, you know, from my understanding of AI, it's quite a foundational technology, right? It's almost like you have a horse, now you have a car. How are you going to now start talking on traveling now that you have a car? So, how does AI go into these spaces and restructure the organization, restructure the way that they think about information and organization?

Raghu Banda: 

Yeah, so that's a great question, right? So, because now that we know the technology is out there... if I go back maybe four or five years back in the lane, ever since I started my journey in this AI-ML space for the last eight to ten years, from what they are doing and how we can enhance their user experience, can we do that by adding additional technology into it or additional understanding their behavior, their patterns, and then provide some recommendations or provide some insights and say that, "Hey, this is what you're doing, not so good. This is what you're doing better and how you can enhance that experience." So, this is where we have been doing it without even naming that it is AI. Obviously, now with technology enhancing, so there are three big pillars I would say as part of when we are working with AI: right people, processes, technologies, and the underlying layer being data. We all know that... the topic you asked about these enterprises, right? All of these enterprises have humongous amount of transactions or interactions with the other enterprises or other suppliers or vendors and their customers. So, there are a lot of inflection points, there are a lot of integration points. This is all data. So, data, you got a lot of data being an enterprise firm or being an enterprise. All these enterprises, they have designed and they have set up huge processes which are implemented by the customers who are using it. So, the companies are also now good in understanding what kind of processes need to be there, whether it is a procure-to-pay scenario or an order-to-cash scenario. If you take a particular business process, now you have tremendous amount of data. Yes, you have the processes that are set up, and now with the advent of the different computing, the computing resources are increasing at a faster pace. 

Whether it is with the CPU processing power, like according to Moore's law, every—the processing speed is getting doubled, increased by 10 times, and now with the GPU power as well, the graphical processing unit, you could at a faster pace, you can analyze what is happening. So, in a way, your technology is also improving at a faster pace, and hence the algorithms that I've explained earlier, the manual "if-then-else," it has now evolved into building the decision-level algorithms, whether we might call it neural networks or decision trees. And that is how the pace at which you could build the algorithms also increased because the underlying technology pieces like the CPUs or the GPUs that you are using has increased. So now that you have addressed, yeah, you have a lot of these processes that you have set up, the technology is improving. The big block or the big thing is that the people part. How do you make sure that people understand how to use these processes? How to use this technology and the data and how you can improvise the business, the process, so that the end user experience improves, whether it is a consumer who is using the technology, whether it is an enterprise using the technology, so it all ties around it how the people play into this. Right? So if less deep, I mean we can deep down, let's deep dive into, um, the process a little bit, right? So how do you, how do you see a company being able to think about what it would mean to have AI be the center of their process or just be part of their process in the first place? Like so how do you say you have this tool? So what are you constructing? This is before you know exactly how people will use it, maybe you just want to just have something small and, uh, that can grow as people can grow with the understanding. Um, so how do you, so how do you have a company, an enterprise, inject that into their ecosystem? Would it be the best place? How to do it maybe you do, you know what rate, how do you, and what is the process, right? And I guess the process will depend, but what is the fundamental idea if there is one? And I can also see how that is a hard question to ask without understanding the people and who are you who are using it, right? Is going to be different from a salesperson and how they value this technology and what they’re going to try to get out of it, right? Um, so I understand all that as well. So if I have a large company, how do I see myself first with the process just to, you know, just to get a little bit of an understanding of what that would mean independent of the people? Uh, and, and start thinking about because that those pieces when they sit, they sit for a long time. The people are kind of movable and people, by definition, are a bit more flexible. So even if they can’t do it now, they can be taught, but the processes, the processes, once you have them there, the idea is that they’re going to stay for quite a while. Um, fortunately or unfortunately wrong.

Right. So let us, uh, unpack that a bit of deep dive into the process bit a little more, right? Like, you know, let us pick up a particular process like for cube to pay. This is a particular procurement organization or a procurement unit wherein you have your list for the people. 

Host:

Can you explain what procurement is?

Raghu Banda:

Yeah, sure. So procurement is a particular business unit in an organization where they are responsible for overseeing how you are buying and selling habits, whether it is items, whether it is whatever is needed for your organization. So you will have the procurement specialist or the procurement organization overseeing how much you are spending and identifying the different items that are needed, whether it is. So that is the reason I would want to take a particular example in a particular organization. You have, let us take the example of a plant manager in a particular plant organization and they would want to, they are running out of some spare parts or a particular item when they are working in their, uh, in their plant or relocation. So obviously, they would want to order this particular item. So when you want to order this particular item, you will need to have, you will need to understand if there is a particular RFP process available, the request for proposal process, or if there are some contracts already aligned so that I can directly place a contract. I can use this particular contract that I have and I can directly reach out to the supplier or the vendor of this spare part or this item and if this particular item is available, it will be delivered. If it is not available, maybe there will be some communication that happens between the purchasing unit and also the supplier who's going to supply, and maybe they will say that, "Hey, this item is not available in XYZ location of our factory or maybe we will get it from some other location around the globe or they might say it is not even available. I will provide you with a substitute." So all and then once it is available, the item is, uh, uh, placed for, uh, uh, order is placed. The item can now be purchased and then it can be delivered. Once the item is delivered, you'll have to see—so this is a whole process—then the payment happens.

So how can AI in corporate business help improvise this process, making all these interactions very smooth? And if there are any hiccups in this process, how can you understand and predict ahead of time, saying that, "Hey, at this particular location, suppose my contract is expired with this particular vendor or supplier, do I need to extend the contract, or do I leave it out?" With AI in corporate business, you could streamline this entire process, allowing for better predictions and smoother decision-making, ultimately improving procurement operations. Additionally, the automation of business process through AI could ensure that repetitive tasks, such as vendor communication or order tracking, are handled automatically, saving time and reducing human error. The use of predictive analytics in business could also provide valuable insights into trends, helping procurement teams forecast demand for various items more accurately. In this way, AI for competitive advantage becomes a crucial tool for organizations looking to stay ahead in the market by optimizing their procurement strategies. As we look toward the future of business AI, it is clear that AI-driven solutions will continue to evolve, making business operations more efficient and cost-effective across industries.

‍

Raghu Banda: 

Yeah, so another example is I ordered this item, and the supplier has said that it is available. We have placed the order, but it is getting delayed. I did not get it. I would like to know when this particular item is going to be delivered so that it can go into my production process. So, I have to do a lot of planning. So in all these different steps, based on the historic information or the historic correspondence with these suppliers, and because I have all the data, I can say that, "Hey, in this particular region, in this particular supplier, there is always some kind of a delay happening." So, you could provide some kind of predictions.

Host: 

Sometimes what happens is that if there is a delay and the product has already been shipped, you could also have added additional sensor technology, like the IoT technology, where you can even understand that, "Hey, on this particular route, there is an accident or there are bad weather conditions," so again, it is getting delayed. So, all these kinds of information—this whole process for the end procurement organization or the procurement specialist who has ordered this particular item—it is available, so that the whole process can be enhanced, and the experience is enhanced by using AI technologies.

Raghu Banda: 

And when I talk about AI technologies, it's not only the machine learning-related algorithms and doing the predictions. Maybe in some scenarios, we also use some kind of automation technologies—automatic bots, which can also automatically... the interaction happening between the supplier and the purchaser. The emails can be automatically read and uploaded into the system.

Host: 

So those kinds of things—the emails can be automatically read, or red and uploaded into the system, right? So, those things you could do with automation solutions.

Raghu Banda: 

Yeah, right. So, this whole process, you could have different AI technologies used at different steps.

Host: 

Yeah, so you were asking something about it?

Raghu Banda: No, I’m curious about, you know, sort of—you draw that landscape, you know, lots of essentially transactions, where the infra, goods, services, and variables that interact with goods and services—being the things that are being procured, the state of the procurement is being identified, and contracts, and various things that have to do with business relationships. Then, under the condition that there is some sort of relationship, and there is some sort of procurement in principle, then you start considering whether or not that will be happening at the rate that you want, as a kind of a company—whether this service or this product or this as a product-service provider is actually giving this to you in different environments. Ideally, so you have this sort of landscape, and then there’s lots of opportunity for education.

Host: 

If nothing else, being able to sense the state of this whole relationship. So where is AI—where can AI, you know, create a qualitative improvement? Because there’s opportunity for inspection at different points and opportunity to raise red flags, but quite, you know, sort of the qualitative improvements are, you know, and there’s also, like, aggregating information, right? So over time, all this information has been collected about the way that this particular vendor behaves. Can we now say that most likely this is the outcome or these are the positive outcomes we can leverage for this vendor versus this vendor?

Raghu Banda: 

Yeah, so independent, you know, that we did—or this is something that we found that this vendor was able to do, but they are no longer able to do. Or over a period of time, they have reduced in productivity. So, those are sort of, there's a bit more qualitative, right, where it's an aggregate of the outcome of multiple data points. So where can AI in that space inject itself either to increase the qualitative improvement to an even higher degree or reduce the complexity of the interaction of being able to extract value from all the multiple interactions with data points being recorded? Sort of backwards or forwards in a way, but you know, sort of where are you and how is it providing value there? But that's sort of my question if that was clear.

Host: 

Yeah, so I got it. I got it. So you're asking about how does this provide better business outcomes to the... yes, exactly, end user, whether it is a procurement specialist or the purchasing manager. So there are, at different points, like for example, I'll take three different steps in here. There's one step which I did not even mention even before a particular order is placed. There are situations that you could take a picture of a particular spare part that has even gone bad or a similar spare part, even if you will do a particular feature of this spare part item, and then you could upload it into the system. This is where there could be huge quality improvements rather than searching their whole set of category of the items that are not available and trying to match. I directly upload a picture of a particular spare part that I'm looking for. So here, right, you can improvise on the shorter time to respond. Right, like you can find this particular... you can do an image search or image match, and then you could increase the time to find, reduce the time to write it. That is one rank. The second thing is...

Raghu Banda: 

Yeah, yeah, go ahead.

Host: 

No, no, go ahead, please.

Raghu Banda: 

So, the second thing is I mentioned that you have particular contracts set up with these different suppliers, maybe, right? If you're running out of... if you have a particular contract that is expiring with a supplier, you need to get into a contractual agreement. If not, you might run into a situation where you are under a particular contractual obligation with a supplier, and you are getting the goods delivered at a better price. So if you are running out of the contract, so this is again another place where AI could be used. Like your system will let you know ahead of time, hey, my contract is getting expired or running out. There is a second situation.

Host: 

So, but how does that differ from the timer, right, where you know that this is six months or a reminder on your calendar? How does AI create the kind of qualitative improvement you identified in the previous example, where being able to identify things is now trivial, which increases the time—well, sorry, reduces the time that it takes to find something at the increasing the amount of time you have to deliver something or just even in general, throughout the whole ecosystem, it reduces, you know, if there is any issue, any red flag that happened, identifying the solution was just faster through following the board. Right. So for this particular situation, how are you different from a sort of like a calendar reminder, right, where you said six months from now, I need to renew this?

Raghu Banda: 

So it's not... yeah, like I explained in the first example and the first step, where you have been doing the image search, it's reducing the time to response. So there, you're getting a huge boost, and in this contract expiration situation, it's not just merely the time or a calendar set, right? Like because you have X amount of contractual obligations set up with this XYZ supplier, it's not only the time period but the amount of contract that you have used. If you have consumed this particular amount of contract in each of these contractual obligations, you will have subcategories with different items, subitems available. It's not only just a calendar, but also the contractual amount, so that as well will... you will need to constantly understand how much of that contractual agreement, how much of this is available, and then provide some kind of timely alert saying that if you do not renew the contract in a better shape, you might be running into a situation of spending more, but so, so if you have, even if you do have contractor agreements where you have multiple things and different tiers of, you know, it can be from a car to a single bolt in the car, just an example, uh, where if you, assuming you're a car manufacturer, right, you're running out of—because you usually run out of both quickly—but your contract is from all the pieces of the car, so you need to react, you need to resign for more car pieces regardless because you can get it at scale, you need it anyway, uh, and Baltimore are fast, so you may as well just double up contract, right? Even that you can consider that to be sort of like maybe not time-based but quantity-based, where as long as the input value is in threshold, you could trigger a re—reward, a trigger at least for the need for renegotiation or, you know, an update on the contract.

So how can AI in corporate business help in that situation? Like, uh, and even if it's just an aggregate of being able to plan that, you know, there's lots of ways, but I just want to know what you think. So how do you—you know, I'm looking for qualitative updates, like, uh, qualitative improvements, updates, improvements, but qualitative improvements in AI in corporate business in the same way where you know that that says okay, this is definitely going to reduce the line. The automation of business process could come into play here by reducing the manual tracking of contract statuses, automatically sending alerts when it's time to renegotiate or update contracts. Additionally, leveraging predictive analytics in business would help forecast contract usage, ensuring that companies can proactively order more materials before they run out, potentially avoiding delays. By incorporating AI for competitive advantage, a business can leverage AI-powered systems to not only track contract limits but also negotiate better terms based on predictive insights, gaining an edge over competitors who might be reactive instead of proactive. Looking at the future of business AI, these systems will likely become even more sophisticated, with AI systems integrating deeper into supply chain management, creating more seamless, efficient workflows.

‍

Raghu Banda: 

Here I would say, like you've explained, I think, uh, there are different input variables that will get into the calculation of this, uh, contractual negotiation of the contractual expiration, uh, and when we are putting these different input variables, it would take the form of a regression kind of an algorithm where you will have to take these different input parameters and come up with this regression algorithm and understand that hey, maybe this particular contractual amount for all these sub-items and depending on the time that you have set will provide this much amount of your contract has been expired or is expiring. Uh, so it is a, uh, you will have a quantity as an output providing that okay, this much amount of this is expired or expiring, and then you could make some decision based out of, no, I do not want to further, uh, expand this contract, but I will go with another supplier or I will leave this particular item and I get into a different requested proposal. So that is where if I... it might go, yeah.

Host: 

So let's, let's—I thank you for that—so let's move on to, just because of time, sales and sort of how your, um, understanding of AI affects the sales process, which is, you know, very much the input of a business. So how do you see this technology affecting the way people reach out to other people, businesses reach out to other people, businesses reach out to customers to identify the value that they have and being able to do that, um, at scale and also sort of bring people down that sales funnel? So, how do you see AI affecting those areas? But specifically—and I think small businesses have been very flexible when it came to using chat to BT for their copy and a lot of that has plotted the landscape, uh, business landscape. But how, how are Enterprise companies that, uh, you know, sort of their engagements or their sales are a lot more nuanced and complicated, how is AI hoping that—how is it AI helping them, you know, shake hands with other large numbers, um, to provide value?

Raghu Banda: 

Yeah, sure, sure. I'm glad that you brought up the chat GPT component as well because there is this particular, uh, sense inquiry process: sales inquiry to sales quotation to sales order and sales order completion. In this particular process, I'll also explain where chat GPT or similar LLM technology can also play a role. For example, an internal sales representative who works there are like maybe three to five sales representatives who work for a sales manager. So their everyday, kind of a day-to-day, day-in, day-out job is to identify what are the number of incoming sales inquiries that are coming in and understand which of these sales inquiries ideally will be converted into a sales code and out of this sales code, I can create a sales order. So once  the sales order is created, obviously you can follow up on this sales order and then you can close them, and obviously, you get paid for that. And then you will know, so maybe there are 100 inquiries, out of that maybe 60 quotes are created, and out of those, maybe 20 to 25 sales orders are really created. So this whole process, you see, it's like a funnel, right? Like, so yes, you see these different sales inquiries when they come, they can come via different forms. It can be like a fax, or it can be like an email, or like a text message, or these different kinds of things. You might get these different requirements, and all this information has to be now analyzed or parsed, and you have to be accepted into your system.

I'll take the example of, I'll leave the sales inquiry part and get into the sales order because that is easier to discuss in the example that I'm going to talk about. The sales representatives, they get emails, different emails, and each of the emails, they might have these purchase orders coming via a document like a Word doc or a PDF, or maybe normal email text. Now, this information has to be parsed and uploaded into the system. In some scenarios, once you upload it to the system, this particular purchase order document has to be converted into a sales order. The conversion process — so getting the email and uploading it to the system is a manual task generally.

So this manual task can be easily handled by using a bot, an RPA bot that can just read an email, take the purchase order PDF file or a Word document, and upload it to the system. That is the first step. The second step now, once it is uploaded to the system, it is still a PDF file, a purchase order PDF. It has to be converted into a sales order. So what you have to do is, this is where you have your text recognition machine learning service which comes into play. Okay, it has to scan through the complete document and understand the different fields, convert it into the sales order item. You can convert it to the sales order related information if I have got all the information needed to create a sales order.

And in some scenarios, what happens is that you got all the basic information or the elements to create the sales order, but you need to create the description. In some scenarios, the description that is sent via the PDF purchase order PDF or via the email is not well-structured. So this is where you can use a large language model service (LLM), where it can summarize that information and create a nice sales order description text. You complete the whole sales order and then create a sales operative system.

Once the sales order is created in the system, there might still be some kind of a notification or alert saying that, "Hey, yes, you have done all this, but there is still some information missing." Again, this is triggered by an AI in corporate business service, saying that you have passed the complete document and you have taken the elements, you have also taken the LLM service and created the sales order description, but there are still some fields that are missing.

Then, so you have greatly reduced the work that is needed for the internal sales representative. Rather than processing hundreds of documents, he will just see that, "Okay, 80 percent of these documents are processed, and recommendations are provided to him." Okay, all this is there, I can recommend creating a sales order for this quote, or he would say no. So, you will get the recommendation saying that these are the different things. And in some scenarios, if the sales order is not created because there is some information missing, he will get into the system, fill out what is the much-needed information, and then create the sales order. Right? So that is — that saves almost 80 percent of the time needed.

To identify and another thing is even just being able to identify that this is an intent to buy on emails could be done by an LLM where it reads and identifies some probability that there is some intent to buy and I did, you know, and at least even it allows you to do a follow-up if necessary or just sort of ensure that you are able to capture the maximum amount of intent, you know, and then that in and of itself can increase the conversion rate on your sales funnel which has a lot of value. This is a great example of AI in corporate business improving efficiency and driving better results for organizations by automating routine tasks and enhancing decision-making processes.

‍

Host: 

So now we've talked, you know, we talked about procurement, we talked about sales. One last thing I would love to ask you about, and for those who don't know, you've actually written a book on all these ideas when it comes to AI and how that integrates machine learning, how that integrates into Enterprise, um, Enterprise companies. But, you know, moving back to the finance part for finance, how do you take advantage of this technology? How do you use it in the finance area, again in sort of in more in the enterprise space and very, very large companies?

Raghu Banda: 

Yeah, so in the finance area again, um, there are a lot of different applications where it can definitely help the finance line of business whether it is the accounts payable Clerk or the accounts receivable Clerk or the accounts payable AP manager manager, right? When they are closing the books at the end of a quarter, there are a lot of reconciliation that you have to do between your various accounts. This is where machine learning can do, or AI services that you are building can greatly enhance and match your documents. You have maybe XYZ accounts that have to be reconciled to make sure that all the information is really closed, and this is where we have different, in each of that there are numerous number of use cases like I mentioned, like you mentioned, yes, I got this book here which is where I have highlighted some of these different aspects in finance line of business or procurement at each, uh, when you get it with each of these use cases I explained on the business side of how it can greatly enhance the finance manager whether it is the AR Clerk or AP clerk in closing their books faster. They can greatly reduce their amount, the amount of time they spend and which will save it costs of dollars for the organizations, 100 pounds.

Host: 

So, so I like to talk about in general, where do you feel like AI is going to go? Where do you think it's going to—I mean we've talked very specifically about very specific parts of Enterprise organizations and sort of where that technology is going to massage itself into something that is unique for that organization and how they need to think about it and how that is going to provide value, but do you have, uh, I'd love, I'd love to know in general what you think about this technology because you've seen it grow and it's been exponential. I mean there's been a lot of exponentials happening over the last, uh, over the last 10 years. I think this last 10 years have been the years of exponentials on cars, on kinds of currency, on some real foundational technology, um, on media, the cut, you know, there's a restructuring and media that's happening, you know, quite literally as we speak. Um, so it's been a very interesting few decades and so where do you know AI has sort of maximized on this utility? It's, it's sort of, uh, yeah, you know, it's, uh, it has been working towards becoming something that has utility, uh, its potential was clear, but now it truly has a surface area that most people can engage with. So now where do you see that going, or how do you see that, what do you see that turning into relative to the experiences you've already had?

Raghu Banda: 

So to be, for the last 10 years, right, I think so this is, this is a, this is a burning question, right? All of us have now, I think there's a lot of these things going on around the—it's no longer a hype, AI in business operations is no longer a hype, it's almost a buzzword now. We see this, uh, affecting our lives in many different ways, so you can take any application whether it is the uh initially the thought process was like 25, 30 years back when I was in my school or my doing my undergrad we all had this fascination of AI, the software, right, robotics, the hardware working in tandem and creating a future where you have these flying cars or where it is the robots coming and taking over. That has not yet happened. Maybe that will happen in the future of business AI; it would take time because hardware working with software and creating this will take time. So I would say the blue-collar jobs I think, yeah, it would take time. But for the white-collar jobs, currently, the knowledge information or the knowledge worker... I'm not saying that the knowledge worker is going to be replaced, but AI for competitive advantage can play a huge role.

For example, this conversation that we are having here, preparing for this conversation, earlier it used to take me a lot of time, a lot of effort to prepare, but now I could use the tools like ChatGPT or other AI. I put my thoughts yesterday night and what I wanted to discuss. It gave me a nice summarization of this. So it explains how a knowledge worker or an IP employee, or a programmer or a product manager, how can they influence their job.

So I would say, rather than jobs getting changed, job role descriptions will change. So for me, earlier, as a developer or as a marketing expert, there are few things I used to do, but now everybody can benefit out of using the AI-assisted tool. I can do my work in a better way, enhance my work in a better way, and help the marketing professional or the development effort or team in a better way. So the way I'm saying it is that the next five to seven years, we will see a lot of these inception of these large language models. Large language models are nothing but, again, foundation models. They are dependent on the foundation models. This is nothing but your chance of AI brains sitting in each of these different organizations or different industries, taking all that information, making it available. So we will be getting to a stage where people using the AI-assisted tools will be in a better shape to handle things in a better shape rather than people staying away. So I would ask people to go ahead and try to understand what is happening, what is possible with the automation of business process. There are numerous amounts of websites out there, information out there, like, you can try to understand what is happening and how you can better use it for your growth and not only for your growth but also for your firm or for your company, how you can enhance the user experience through predictive analytics in business.

Host: 

Well, Raghu, thank you so much for teaching us about AI. We've been doing this for a while, so we really appreciate you sort of downloading your knowledge. And if people want to hear more about you, I know that you're doing lots of different things, so if people want to hear more about you, where can they go?

Raghu Banda: 

So one important thing, yeah, so there are a lot of, like you mentioned, yeah, I'm as passionate about learning about AI as well as doing AI. So, in addition to writing the book, I continue the... because the book is once you have done, you have to keep upgrading the things with their book. So what I have to do is that I have to continue doing what I am doing with the book and continuing my conversation. So I also have a blog series, so that I have my continued conversation and to keep things in perspective. I'm also running a podcast, so what I do is that I also, like you are doing, I also invite guests from various different fields or various different AI domains and to keep things in perspective, I have conversations with people.

So for me, there is a huge amount of information out there, right? I think whether it is, there are beautiful blogs written by many people, there are beautiful podcasts done by many people, like you and many other people, and there are numerous amounts of information out there. So these are all the resources that you could go, you could all my podcast series is called Extra AI (X-T-R-A W-E-I). One of the things I’ve been reflecting on a lot recently is how AI in corporate business is rapidly evolving. Businesses are adopting AI for streamlining operations, improving customer experiences, and enhancing decision-making. It’s fascinating to see how companies integrate AI across different functions like marketing, sales, and even HR, transforming traditional processes into more efficient and automated systems.

And then, if we look beyond business, another key focus is on AI in corporate business driving innovation. For instance, AI is playing a huge role in product development, helping companies create smarter and more personalized solutions. It’s clear that AI is not just a tool but a game-changer for the corporate world, reshaping the way companies operate and compete in the market.

‍

Host: 

What's the name of your book?

Raghu Banda: 

The name of the book is "Implementing Machine Learning with SAP S4HANA Core Returns here for SAP."

Host: 

Okay, so this is where we are. There is a lot of information out there and enjoy with their tools and technologies that we have around. Thank you so much for joining.

Raghu Banda: 

Thanks for the opportunity. I had to, and I'll keep a close watch on this podcast as well. I know you have been doing some great conversations and thank you.

‍